From 2e8926802ec2eab58b18284f6ce6582cc6daf002 Mon Sep 17 00:00:00 2001 From: Claude Date: Tue, 30 Jun 2026 22:56:59 +0000 Subject: [PATCH 1/4] Add manuscript scaffold for agentic data-harmonization evaluation Create docs/manuscript/ with a full pre-results paper skeleton plus literature-grounded background: - 03_methods.md: drafted from the locked eval design (config/, src/, skills/) - 01_introduction.md, 02_background.md: drafted and cited from a deep literature survey of (a) the latest agentic-curation / LLM-for-data-work research (2024-2026) and (b) environmental harmonization, data cleaning, ontologies, provenance, and benchmarks - 00_abstract.md, 04_results.md, 05_discussion.md, 06_conclusion.md: scaffolds with [[PLACEHOLDER]] markers tied to the metrics/figures the analysis pipeline will generate - references.md: ~90 sources with per-source verification status; every in-text key resolves - research_notes.md: survey provenance, comparable-work ranking, and pre-submission verification TODOs - figures_and_tables.md, README.md: display-item manifest and section index Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_01FD9C8gFuJGBBYqemVFtwTx --- docs/manuscript/00_abstract.md | 29 +++ docs/manuscript/01_introduction.md | 98 +++++++++ docs/manuscript/02_background.md | 202 ++++++++++++++++++ docs/manuscript/03_methods.md | 235 +++++++++++++++++++++ docs/manuscript/04_results.md | 169 +++++++++++++++ docs/manuscript/05_discussion.md | 112 ++++++++++ docs/manuscript/06_conclusion.md | 19 ++ docs/manuscript/README.md | 58 ++++++ docs/manuscript/figures_and_tables.md | 37 ++++ docs/manuscript/references.md | 285 ++++++++++++++++++++++++++ docs/manuscript/research_notes.md | 150 ++++++++++++++ 11 files changed, 1394 insertions(+) create mode 100644 docs/manuscript/00_abstract.md create mode 100644 docs/manuscript/01_introduction.md create mode 100644 docs/manuscript/02_background.md create mode 100644 docs/manuscript/03_methods.md create mode 100644 docs/manuscript/04_results.md create mode 100644 docs/manuscript/05_discussion.md create mode 100644 docs/manuscript/06_conclusion.md create mode 100644 docs/manuscript/README.md create mode 100644 docs/manuscript/figures_and_tables.md create mode 100644 docs/manuscript/references.md create mode 100644 docs/manuscript/research_notes.md diff --git a/docs/manuscript/00_abstract.md b/docs/manuscript/00_abstract.md new file mode 100644 index 0000000..eab406a --- /dev/null +++ b/docs/manuscript/00_abstract.md @@ -0,0 +1,29 @@ +# Abstract + +> **Status: SCAFFOLD.** To be written last, once results exist. Target ~200–250 +> words. Numbers are `[[PLACEHOLDER]]`. + +**Motivation.** Harmonizing heterogeneous environmental datasets into analysis-ready +form is a labor-intensive, expertise-bound bottleneck in Earth-science synthesis. +`[[1–2 sentences on the soil-moisture harmonization problem and its cost.]]` + +**Approach.** We evaluate a two-agent large language model (LLM) workflow — a +*curator* that ingests a dataset identifier and decides what and how to harmonize, +and a *harmonizer* that generates executable transformation code and a documented +change-mapping — on soil-moisture datasets from the DOE ESS-DIVE repository. We +benchmark it against a domain expert who harmonized 19 Watershed Function SFA +datasets, using grouped leave-one-out cross-validation (retrospective) and blind +prospective evaluation on novel datasets. Our primary endpoint is *output-data +equivalence*: agent and expert code are executed on identical raw inputs and the +resulting tables are compared cell-by-cell, so two correct harmonizations score +equally regardless of coding style. We additionally score each agent in isolation, +the harmonizer under an oracle bundle, and the full pipeline, enabling explicit +attribution of error between the two agents. + +**Results.** `[[PLACEHOLDER: headline end-to-end output equivalence; per-skill +performance; error-propagation gap; dominant failure mode; comparison to +single-call and no-exemplar baselines; similarity-generalization threshold.]]` + +**Conclusions.** `[[PLACEHOLDER: the workflow can/cannot reliably draft expert-grade +harmonizations; recommended human-in-the-loop operating point; implications for +scaling FAIR environmental data curation.]]` diff --git a/docs/manuscript/01_introduction.md b/docs/manuscript/01_introduction.md new file mode 100644 index 0000000..03070bc --- /dev/null +++ b/docs/manuscript/01_introduction.md @@ -0,0 +1,98 @@ +# Introduction + +> **Status:** Drafted. Prose is near-final; citations use author-year keys resolved +> in `references.md`. Any forward reference to a result is marked `[[PLACEHOLDER]]`. + +## The harmonization bottleneck in environmental science + +Synthesizing environmental observations across studies depends on first making +heterogeneous datasets comparable — a process of *harmonization* that maps disparate +source files onto a common schema with consistent variables, units, coordinates, and +time references. For soil moisture, the variable of interest in this work, the same +physical quantity is reported as volumetric water content (in % or m³ m⁻³), +gravimetric water content (g g⁻¹), or matric/water potential (kPa, MPa, or bar); +the gravimetric-to-volumetric conversion itself requires a bulk-density value that is +frequently missing [CEOS-SM; NIDIS-2024]. Sites are georeferenced in incompatible +coordinate systems (UTM vs. latitude/longitude, differing datums), depths follow +different conventions, timestamps are recorded in assorted local time zones, and +tables arrive in both wide and long layouts. Layered on top of these mechanical +differences are scientific judgments: whether a dataset measures soil moisture +directly or derives it from a model, whether observations form a continuous sensor +time series or a discrete sampling campaign, and whether an experimental manipulation +(warming, irrigation) confounds the natural signal. + +Resolving these issues is expert, labor-intensive work, and at scale it remains +largely manual. The International Soil Moisture Network — the field's canonical +harmonization effort, integrating thousands of stations to a common volumetric, UTC +standard — documents that data collection from providers is still done "either +manually (mostly by email) or is automated, depending on the degree of automation at +the data provider side" [Dorigo-2021]. Community pipelines such as AmeriFlux BASE +[Chu-2023] and FLUXNET ONEFlux [Pastorello-2020] show that turning heterogeneous +submissions into a standardized product requires substantial, purpose-built curation +infrastructure. More broadly, data practitioners report spending a large majority of +their time on data preparation and cleaning rather than analysis +[Kandel-2012; Lohr-2014]. This curation cost is a first-order bottleneck on the +reuse of the long tail of environmental data deposited in repositories such as the +U.S. Department of Energy's ESS-DIVE [Crystal-Ornelas-2022]. + +## LLM agents as a route to automated curation + +Large language models, embedded in tool-using *agentic* workflows, have rapidly +become candidates for exactly this kind of work. Recent systems apply LLMs to data +standardization [Qi-2024], autonomous data engineering [Wang-2025], +data preprocessing [Zhang-2024], schema and entity matching [KcMF-2024], and the +generation of executable data-transformation and ETL code [TabulaX-2024; +Prompt2DAG-2025], extending a long programming-by-example lineage from FlashFill +[Gulwani-2011] and the Wrangler/Trifacta tools. Most directly related to the present +work, *Harmonia* casts data harmonization itself as LLM-agent pipeline synthesis, +mapping diverse datasets to a target schema — though it is interactive +(human-in-the-loop), demonstrated on clinical data, and not evaluated against expert +ground truth via cross-validation [Santos-2025]. In parallel, benchmarks for +LLM agents on data tasks consistently find that *end-to-end* data pipelines remain +hard: KramaBench, which spans real-world pipelines including environmental science, +reports that the best system reaches only ~55% end-to-end accuracy [KramaBench-2025]. + +Two further threads frame our design. First, decomposing a task across multiple +agents introduces its own failure modes; recent taxonomies of multi-agent LLM +failures emphasize inter-agent misalignment and the difficulty of *attributing* a +failure to the responsible agent [Cemri-2025]. Second, evaluating stochastic LLM +systems reliably is non-trivial: LLM-as-judge scoring is unstable +[JudgeReliability-2026], motivating automated, execution-grounded metrics with known +answers [Zhang-2025-DataSci]. Our evaluation is built around both concerns. + +## This work + +We study a two-agent workflow, implemented as Anthropic Agent Skills +[AgentSkills-2025], for harmonizing soil-moisture datasets from ESS-DIVE. A +**curator** agent ingests a dataset identifier, inspects the package, decides whether +and how the dataset should be harmonized, and emits a structured bundle; a +**harmonizer** agent consumes that bundle and produces (a) executable Python +transformation code and (b) a documented change-mapping. We benchmark the pair +against a domain expert who harmonized 19 Watershed Function SFA datasets, combining +retrospective grouped leave-one-out cross-validation with prospective blind +evaluation on novel datasets. + +Our evaluation makes three methodological commitments designed to produce a credible, +decomposable estimate of real-world performance: + +1. **Output-data equivalence as the primary endpoint.** Rather than judging code text + or relying on an LLM judge, we *execute* both the agent's and the expert's code on + the same raw inputs and compare the resulting tables cell-by-cell. Two correct + harmonizations score identically regardless of how the code is written. +2. **Decomposed error attribution.** Each agent is scored in isolation, the + harmonizer is scored under an oracle (gold) bundle, and the full pipeline is scored + end-to-end. The gap between the oracle and end-to-end conditions quantifies how + much curator error propagates downstream, and an explicit taxonomy attributes each + end-to-end failure to the curator, the harmonizer, their interface, or genuine + ambiguity. +3. **Honest generalization.** Cross-validation holds out whole source/instrument + clusters, and a per-dataset similarity covariate converts the in-context-exemplar + leakage risk into a measured relationship between dataset novelty and performance. + +Our contributions are: (i) a formal, reproducible evaluation framework for agentic +environmental data harmonization, with an output-equivalence primary endpoint and a +cross-agent error-propagation taxonomy; (ii) retrospective and prospective estimates +of how closely the agent pair reproduces expert harmonization `[[PLACEHOLDER: +headline numbers]]`; and (iii) an analysis of where the workflow succeeds, where and +why it fails, and the dataset-novelty boundary beyond which it should defer to a human +`[[PLACEHOLDER]]`. diff --git a/docs/manuscript/02_background.md b/docs/manuscript/02_background.md new file mode 100644 index 0000000..8724787 --- /dev/null +++ b/docs/manuscript/02_background.md @@ -0,0 +1,202 @@ +# Background and Related Work + +> **Status:** Drafted from the literature research pass (see `research_notes.md` for +> per-source verification status). Author-year keys resolve in `references.md`. +> Sources flagged "snippet-only" there should be confirmed before submission. + +This work sits at the intersection of three literatures: (1) environmental and +Earth-science **data harmonization** practice and the standards it targets; (2) the +classical computer-science literature on **data cleaning, schema matching, and +integration**; and (3) the rapidly moving frontier of **LLM agents for data work**. +We review each in turn and then position our contribution. + +## 1. Harmonization in environmental and Earth science + +### 1.1 The problem and its cost + +Reusing observational environmental data requires reconciling deep heterogeneity in +how the same quantity is recorded. For soil moisture specifically, the same physical +property appears as volumetric water content (% or m³ m⁻³), gravimetric water content +(g g⁻¹), or matric/water potential (kPa, MPa, bar, or pF); converting between +gravimetric and volumetric bases requires a bulk-density value that source datasets +frequently omit [NIDIS-2024; CEOS-SM]. The canonical community effort, the +International Soil Moisture Network (ISMN), addresses this by converting all soil +moisture to volumetric m³ m⁻³ and resampling all series to an hourly UTC reference, +applying automated quality flags rather than deleting suspect values +[Dorigo-2011; Dorigo-2021; Dorigo-2013]. Tellingly for the motivation of this paper, +the ISMN reports that data ingest from providers is still performed "either manually +(mostly by email) or is automated, depending on the degree of automation at the data +provider side" [Dorigo-2021] — i.e., the schema/unit mapping itself remains +substantially manual, with automation largely confined to downstream QC. + +The same pattern holds across networks. AmeriFlux BASE [Chu-2023] and the FLUXNET2015 +ONEFlux pipeline [Pastorello-2020] standardize variable names and units across +hundreds of sites, but only through purpose-built, community-engineered processing +infrastructure. Cosmic-ray neutron networks (COSMOS-UK [Cooper-2021], +COSMOS-Europe [Bogena-2022]) require multi-stage, network-specific correction chains +and harmonized calibration to produce comparable products, and even within a single +"standardized" network (e.g., NEON soil water content [NEON-DP1]) changes in +calibration method over time introduce step discontinuities that complicate reuse. + +This curation burden is an instance of a general phenomenon. Qualitative study of +analysts finds data wrangling and integration to be a dominant, painful bottleneck +[Kandel-2012], and the widely repeated claim that practitioners spend 50–80% of their +time on data preparation traces to a *New York Times* article attributed only to +"interviews and expert estimates" [Lohr-2014] and a small industry survey +[CrowdFlower-2016]; the figure's weak provenance is itself documented [Dodds-2020], +so we cite it as illustrative rather than established. More rigorously, Heidorn's +"dark data in the long tail of science" frames the core risk: the heterogeneous, +hand-curated data of many small projects — exactly the kind harmonized here — is the +most expensive to integrate and the most likely to be lost [Heidorn-2008], and +quantified digitization-labor studies show curation cost is dominated by skilled human +time [Blagoderov-2012]. + +### 1.2 The standards being targeted + +Harmonization is harmonization *toward* something. The datasets in this study come +from the DOE **ESS-DIVE** repository [ESS-DIVE], whose team has developed a family of +community **reporting formats** — the product of reviewing 112 existing standards and +iterating with 247 representatives from 128 institutions — spanning CSV file structure, +file-level metadata, hydrologic monitoring, location metadata, and sample identifiers +[Crystal-Ornelas-2022; Velliquette-2021-CSV; Goldman-2021; Damerow-2021]. Notably, +ESS-DIVE has **no dedicated soil-moisture reporting format**, which is part of what +motivates an automated mapping to a project-defined target schema. The ground-truth +data derive from the Watershed Function SFA and its East River, Colorado community +observatory [Hubbard-2018; Varadharajan-2022], a multiscale hydro-biogeochemical +testbed whose sensor and sample datasets exhibit precisely the file-, unit-, and +coordinate-level heterogeneity our agents must resolve. + +More broadly, environmental data interoperability is served by a stack of information +models, vocabularies, and unit systems that any harmonizer implicitly reconciles: +information models such as ODM2 [Horsburgh-2016], CUAHSI WaterML/HydroShare +[CUAHSI-HIS], OGC Observations & Measurements / SensorThings [OGC-SensorThings], and +the SOSA/SSN ontology [Janowicz-2019]; variable vocabularies such as the CF Standard +Names (which provide canonical identifiers and units for soil-moisture variables) +[CF-Conventions], the NERC vocabulary server [NERC-NVS], ENVO [Buttigieg-2013], and +AGROVOC [AGROVOC]; and unit ontologies such as UCUM [UCUM], QUDT [QUDT], and OM +[Rijgersberg-2013]. All of this is underwritten by the FAIR principles +[Wilkinson-2016], whose interoperability and reusability tenets are, in effect, a +statement of the harmonization problem. Comparable common-schema efforts in adjacent +fields — Darwin Core in biodiversity [Wieczorek-2012], OMOP/OHDSI in health +[OHDSI-OMOP], the Maelstrom guidelines for retrospective cohort harmonization +[Fortier-2017], and large ecological syntheses such as plant-trait harmonization +[Schneider-2020] and BioTIME [Dornelas-2018] — establish both the importance of the +task and the predominantly manual, expert-driven way it is done today. + +### 1.3 Specific heterogeneity hazards + +The transformations our harmonizer must get right are exactly those with a long +history of silent, costly error. Unit mismatches are a classic failure mode in +quantitative science — the loss of NASA's Mars Climate Orbiter to a +pound-force-seconds vs. newton-seconds mismatch being the canonical example +[Mars-Orbiter] — motivating machine-actionable unit systems and conversion libraries +(UCUM, QUDT, OM, Pint) that surface dimensional errors rather than hiding them +[UCUM; QUDT; Rijgersberg-2013; Pint]; in soil science the volumetric–gravimetric +conversion's dependence on bulk density makes this especially treacherous +[Dorigo-2021]. Coordinate-reference heterogeneity (mixed UTM zones, datum +differences, axis order) is handled in software by PROJ/EPSG [PROJ], and automated +geocoordinate cleaning of biodiversity records shows such errors are pervasive at +scale — CoordinateCleaner flagged ~3.6% of ~91M GBIF plant records, including +systematic coordinate-conversion artifacts [Zizka-2019]. Time-zone and UTC conversion +of sensor time series is a frequent multi-hour error source absent explicit ISO-8601 +offsets [ISO-8601], depth conventions differ across sources (standardized depth +intervals such as GlobalSoilMap's being one response [Poggio-2021]), and the +reshaping between wide and long ("tidy") table layouts is itself a recognized, +error-prone operation [Wickham-2014]. Finally, distinguishing genuinely observational +data from experimentally manipulated data (warming, irrigation) is a scientific +judgment with direct downstream consequences: experimental and observational studies +can yield *opposite* directional responses, so naively pooling them confounds any +synthesis [Yue-2017] — a distinction our curator must make explicitly. + +### 1.4 Gold standards and benchmarks + +Evaluating automated harmonization against expert ground truth follows an established +paradigm: schema-matching, ontology-alignment, and entity-resolution work is scored +against expert-curated reference alignments using precision, recall, and F-measure +[OAEI]. The data-management community has built benchmarks for adjacent tasks — +error detection [Abedjan-2016], the impact of cleaning on downstream ML (CleanML) +[CleanML-2021], controlled error injection with known ground truth (BART) +[Arocena-2015], entity matching (DeepMatcher, WDC Products) +[Mudgal-2018; Peeters-2024], and semantic-type / table-to-KG interpretation +(Sherlock, SemTab) [Hulsebos-2019; SemTab]. Most directly, recent benchmarks for +*LLM agents* on data-science and data-engineering tasks — DSBench [Jing-2024], +InfiAgent-DABench [Hu-2024], and KramaBench [KramaBench-2025] — report that current +agents fall well short of expert performance (e.g., the best DSBench agent solves +~34% of data-analysis tasks). These results both motivate our study and frame our +methodological choice to benchmark against a real expert-harmonized corpus with +execution-grounded agreement metrics. + +## 2. Data cleaning, schema matching, and integration + +Our harmonizer is, in classical terms, a system that performs schema matching, value +transformation, and entity/record reconciliation, then emits a transformation +program. Each has a deep literature. Interactive cleaning and wrangling systems — +Potter's Wheel [Raman-2001], Wrangler/Trifacta [Kandel-2011], and OpenRefine +[OpenRefine] — established the value of generating reusable, inspectable +transformation scripts rather than one-off manual edits, and are the conceptual +ancestors of an agent that writes auditable harmonization code. Schema matching has a +canonical taxonomy [Rahm-2001] revisited a decade later [Bernstein-2011]; data +quality and constraint-based repair are surveyed in [Ilyas-2019] and exemplified by +HoloClean's probabilistic-inference approach [Rekatsinas-2017]; and entity resolution +descends from the Fellegi–Sunter model [Fellegi-1969] through managed pipelines such +as Magellan [Konda-2016]. Programming-by-example synthesis of transformations, from +FlashFill onward [Gulwani-2011], is the direct precursor to LLM code generation for +data wrangling. + +The transition to learned and then foundation-model methods is recent and rapid: DITTO +brought pre-trained language models to entity matching [Li-2020]; Narayan et al. showed +that a single foundation model can match or beat task-specific systems across five +cleaning/integration tasks with no fine-tuning ["Can Foundation Models Wrangle Your +Data?"] [Narayan-2022]; Table-GPT improved LLMs' tabular reasoning via table-tuning +[Li-2023]; and empirical studies of LLM schema matching find that prompt context must +be balanced — too little or too much degrades quality [Parciak-2024]. These works +establish that LLMs are credible engines for the *sub-tasks* of harmonization; +our contribution is to evaluate a structured *agent pair* that composes them +end-to-end against executed expert ground truth. + +## 3. LLM agents for data curation and harmonization + +The closest contemporary work casts harmonization itself as agentic pipeline +synthesis. **Harmonia** uses LLM reasoning plus an interactive UI and a library of +harmonization primitives to map diverse datasets to a standard schema, demonstrated on +clinical data [Santos-2025]; it is the most directly comparable system, but it is +human-in-the-loop, in a different domain, and is not evaluated with cross-validation +or cell-level output equivalence. Single-agent standardization and data-engineering +systems are also emerging: CleanAgent decomposes data standardization into iterative +LLM steps [Qi-2024]; Dataforge adds grounding-validation and dual-loop optimization +with cross-validated evaluation [Wang-2025]; Jellyfish instruction-tunes local models +as universal preprocessors [Zhang-2024]; and a line of schema/entity-matching and +ETL-generation work applies LLMs to the mapping and code-generation steps +[KcMF-2024; TabulaX-2024; Prompt2DAG-2025]. On the scientific-curation side, LLMs are +being used to extract and *validate* metadata against ground truth +[MOLE-2025; DCAT-LLM-2025], directly analogous to our conformance and accuracy checks. + +Three further threads shape our methodology. First, **multi-agent failure**: recent +taxonomies (MAST) catalogue how chained LLM systems fail through inter-agent +misalignment and weak verification, and emphasize the difficulty of *attributing* a +failure to the responsible agent [Cemri-2025] — exactly what our oracle-vs-end-to-end +decomposition and error-propagation taxonomy operationalize for a two-agent pipeline. +Second, **evaluation of agents on data tasks**: benchmarks such as DataSciBench +[Zhang-2025-DataSci] and KramaBench [KramaBench-2025] grapple with uncertain ground +truth and execution-based scoring, and KramaBench — which spans environmental science — +finds the best system reaches only ~55% end-to-end accuracy, underscoring that this is +an unsolved problem. Third, **the unreliability of LLM-as-judge** scoring +[JudgeReliability-2026] motivates our reliance on automated, execution-grounded +cell-level metrics with expert validation rather than model-graded outputs. Our agents +are implemented as Anthropic **Agent Skills** [AgentSkills-2025], a structured +instruction-and-tool packaging in which the in-context exemplars (the 19 expert +harmonizations) function as retrievable references. + +## 4. Positioning of this work + +Against this backdrop, our contribution is methodological and empirical rather than a +new system per se. Where Harmonia and CleanAgent demonstrate feasibility, and where +data-task benchmarks report aggregate success rates, we provide a **rigorous, +decomposable evaluation** of an autonomous curator–harmonizer pair on real +environmental data: an output-data-equivalence primary endpoint (executed code +compared cell-by-cell against an expert's), grouped cross-validation with a +similarity covariate to give honest generalization estimates, and a cross-agent +error-propagation taxonomy that attributes each failure to its source. To our +knowledge this is the first such evaluation in the environmental-data-harmonization +setting, and it directly addresses the verification and attribution gaps highlighted +by the multi-agent and agent-evaluation literature. diff --git a/docs/manuscript/03_methods.md b/docs/manuscript/03_methods.md new file mode 100644 index 0000000..ac7fd60 --- /dev/null +++ b/docs/manuscript/03_methods.md @@ -0,0 +1,235 @@ +# Methods + +> **Status:** Drafted from the locked evaluation design (`config/`, `src/`, `skills/`). +> All numbers, fold counts, and model identifiers are placeholders pending the +> final experimental run and should be reconciled against `config/experiment.yaml` +> and `config/cv_folds.yaml` before submission. + +## 1. Overview + +We evaluate a two-agent large language model (LLM) workflow that automates the +harmonization of heterogeneous soil-moisture datasets into a single analysis-ready +schema. The workflow is a *curator* agent (Skill 1) followed by a *harmonizer* +agent (Skill 2). The curator ingests a dataset identifier, retrieves and inspects +the package, makes an inclusion/exclusion decision, and emits a structured bundle; +the harmonizer consumes that bundle and emits (a) executable Python transformation +code and (b) a documented change-mapping. We measure how well this pair reproduces +the work of a domain expert, both retrospectively (cross-validation against a fixed +gold standard) and prospectively (blind comparison on previously unseen datasets). + +The design has three deliberate properties: + +1. **Output-centric scoring.** The primary endpoint is *output-data equivalence* — + the harmonized table produced by the agent's code is compared cell-by-cell with + the table produced by the expert's code on the *same* raw inputs. Two correct + harmonizations score identically even if the underlying Python differs entirely, + so we measure *what the code does*, not how it is written. +2. **Decomposable error attribution.** Each skill is scored in isolation and the + pipeline is scored end-to-end, allowing us to separate curator error from + harmonizer error and to quantify error propagation between the two agents. +3. **Honest generalization estimates.** Cross-validation holds out whole + source/instrument clusters, and a per-dataset similarity covariate converts the + in-context-exemplar leakage risk into a measured effect rather than an + uncontrolled confound. + +## 2. The agent pair under evaluation + +Both agents are implemented as Anthropic *Agent Skills* — versioned instruction +bundles (system prompt plus reference context) invoked inside an agentic +tool-using loop. Skill versions are pinned per run (`skill1_version`, +`skill2_version` in `config/experiment.yaml`). + +### 2.1 Skill 1 — Curator (`essdive_sm_curator`) + +Given one or more DOIs / ESS-DIVE package identifiers, the curator: + +- normalizes identifiers and retrieves package metadata (local cache first, then + the ESS-DIVE API); +- classifies each file into *data payload*, *location metadata*, *sensor metadata*, + or *documentation*; +- inspects payload headers to locate timestamp, moisture (VWC / GWC / water + potential), depth, site, and replicate columns; +- resolves site coordinates through a fixed fallback cascade (payload → ancillary + file → package metadata → external location registry → unresolvable), assigning a + `qc_flag` recommendation; +- infers time-series vs. discrete sampling and, for time series, a sampling + interval; +- detects experimental manipulations (warming, irrigation, fertilization, etc.) and + recommends include / exclude / flag; +- selects the most structurally similar gold exemplar; and +- emits a structured **curator bundle** (`src/schemas/skill1_bundle.py`) with an + `INCLUDE` / `EXCLUDE` / `FLAG_FOR_REVIEW` decision and a list of open questions. + +### 2.2 Skill 2 — Harmonizer (`wfsfa_sm_harmonization`) + +Given a curator bundle (or an operator-supplied equivalent), the harmonizer: + +- re-confirms the inclusion decision under a strict rule set; +- maps each source variable to the target schema, including unit conversions; +- resolves location, depth encoding, timestamp format/timezone, and replicate + encoding; +- generates a Python code block following fixed conventions (shared helper + functions, wide-to-long reshaping, UTC conversion, column-order enforcement); and +- emits a JSON **change-mapping** entry (`src/schemas/skill2_mapping.py`) + documenting every transformation with source pattern, destination variable, + transformation description, and unit conversion. + +### 2.3 Target schema + +Every harmonized table must contain exactly nine columns +(`src/schemas/target_schema.py`): + +`datetime_UTC`, `site_id`, `depth_m`, `replicate`, `is_timeseries`, +`interval_min`, `volumetric_water_content_m3_m3`, +`gravimetric_water_content_gH2O_gs`, `water_potential_kPa`. + +Canonical units are meters (depth), m³ m⁻³ (VWC), g H₂O g⁻¹ soil (GWC), and kPa +(water potential, negative). A controlled `qc_flag` vocabulary records depth +approximation (`d1`) and coordinate provenance (`g1` retrieved from an external +registry, `g2` unresolvable). Schema conformance is validated programmatically +(column presence/exclusivity, dtypes, value ranges, required-field non-nullity). + +## 3. Ground truth + +The reference standard is a corpus of **19 expert-harmonized** Watershed Function +SFA (WFSFA) soil-moisture datasets drawn from ESS-DIVE. For each, the expert +produced (a) a harmonized output table, (b) modular, documented Python +transformation code (`data/gold/expert_code/harmonize_sm/`, one +`dataset_NN.py` per dataset plus shared `common.py`), and (c) a structured + +free-text change-mapping (`data/gold/sm_data_harmonization_mapping.json`). The +mapping file also encodes excluded datasets with their plain-language exclusion +reasons, providing gold labels for the curator's decision task. + +> **TODO (results-time):** confirm the exact count of *included* vs. *excluded* +> datasets in the final gold set and report it here; the mapping JSON currently +> enumerates indices 0–27 with several excluded/auxiliary entries. + +## 4. Study design + +The study has two phases; within each, the agent pair is evaluated in three modes. + +### 4.1 Phase A — Retrospective cross-validation + +We perform **grouped leave-one-out cross-validation** over the gold datasets. +Datasets are grouped into source/instrument **clusters** (`config/cv_folds.yaml`) +so that datasets sharing a lab, site, or instrument family are held out together; +this prevents a near-duplicate sibling from leaking into the exemplar pool and +inflating apparent performance. For each fold, the held-out dataset(s) are +harmonized using only the *other* datasets as exemplars, and outputs are scored +against the expert version. + +Phase A serves a second purpose: it **validates the automated metrics** against +expert judgment. If automated output-equivalence scoring agrees with an expert +rubric on the gold data, we can rely on the cheaper automated scoring in Phase B. + +### 4.2 Phase B — Prospective blind evaluation + +Using all gold datasets as exemplars, the agent harmonizes genuinely new ESS-DIVE +datasets. In parallel and blind to the agent's output, the domain expert harmonizes +the same datasets; the expert output becomes the reference. This estimates +real-world deployment performance on data neither the agent nor the expert has +previously processed. + +### 4.3 Evaluation modes + +Both phases score the system three ways to localize error: + +1. **Skill 1 in isolation** — the curator bundle is scored field-by-field against + expert labels. +2. **Skill 2 with oracle input** — the harmonizer is fed the *correct* (gold) + curator bundle, so transformation quality is measured independently of curator + error. +3. **End-to-end** — the full pipeline runs and errors propagate. The gap between + modes (2) and (3) quantifies how much curator error degrades the final output. + +### 4.4 Baselines and controls + +- **Non-agentic single-call LLM** — one prompt, no skill structure or tool loop. +- **No-exemplars agent** — the agent with the gold exemplars removed, isolating the + value of the 19 references. +- **Naïve heuristic** — a string-matching baseline. +- **Similarity covariate** — for each evaluation dataset we compute similarity to + its nearest available exemplar and model it explicitly, turning a leakage risk + into a reported finding ("performance holds up to similarity *X*, then degrades"). +- **Reproducibility controls** — model ID/version, both skill versions, sampling + parameters, and random seed are pinned per run (`config/experiment.yaml`). + +> **TODO (results-time):** record the exact model identifier and version actually +> used (config currently pins a Claude model), temperature, `max_tokens`, and the +> number of stochastic repeats per dataset/mode (`n_stochastic_runs`). + +## 5. Metrics + +All scoring lives in `src/metrics/` and is phase-agnostic. + +### 5.1 Skill 1 (curator) + +Decision accuracy / precision / recall / F1 over `INCLUDE` / `EXCLUDE` / `FLAG`; +file-set precision/recall/F1 for payload, location, and sensor files; time-series +binary accuracy and sampling-interval numeric error; location-source and `qc_flag` +accuracy; manipulation-detection accuracy; exemplar-selection match quality; and +deferral (FLAG) calibration against expert confidence. A composite curator score +aggregates these with fixed, versioned weights (`config/metrics_weights.yaml`). + +### 5.2 Skill 2 (harmonizer) + +- **Primary endpoint — output-data equivalence.** Agent and expert code are each + executed on the same raw inputs in a sandbox; the resulting tables are aligned on + the natural key (`datetime_UTC`, `site_id`, `depth_m`, `replicate`) and compared. + We report row-alignment precision/recall/F1 and per-column cell agreement + (numeric columns within a float tolerance; categorical columns by exact match), + with sub-scores for datetime, coordinate, depth, and unit-conversion accuracy. +- **Schema conformance** — columns, dtypes, ranges, required fields. +- **Semantic mapping accuracy** — precision/recall/F1 of the change-mapping against + the expert's, over a controlled transformation-type vocabulary. +- **Code executability** — does the generated code run and reproduce its claimed + output. +- **Documentation completeness** — does the change-mapping cover all and only the + actual transformations. +- **Ontology / controlled-vocabulary validity** — correct use of schema terms and + `qc_flag` codes. + +A composite harmonizer score weights output equivalence most heavily +(`config/metrics_weights.yaml`). + +### 5.3 End-to-end + +Task success (output equivalence above threshold) plus an **error-propagation +taxonomy** classifying each failure as (i) Skill-1 error propagated, (ii) Skill-2 +error given correct input, (iii) inter-skill interface inconsistency, or (iv) +genuinely ambiguous (expert also uncertain). Classification logic is in +`src/analysis/error_taxonomy.py`. + +## 6. Statistical analysis + +The data are nested: stochastic runs within datasets within source-clusters. To +avoid inflating the effective sample size we: + +- report **cluster-level bootstrap confidence intervals** (resampling clusters, not + individual runs); +- fit **mixed-effects models** with dataset and cluster as random effects and mode + and similarity as fixed effects + (`metric ~ mode + similarity + (1|dataset) + (1|cluster)`); +- report the **error-propagation gap** (oracle minus end-to-end) as a paired + difference with a cluster-bootstrap CI; +- emphasize **effect sizes with CIs** over *p*-values, given the small number of + gold datasets (N ≈ 19); and +- report the **distribution across stochastic repeats** (variance, and pass@k where + appropriate), not only means. + +Inter-rater reliability between automated metrics and the expert rubric is reported +for the metric-validation analysis (`src/analysis/irr.py`). + +## 7. Reproducibility and provenance + +Per-run artifacts (raw agent outputs, scored tidy CSVs, figures, tables) are written +under `results/`. Each run records the model identifier/version, both skill +versions, sampling parameters, and the random seed. The gold corpus is treated as +read-only. Code execution occurs in a sandbox (`src/execution/sandbox.py`) and +output comparison in `src/execution/output_loader.py`. + +> **Reporting checklist (fill at results-time):** model + version; skill versions; +> temperature / sampling; seeds; `n_stochastic_runs`; number of folds and held-out +> clusters; counts of included/excluded gold datasets; number and provenance of +> prospective (Phase B) datasets; software/library versions; total compute / token +> budget. diff --git a/docs/manuscript/04_results.md b/docs/manuscript/04_results.md new file mode 100644 index 0000000..8465a62 --- /dev/null +++ b/docs/manuscript/04_results.md @@ -0,0 +1,169 @@ +# Results + +> **Status: SCAFFOLD ONLY — no results yet.** This section is structured so that +> figures, tables, and numbers can be dropped in once the experimental runs +> complete. Every `[[PLACEHOLDER]]` marks a value, table, or figure to be generated +> by the analysis pipeline. Do not report any number here that is not produced by +> `experiments/` and written under `results/`. + +## 0. Reporting conventions + +All point estimates are reported with 95% cluster-level bootstrap confidence +intervals in brackets. Cell-agreement and F1 metrics are fractions in [0, 1]. +"Pass" denotes output equivalence at or above the configured threshold +(`output_equivalence_min`, currently 0.95). Unless stated otherwise, per-mode +numbers pool stochastic repeats and we also report run-to-run variance. + +## 1. Gold corpus characterization + +> Descriptive, generated by `notebooks/01_explore_gold.ipynb`. + +- Number of gold datasets, and the include/exclude split: `[[PLACEHOLDER]]`. +- Distribution of structural types (wide vs. long; time-series vs. discrete; + coordinate system; unit conventions; presence of separate location/sensor files): + **Table 1** `[[PLACEHOLDER]]`. +- Cluster membership and sizes (`config/cv_folds.yaml`): `[[PLACEHOLDER]]`. +- Pairwise dataset-similarity distribution (motivates the grouped-CV design): + **Figure 1** `[[PLACEHOLDER]]`. + +## 2. Phase A — Retrospective cross-validation + +### 2.1 Skill 1 (curator) in isolation + +> Generated by `experiments/phase_a_crossval.py` (mode `skill1_isolated`), +> scored by `src/metrics/skill1_metrics.py`. + +**Table 2 — Curator performance.** `[[PLACEHOLDER]]` + +| Metric | Score [95% CI] | +|---|---| +| Decision accuracy (INCLUDE/EXCLUDE/FLAG) | `[[ ]]` | +| Decision macro-F1 | `[[ ]]` | +| File-selection F1 — payload / location / sensor | `[[ ]]` | +| Time-series binary accuracy | `[[ ]]` | +| Sampling-interval error (min) | `[[ ]]` | +| Location-source accuracy | `[[ ]]` | +| `qc_flag` accuracy | `[[ ]]` | +| Manipulation-detection accuracy | `[[ ]]` | +| Exemplar-selection match | `[[ ]]` | +| Deferral (FLAG) calibration | `[[ ]]` | +| **Composite curator score** | `[[ ]]` | + +- Confusion matrix over the three decisions: **Figure 2** `[[PLACEHOLDER]]`. +- Where does the curator most often err (file selection vs. time-series vs. + location)? `[[PLACEHOLDER narrative]]`. + +### 2.2 Skill 2 (harmonizer) with oracle input + +> Mode `skill2_oracle`; primary scoring by `src/metrics/skill2_output_equiv.py`. + +**Table 3 — Harmonizer performance given correct curator input.** `[[PLACEHOLDER]]` + +| Metric | Score [95% CI] | +|---|---| +| **Output-data equivalence (overall cell agreement)** | `[[ ]]` | +| Row-alignment F1 | `[[ ]]` | +| Datetime accuracy | `[[ ]]` | +| Coordinate accuracy | `[[ ]]` | +| Depth accuracy | `[[ ]]` | +| Unit-conversion accuracy | `[[ ]]` | +| Schema conformance (pass rate) | `[[ ]]` | +| Semantic mapping F1 | `[[ ]]` | +| Code executability (pass rate) | `[[ ]]` | +| Documentation completeness | `[[ ]]` | +| **Composite harmonizer score** | `[[ ]]` | +| **Output-equivalence pass rate (≥ threshold)** | `[[ ]]` | + +- Per-column cell-agreement breakdown: **Figure 3** `[[PLACEHOLDER]]`. +- Distribution of per-dataset output equivalence (how many are near-perfect vs. + partial): **Figure 4** `[[PLACEHOLDER]]`. + +### 2.3 End-to-end pipeline + +> Mode `end_to_end`. + +**Table 4 — End-to-end vs. oracle.** `[[PLACEHOLDER]]` + +| Metric | Oracle (Skill 2) | End-to-end | Gap [95% CI] | +|---|---|---|---| +| Output-data equivalence | `[[ ]]` | `[[ ]]` | `[[ ]]` | +| Output-equivalence pass rate | `[[ ]]` | `[[ ]]` | `[[ ]]` | +| Composite score | `[[ ]]` | `[[ ]]` | `[[ ]]` | + +The **error-propagation gap** (oracle − end-to-end) is `[[PLACEHOLDER]]` +(`src/analysis/stats.py::error_propagation_gap`), quantifying the fraction of +end-to-end degradation attributable to curator error. + +### 2.4 Error-propagation taxonomy + +> `src/analysis/error_taxonomy.py`. + +**Figure 5 / Table 5 — Attribution of end-to-end failures.** `[[PLACEHOLDER]]` + +| Failure category | Share of failures | +|---|---| +| Skill-1 error propagated | `[[ ]]` | +| Skill-2 error (correct input) | `[[ ]]` | +| Inter-skill interface inconsistency | `[[ ]]` | +| Genuinely ambiguous (expert also uncertain) | `[[ ]]` | + +### 2.5 Baselines + +> `config/experiment.yaml::baselines`. + +**Table 6 — Agent pair vs. baselines** (output equivalence, composite). +`[[PLACEHOLDER]]` + +| System | Output equivalence [95% CI] | Pass rate | +|---|---|---| +| Agent pair (end-to-end) | `[[ ]]` | `[[ ]]` | +| Single-call LLM (non-agentic) | `[[ ]]` | `[[ ]]` | +| Agent without exemplars | `[[ ]]` | `[[ ]]` | +| Naïve string-matching heuristic | `[[ ]]` | `[[ ]]` | + +Key contrasts to report: agent-pair vs. single-call (value of agentic structure); +agent vs. no-exemplars (value of the 19 references). + +### 2.6 Similarity–performance relationship + +> `src/analysis/similarity.py`; mixed-effects model in `src/analysis/stats.py`. + +- Output equivalence vs. nearest-exemplar similarity, with regression fit: + **Figure 6** `[[PLACEHOLDER]]`. +- Mixed-effects coefficients (mode, similarity) with CIs: **Table 7** + `[[PLACEHOLDER]]`. +- Interpretation: the agent generalizes up to similarity `[[X]]`, beyond which + output equivalence degrades to `[[Y]]`. `[[PLACEHOLDER narrative]]`. + +### 2.7 Stochasticity + +- Run-to-run variance per dataset/mode and pass@k: **Figure 7** `[[PLACEHOLDER]]`. + +## 3. Metric validation + +> `experiments/metric_validation.py`; `src/analysis/irr.py`. + +Agreement between automated output-equivalence scoring and an expert rubric on the +gold data: correlation `[[PLACEHOLDER]]`, inter-rater reliability +`[[PLACEHOLDER]]`. This determines whether automated scoring can stand in for +expert judgment in Phase B. **Figure 8** `[[PLACEHOLDER]]`. + +## 4. Phase B — Prospective blind evaluation + +> `experiments/phase_b_prospective.py`. To be run once novel datasets and the +> expert's blind harmonizations are available. + +- Number and provenance of prospective datasets: `[[PLACEHOLDER]]`. +- Agent vs. expert output equivalence on novel data: **Table 8** `[[PLACEHOLDER]]`. +- Comparison of Phase B performance to the Phase A end-to-end estimate (does + retrospective CV predict prospective performance?): `[[PLACEHOLDER narrative]]`. +- Curator decision agreement with the expert on novel inclusion/exclusion calls: + `[[PLACEHOLDER]]`. + +## 5. Summary of findings + +> One-paragraph synthesis to be written once the tables above are populated. +> Anticipated structure: headline output-equivalence number end-to-end; the size of +> the curator-induced error-propagation gap; the dominant failure category; the +> value added by agentic structure and by exemplars (vs. baselines); and the +> similarity threshold beyond which the system should defer to a human. diff --git a/docs/manuscript/05_discussion.md b/docs/manuscript/05_discussion.md new file mode 100644 index 0000000..f550363 --- /dev/null +++ b/docs/manuscript/05_discussion.md @@ -0,0 +1,112 @@ +# Discussion + +> **Status:** Structural scaffold. Argument skeleton and limitations are derivable +> from the design now; quantitative claims (`[[PLACEHOLDER]]`) are filled at +> results-time. Cross-references to related work (`§` in `02_background.md`) are +> marked `[[cite]]` and resolved against `references.md`. + +## 1. Principal findings + +> One or two paragraphs once results exist. Anticipated claims, each contingent on +> the data: + +- The agent pair reproduces expert harmonization with end-to-end output equivalence + of `[[PLACEHOLDER]]`, with `[[PLACEHOLDER]]`% of datasets meeting the strict + cell-agreement threshold. +- Given correct curator input, the harmonizer alone reaches `[[PLACEHOLDER]]`; the + error-propagation gap of `[[PLACEHOLDER]]` shows that `[[most / little]]` of the + end-to-end shortfall originates upstream in curation. +- The dominant failure mode is `[[category]]`, concentrated in + `[[structural type / unit / location resolution]]`. + +## 2. Why decompose the workflow + +Scoring each skill in isolation, the harmonizer under an oracle bundle, and the full +pipeline lets us attribute error rather than report an opaque success rate. This +matters for deployment: a system that fails mostly through *curator* errors (file +selection, inclusion decisions) calls for a different intervention — a human gate at +curation — than one that fails through *harmonizer* errors (unit conversions, +timezone handling). The error-propagation taxonomy operationalizes that distinction +and connects to broader work on chained/multi-agent LLM pipelines `[[cite]]`, where +compounding error across stages is a known failure mode. + +## 3. Output equivalence as the right endpoint + +Comparing executed outputs rather than code (or rather than an LLM-judge score) +sidesteps two problems: stylistic code variation (two correct programs differ +arbitrarily) and the reliability concerns of LLM-as-judge evaluation `[[cite]]`. It +also gives a domain-meaningful unit — the fraction of harmonized cells a downstream +scientist could trust. The cost is that output equivalence requires both code +artifacts to *execute* on shared raw data, which is why code executability is scored +separately and why sandboxed execution is part of the harness. + +## 4. Generalization and the exemplar question + +In-context exemplars are the mechanism by which the agent learns the target schema, +yet they also create a leakage risk under cross-validation: a held-out dataset with +a near-duplicate sibling in the exemplar pool would overstate performance. Grouped +cross-validation (holding out whole source/instrument clusters) and the explicit +similarity covariate convert this risk into a measured relationship. The +similarity–performance curve `[[PLACEHOLDER]]` indicates the boundary of reliable +generalization and, practically, where the system should defer to a human. The +no-exemplars baseline `[[PLACEHOLDER]]` separately quantifies how much the reference +corpus contributes at all. + +## 5. Relation to prior work + +> Resolve against `02_background.md` once research citations are integrated. + +- **Vs. classic data-wrangling / ETL synthesis** (`[[cite]]`): how the agentic, + schema-targeted, exemplar-driven approach differs from program-synthesis and + rule-based wrangling tools. +- **Vs. LLM-based schema/entity matching and data cleaning** (`[[cite]]`): we + evaluate end-to-end *transformation-code generation against executed expert + ground truth*, not a matching/labeling decision in isolation. +- **Vs. environmental-science harmonization practice** (`[[cite]]`): contrast with + manual, script-per-dataset harmonization and with standardized reporting formats / + ontologies; position the agent pair as an accelerator that still produces + human-auditable code and documented mappings. + +## 6. Limitations + +- **Small gold corpus.** With ≈19 expert-harmonized datasets, estimates are + uncertain; we report effect sizes with cluster-bootstrap CIs and avoid + over-reliance on *p*-values, but power is inherently limited. +- **Single domain, single repository.** The corpus is WFSFA soil moisture on + ESS-DIVE. Generalization to other variables, repositories, or Earth-science + domains is untested and is an explicit direction for future work. +- **Expert as gold standard.** The reference is one expert's harmonization; some + decisions are genuinely ambiguous (captured by the taxonomy's fourth category and + by expert-confidence labels), and a second independent expert would strengthen the + ground truth. +- **Model and skill drift.** Results are pinned to a specific model version and + skill versions; LLM and skill updates may change performance, so we report exact + versions and seeds for reproducibility. +- **Stochasticity.** LLM outputs vary run-to-run; we report variance and pass@k, but + any single deployed run may differ from the reported mean. +- **Automated-metric validity.** Phase B leans on automated scoring; its credibility + rests on the Phase A metric-validation agreement `[[PLACEHOLDER]]`, which bounds + how far the automated numbers can be trusted. +- **Prospective scope.** Phase B covers `[[N]]` novel datasets; broader prospective + deployment is needed to confirm the retrospective estimate. + +## 7. Implications and future work + +- **Human-in-the-loop deployment.** The natural operating point is agent-drafts / + expert-reviews, with automatic deferral above the similarity threshold and on + low-confidence curator decisions (the FLAG path). Report the implied human-effort + savings `[[PLACEHOLDER]]`. +- **Beyond soil moisture.** Extend to other ESS-DIVE reporting formats and to other + environmental variables; test whether the curator/harmonizer split transfers. +- **Richer provenance.** The change-mapping JSON is a step toward machine-readable, + FAIR-aligned transformation provenance `[[cite]]`; future work could emit + standardized provenance directly. +- **Active exemplar selection.** Given the similarity findings, selecting or + synthesizing exemplars to cover structural gaps may extend the generalization + frontier. + +## 8. Conclusion + +> Short — typically merged into `06_conclusion.md`. Restate the headline +> output-equivalence result, the decomposed error attribution, and the practical +> deployment recommendation, all contingent on the final numbers. diff --git a/docs/manuscript/06_conclusion.md b/docs/manuscript/06_conclusion.md new file mode 100644 index 0000000..1711833 --- /dev/null +++ b/docs/manuscript/06_conclusion.md @@ -0,0 +1,19 @@ +# Conclusion + +> **Status: SCAFFOLD.** Write last, ~1 paragraph. Restate the contribution and the +> headline result without hedging once numbers exist. + +We presented a rigorous evaluation of a two-agent LLM workflow — a curator and a +harmonizer — for automating the harmonization of heterogeneous environmental +(soil-moisture) datasets into an analysis-ready schema, benchmarked against expert +ground truth via retrospective grouped cross-validation and prospective blind +evaluation. By scoring output-data equivalence (executed code, compared cell-by-cell +against the expert's), decomposing performance across the two agents, and attributing +end-to-end failures with an explicit taxonomy, the study `[[PLACEHOLDER: states what +was learned — the achievable accuracy, where error originates, and the operating +point at which the system is deployable with human oversight]]`. + +`[[PLACEHOLDER closing sentence on broader significance: a path toward scalable, +auditable, FAIR-aligned curation of long-tail environmental data, with the agent +producing human-reviewable transformation code and documented mappings rather than +opaque outputs.]]` diff --git a/docs/manuscript/README.md b/docs/manuscript/README.md new file mode 100644 index 0000000..38ad6aa --- /dev/null +++ b/docs/manuscript/README.md @@ -0,0 +1,58 @@ +# Manuscript + +Working drafts for the paper describing this study: +**Evaluating an LLM Agent Pair for Environmental Data Harmonization.** + +> **Status: pre-results scaffold (2026-06).** The methods and background are +> drafted; results, discussion, and the abstract contain `[[PLACEHOLDER]]` markers +> to be filled once the experimental pipeline (`experiments/`) has produced scored +> outputs under `results/`. Nothing here reports a number that the analysis code has +> not yet generated. + +## Files + +| File | Section | Status | +|---|---|---| +| `00_abstract.md` | Abstract | Scaffold — write last | +| `01_introduction.md` | Introduction & motivation | Drafted (cited) | +| `02_background.md` | Background / related work | Drafted (cited) | +| `03_methods.md` | Methods | Drafted from locked design | +| `04_results.md` | Results | Scaffold — tables/figures stubbed | +| `05_discussion.md` | Discussion | Scaffold — argument skeleton | +| `06_conclusion.md` | Conclusion | Scaffold | +| `figures_and_tables.md` | Display-item manifest | Mapped to analysis code | +| `references.md` | Bibliography | Drafted (research-backed) | +| `research_notes.md` | Raw literature research (provenance/verification) | Reference material | + +Read in numeric order for a continuous draft. The intended journal/venue is TBD +(candidates: *Scientific Data*, *Environmental Modelling & Software*, *PVLDB / +SIGMOD data-systems venues*, or an agentic-AI venue — decide once results frame the +strongest contribution). + +## How this maps to the codebase + +- **Methods** are derived from `config/` (experiment, folds, metric weights), + `src/schemas/` (target schema, curator/harmonizer bundles), `src/metrics/`, + `src/analysis/`, and `skills/` (the two agents under evaluation). +- **Results** display items each name the `experiments/…` or `src/…` code that + generates them (see `figures_and_tables.md`). +- **Ground truth** is `data/gold/` (19 expert-harmonized datasets: code + mappings). + +## Filling in results + +1. Run `experiments/phase_a_crossval.py` → `results/scored/phase_a_results.csv`. +2. Run `experiments/metric_validation.py` (validates automated metrics vs. expert). +3. Run `experiments/phase_b_prospective.py` once novel + expert-blind data exist. +4. Generate figures/tables from `notebooks/` and `src/analysis/`. +5. Replace every `[[PLACEHOLDER]]` in `04_results.md`, then write `00_abstract.md`, + the discussion's quantitative claims, and `06_conclusion.md`. +6. Reconcile the reporting checklist at the end of `03_methods.md` (model version, + seeds, counts). + +## Citation policy + +`references.md` separates sources the research pass **verified by fetching** from +those captured **from search snippets only**; the latter are flagged and must be +confirmed before submission. Author-year keys in the prose (e.g. `[Dorigo-2021]`) +resolve to entries in `references.md`. Do not introduce a citation that is not in +`references.md`. diff --git a/docs/manuscript/figures_and_tables.md b/docs/manuscript/figures_and_tables.md new file mode 100644 index 0000000..4bcd48b --- /dev/null +++ b/docs/manuscript/figures_and_tables.md @@ -0,0 +1,37 @@ +# Figures and Tables — Manifest + +> Planned display items, with the analysis code/path that generates each. Keep this +> in sync with `04_results.md`. Outputs are written under `results/figures/` and +> `results/tables/`. + +## Tables + +| # | Title | Source | Status | +|---|---|---|---| +| 1 | Gold-corpus structural characterization (format, units, CRS, time-series share) | `notebooks/01_explore_gold.ipynb` | TODO | +| 2 | Curator (Skill 1) performance, all sub-metrics + composite | `phase_a_crossval.py` → `skill1_metrics.py` | TODO | +| 3 | Harmonizer (Skill 2, oracle) performance, output-equivalence breakdown | `skill2_output_equiv.py`, `skill2_*` | TODO | +| 4 | End-to-end vs. oracle, with error-propagation gap | `stats.py::error_propagation_gap` | TODO | +| 5 | Error-propagation taxonomy (failure attribution shares) | `error_taxonomy.py` | TODO | +| 6 | Agent pair vs. baselines (single-call, no-exemplar, naïve) | `phase_a_crossval.py` (baselines) | TODO | +| 7 | Mixed-effects model coefficients (mode, similarity) | `stats.py::mixed_effects_comparison` | TODO | +| 8 | Phase B prospective: agent vs. expert on novel data | `phase_b_prospective.py` | TODO | + +## Figures + +| # | Title | Source | Status | +|---|---|---|---| +| 1 | Pairwise dataset-similarity distribution (motivates grouped CV) | `similarity.py` | TODO | +| 2 | Curator decision confusion matrix | `skill1_metrics.py` | TODO | +| 3 | Per-column cell-agreement (harmonizer) | `skill2_output_equiv.py` | TODO | +| 4 | Per-dataset output-equivalence distribution | `phase_a_crossval.py` | TODO | +| 5 | Error-propagation taxonomy (stacked bar / Sankey) | `error_taxonomy.py` | TODO | +| 6 | Output equivalence vs. nearest-exemplar similarity (+ regression) | `similarity.py`, `stats.py` | TODO | +| 7 | Run-to-run variance / pass@k per mode | `phase_a_crossval.py` | TODO | +| 8 | Automated-metric vs. expert-rubric agreement | `metric_validation.py`, `irr.py` | TODO | + +## Schematic (no data needed — can be drafted now) + +| # | Title | Notes | +|---|---|---| +| 0 | Workflow + evaluation design schematic | Curator → bundle → Harmonizer → code+mapping → executed output vs. expert; three modes; two phases. Good candidate for **Figure 1 (main)**. Draft as a diagram (e.g., Mermaid/Figma) before results. | diff --git a/docs/manuscript/references.md b/docs/manuscript/references.md new file mode 100644 index 0000000..9e97679 --- /dev/null +++ b/docs/manuscript/references.md @@ -0,0 +1,285 @@ +# References + +> Author-year keys match the in-text citations in `01_introduction.md` and +> `02_background.md`. Verification status from the literature research pass is noted: +> **[V]** = a primary source was fetched and confirmed; **[S]** = corroborated from +> search snippets / authoritative secondary pages but not directly fetched (confirm +> DOI and author order before submission); **[org]** = official standards/repository +> site. See `research_notes.md` for fuller provenance notes. +> +> A `.bib` export should be generated before submission; this file is the +> human-readable working bibliography. + +## Environmental / Earth-science harmonization, networks, and standards + +- **[Dorigo-2011]** Dorigo, W. A., et al. (2011). The International Soil Moisture + Network: a data hosting facility for global in situ soil moisture measurements. + *Hydrology and Earth System Sciences* 15:1675–1698. **[S]** + https://doi.org/10.5194/hess-15-1675-2011 +- **[Dorigo-2013]** Dorigo, W. A., et al. (2013). Global Automated Quality Control of + In Situ Soil Moisture Data from the International Soil Moisture Network. *Vadose + Zone Journal* 12(3). **[S]** https://doi.org/10.2136/vzj2012.0097 +- **[Dorigo-2021]** Dorigo, W., et al. (2021). The International Soil Moisture Network: + serving Earth system science for over a decade. *Hydrology and Earth System + Sciences* 25:5749–5804. **[V]** https://doi.org/10.5194/hess-25-5749-2021 · + harmonization docs: https://ismn.earth/en/data/harmonization/ +- **[Chu-2023]** Chu, H., Christianson, D. S., Cheah, Y.-W., et al. (2023). AmeriFlux + BASE data pipeline to support network growth and data sharing. *Scientific Data* + 10:614. **[V]** https://doi.org/10.1038/s41597-023-02531-2 +- **[Pastorello-2020]** Pastorello, G., et al. (2020). The FLUXNET2015 dataset and the + ONEFlux processing pipeline for eddy covariance data. *Scientific Data* 7:225. **[S]** + https://doi.org/10.1038/s41597-020-0534-3 +- **[Cooper-2021]** Cooper, H. M., et al. (2021). COSMOS-UK: national soil moisture and + hydrometeorology data for environmental science research. *Earth System Science + Data* 13:1737–1757. **[V]** https://doi.org/10.5194/essd-13-1737-2021 +- **[Bogena-2022]** Bogena, H. R., et al. (2022). COSMOS-Europe: a European network of + cosmic-ray neutron soil moisture sensors. *Earth System Science Data* + 14:1125–1151. **[V]** https://doi.org/10.5194/essd-14-1125-2022 +- **[NEON-DP1]** NSF National Ecological Observatory Network. Soil water content and + water salinity (DP1.00094.001). **[S]** + https://data.neonscience.org/data-products/DP1.00094.001 +- **[CEOS-SM]** CEOS Land Product Validation Subgroup (2020). Soil Moisture Validation + Good Practices Protocol. **[org]** + https://lpvs.gsfc.nasa.gov/PDF/CEOS_SM_LPV_Protocol_V1_20201027_final.pdf +- **[NIDIS-2024]** NIDIS (2024). Soil Moisture Data Quality Guidance. **[org]** + https://www.drought.gov/sites/default/files/2024-12/Soil-Moisture-Data-Quality-Guidance-12-09-2024.pdf + +### ESS-DIVE and the Watershed Function SFA + +- **[ESS-DIVE]** Environmental Systems Science Data Infrastructure for a Virtual + Ecosystem (ESS-DIVE), LBNL/DOE. **[org]** https://ess-dive.lbl.gov · + https://docs.ess-dive.lbl.gov · https://www.re3data.org/repository/r3d100000019 +- **[Crystal-Ornelas-2022]** Crystal-Ornelas, R., Varadharajan, C., et al. (2022). + Enabling FAIR data in Earth and environmental science with community-centric + (meta)data reporting formats. *Scientific Data* 9:700. **[V]** + https://doi.org/10.1038/s41597-022-01606-w +- **[Velliquette-2021-CSV]** Velliquette, T., Welch, J., Crow, M., et al. (2021). CSV + File Structure Reporting Format. ESS-DIVE. **[V]** DOI 10.15485/1734841 · + https://github.com/ess-dive-community/essdive-csv-structure +- **[Goldman-2021]** Goldman, A. E., Ren, H., Torgeson, J., Zhou, H. (2021). Hydrologic + Monitoring Data and Metadata Reporting Format. ESS-DIVE. **[S]** + https://www.osti.gov/biblio/1822940 +- **[Damerow-2021]** Damerow, J. E., Varadharajan, C., et al. (2021). Sample + Identifiers and Metadata to Support Data Management and Reuse in Multidisciplinary + Ecosystem Sciences. *Data Science Journal* 20:11. **[V]** + https://doi.org/10.5334/dsj-2021-011 +- **[Crystal-Ornelas-2021]** Crystal-Ornelas, R., et al. (2021). A Guide to Using + GitHub for Developing and Versioning Data Standards and Reporting Formats. *Earth + and Space Science* 8:e2021EA001797. **[S]** https://doi.org/10.1029/2021EA001797 +- **[Hubbard-2018]** Hubbard, S. S., Williams, K. H., et al. (2018). The East River, + Colorado, Watershed: A Mountainous Community Testbed... *Vadose Zone Journal* + 17:180061. **[V]** https://doi.org/10.2136/vzj2018.03.0061 +- **[Varadharajan-2022]** Varadharajan, C., Hendrix, V. C., et al. (2022). The Colorado + East River Community Observatory Data Collection. *Hydrological Processes* + 36:e14243. **[S]** https://doi.org/10.1002/hyp.14243 + +### Information models, ontologies, vocabularies, units, FAIR + +- **[Wilkinson-2016]** Wilkinson, M. D., et al. (2016). The FAIR Guiding Principles for + scientific data management and stewardship. *Scientific Data* 3:160018. **[V]** + https://doi.org/10.1038/sdata.2016.18 +- **[Horsburgh-2016]** Horsburgh, J. S., et al. (2016). Observations Data Model 2 + (ODM2): A community information model for spatially discrete Earth observations. + *Environmental Modelling & Software* 79:55–74. **[S]** + https://doi.org/10.1016/j.envsoft.2016.01.010 · https://www.odm2.org/ +- **[CUAHSI-HIS]** CUAHSI Hydrologic Information System / WaterML / HydroShare. **[org]** + https://his.cuahsi.org/ · https://www.hydroshare.org/ +- **[OGC-SensorThings]** Open Geospatial Consortium. SensorThings API Part 1: Sensing + v1.1 (18-088); Observations & Measurements (ISO 19156). **[V]** + https://docs.ogc.org/is/18-088/18-088.html +- **[Janowicz-2019]** Janowicz, K., Haller, A., Cox, S. J. D., et al. (2019). The + modular SSN ontology: A joint W3C and OGC standard. *Semantic Web* 10(1). **[V]** + https://www.w3.org/TR/vocab-ssn/ +- **[Buttigieg-2013]** Buttigieg, P. L., Morrison, N., Smith, B., Mungall, C. J., Lewis, + S. E. (2013). The environment ontology: contextualising biological and biomedical + entities. *Journal of Biomedical Semantics* 4:43. **[V]** + https://doi.org/10.1186/2041-1480-4-43 +- **[AGROVOC]** FAO. AGROVOC: the linked data concept hub for food and agriculture. + **[org]** https://www.fao.org/agrovoc/ +- **[CF-Conventions]** NetCDF Climate and Forecast (CF) Metadata Conventions and CF + Standard Name Table. **[V]** https://cfconventions.org/ +- **[NERC-NVS]** NERC Vocabulary Server (BODC P01/P06/P07). **[org]** + https://vocab.nerc.ac.uk/ +- **[UCUM]** Schadow, G., McDonald, C. J. Unified Code for Units of Measure + (Regenstrief Institute). **[org]** https://ucum.org/ · https://ucum.nlm.nih.gov/ +- **[QUDT]** QUDT — Quantities, Units, Dimensions, and Types ontologies. **[org]** + https://qudt.org/ +- **[Rijgersberg-2013]** Rijgersberg, H., van Assem, M., Top, J. (2013). Ontology of + units of measure and related concepts. *Semantic Web* 4(1):3–13. **[V]** + https://www.semantic-web-journal.net/sites/default/files/swj177.pdf +- **[Wieczorek-2012]** Wieczorek, J., et al. (2012). Darwin Core: An Evolving + Community-Developed Biodiversity Data Standard. *PLOS ONE* 7(1):e29715. **[V]** + https://doi.org/10.1371/journal.pone.0029715 · https://dwc.tdwg.org/ + +## Data heterogeneity, provenance, and the cost of curation + +- **[Mars-Orbiter]** NASA Mars Climate Orbiter Mishap Investigation Board (1999). + Phase I Report. **[V]** https://science.nasa.gov/mission/mars-climate-orbiter/ +- **[Pint]** Grecco, H. Pint: a Python units library. **[org]** + https://github.com/hgrecco/pint +- **[PROJ]** PROJ — coordinate transformation software; EPSG registry. **[org]** + https://proj.org/ · https://epsg.org/ +- **[Zizka-2019]** Zizka, A., Silvestro, D., Andermann, T., et al. (2019). + CoordinateCleaner: Standardized cleaning of occurrence records from biological + collection databases. *Methods in Ecology and Evolution* 10(5):744–751. **[S]** + https://doi.org/10.1111/2041-210X.13152 +- **[ISO-8601]** ISO 8601-1:2019, Date and time representations. **[org]** + https://www.iso.org/iso-8601-date-and-time-format.html +- **[Wickham-2014]** Wickham, H. (2014). Tidy Data. *Journal of Statistical Software* + 59(10):1–23. **[V]** https://doi.org/10.18637/jss.v059.i10 +- **[Poggio-2021]** Poggio, L., et al. (2021). SoilGrids 2.0: producing soil + information for the globe with quantified spatial uncertainty. *SOIL* 7:217–240. + **[S]** https://doi.org/10.5194/soil-7-217-2021 +- **[Yue-2017]** Yue, K., et al. (2017). Experimental and observational studies find + contrasting responses of soil nutrients to climate change. *eLife* 6:e23255. **[S]** + https://doi.org/10.7554/eLife.23255 +- **[Heidorn-2008]** Heidorn, P. B. (2008). Shedding Light on the Dark Data in the Long + Tail of Science. *Library Trends* 57(2):280–299. **[S]** + https://muse.jhu.edu/article/262029 +- **[Blagoderov-2012]** Blagoderov, V., et al. (2012). No specimen left behind: + industrial-scale digitization of natural history collections. *ZooKeys* + 209:133–146. **[S]** https://doi.org/10.3897/zookeys.209.3178 +- **[Kandel-2012]** Kandel, S., Paepcke, A., Hellerstein, J., Heer, J. (2012). + Enterprise Data Analysis and Visualization: An Interview Study. *IEEE VAST*. **[S]** + http://vis.stanford.edu/papers/enterprise-analysis-interviews +- **[Lohr-2014]** Lohr, S. (2014). For Big-Data Scientists, 'Janitor Work' Is Key + Hurdle to Insights. *The New York Times*, Aug 18. **[S — claim provenance weak]** + https://www.nytimes.com/2014/08/18/technology/for-big-data-scientists-hurdle-to-insights-is-janitor-work.html +- **[CrowdFlower-2016]** CrowdFlower (2016). Data Science Report. **[S]** + https://visit.figure-eight.com/data-science-report.html +- **[Dodds-2020]** Dodds, L. (2020). Do data scientists spend 80% of their time + cleaning data? Turns out, no? **[org]** + https://blog.ldodds.com/2020/01/31/do-data-scientists-spend-80-of-their-time-cleaning-data-turns-out-no/ +- **[Fortier-2017]** Fortier, I., et al. (2017). Maelstrom Research guidelines for + rigorous retrospective data harmonization. *International Journal of Epidemiology* + 46(1):103–105. **[V]** https://doi.org/10.1093/ije/dyw075 +- **[OHDSI-OMOP]** Observational Health Data Sciences and Informatics. OMOP Common + Data Model. **[org]** https://www.ohdsi.org/data-standardization/ +- **[Schneider-2020]** Schneider, F. D., et al. (2020). Towards an ecological + trait-data standard / harmonizing heterogeneous trait data. *Ecological + Informatics* 60:101206. **[S]** + https://doi.org/10.1016/j.ecoinf.2020.101206 +- **[Dornelas-2018]** Dornelas, M., et al. (2018). BioTIME: A database of biodiversity + time series for the Anthropocene. *Global Ecology and Biogeography* 27(7):760–786. + **[S]** https://doi.org/10.1111/geb.12729 +- **[PROV-O]** Lebo, T., Sahoo, S., McGuinness, D. (eds.) (2013). PROV-O: The PROV + Ontology. W3C Recommendation. **[V]** https://www.w3.org/TR/prov-o/ +- **[RO-Crate-2022]** Soiland-Reyes, S., et al. (2022). Packaging research artefacts + with RO-Crate. *Data Science* 5(2):97–138. **[V]** + https://doi.org/10.3233/DS-210053 + +## Data cleaning, schema matching, entity resolution (classical) + +- **[Raman-2001]** Raman, V., Hellerstein, J. M. (2001). Potter's Wheel: An Interactive + Data Cleaning System. *VLDB*. **[S]** http://www.vldb.org/conf/2001/P381.pdf +- **[Kandel-2011]** Kandel, S., Paepcke, A., Hellerstein, J., Heer, J. (2011). Wrangler: + Interactive Visual Specification of Data Transformation Scripts. *ACM CHI*. **[S]** + https://idl.uw.edu/papers/wrangler +- **[OpenRefine]** OpenRefine (formerly Google Refine). **[org]** + https://openrefine.org/ +- **[Rahm-2001]** Rahm, E., Bernstein, P. A. (2001). A survey of approaches to + automatic schema matching. *The VLDB Journal* 10(4):334–350. **[V]** + https://doi.org/10.1007/s007780100057 +- **[Bernstein-2011]** Bernstein, P. A., Madhavan, J., Rahm, E. (2011). Generic Schema + Matching, Ten Years Later. *PVLDB* 4(11):695–701. **[S]** + https://vldb.org/pvldb/vol4/p695-bernstein_madhavan_rahm.pdf +- **[Ilyas-2019]** Ilyas, I. F., Chu, X. (2019). *Data Cleaning*. ACM Books, Vol. 28. + **[S]** https://doi.org/10.1145/3310205 +- **[Rekatsinas-2017]** Rekatsinas, T., Chu, X., Ilyas, I. F., Ré, C. (2017). HoloClean: + Holistic Data Repairs with Probabilistic Inference. *PVLDB* 10(11):1190–1201. **[V]** + https://doi.org/10.14778/3137628.3137631 +- **[Fellegi-1969]** Fellegi, I. P., Sunter, A. B. (1969). A Theory for Record Linkage. + *Journal of the American Statistical Association* 64(328):1183–1210. **[S]** + https://doi.org/10.1080/01621459.1969.10501049 +- **[Konda-2016]** Konda, P., Das, S., Suganthan G.C., P., Doan, A., et al. (2016). + Magellan: Toward Building Entity Matching Management Systems. *PVLDB* + 9(13):1581–1584. **[V]** https://doi.org/10.14778/2994509.2994535 +- **[Gulwani-2011]** Gulwani, S. (2011). Automating String Processing in Spreadsheets + Using Input-Output Examples (FlashFill). *POPL*. **[S]** + https://doi.org/10.1145/1926385.1926423 + +## LLM / foundation-model methods for data work + +- **[Li-2020]** Li, Y., Li, J., Suhara, Y., Doan, A., Tan, W.-C. (2020). Deep Entity + Matching with Pre-Trained Language Models (DITTO). *PVLDB* 13(12). **[S]** + https://arxiv.org/abs/2004.00584 +- **[Narayan-2022]** Narayan, A., Chami, I., Orr, L., Arora, S., Ré, C. (2022). Can + Foundation Models Wrangle Your Data? *PVLDB* 16(4):738–746. **[V]** + https://arxiv.org/abs/2205.09911 +- **[Li-2023]** Li, P., He, Y., et al. (2023). Table-GPT: Table-tuned GPT for Diverse + Table Tasks. *arXiv:2310.09263*. **[V]** https://arxiv.org/abs/2310.09263 +- **[Parciak-2024]** Parciak, M., et al. (2024). Schema Matching with Large Language + Models: An Experimental Study. *VLDB Workshops (TaDA)*. **[V]** + https://arxiv.org/abs/2407.11852 +- **[Qi-2024]** Qi, D., Miao, R., Wang, J. (2024). CleanAgent: Automating Data + Standardization with LLM-based Agents. *arXiv:2403.08291*. **[V]** + https://arxiv.org/abs/2403.08291 +- **[Wang-2025]** Wang, et al. (2025). Dataforge: Agentic Platform for Autonomous Data + Engineering. *arXiv:2511.06185*. **[V]** https://arxiv.org/abs/2511.06185 +- **[Zhang-2024]** Zhang, H., et al. (2024). Jellyfish: A Large Language Model for Data + Preprocessing. *EMNLP*. **[S]** https://arxiv.org/abs/2312.01678 +- **[KcMF-2024]** (2024). KcMF: A Knowledge-compliant Framework for Schema and Entity + Matching with Fine-tuning-free LLMs. *arXiv:2410.12480*. **[S]** + https://arxiv.org/abs/2410.12480 +- **[TabulaX-2024]** (2024). TabulaX: Leveraging LLMs for Multi-Class Table + Transformations. *arXiv:2411.17110*. **[S]** https://arxiv.org/abs/2411.17110 +- **[Prompt2DAG-2025]** (2025). Prompt2DAG: Modular Methodology for LLM-Based Data + Enrichment Pipeline Generation. *arXiv:2509.13487*. **[S]** + https://arxiv.org/abs/2509.13487 + +## Agentic harmonization, multi-agent systems, and agent evaluation + +- **[Santos-2025]** Santos, A., Pena, E., Lopez, R., Freire, J. (2025). Interactive Data + Harmonization with LLM Agents (Harmonia). *NOVAS'25 (co-located with SIGMOD)*. **[V]** + https://arxiv.org/abs/2502.07132 +- **[Cemri-2025]** Cemri, M., Pan, M., et al. (2025). Why Do Multi-Agent LLM Systems + Fail? (MAST). *NeurIPS Datasets & Benchmarks*. **[V]** + https://github.com/multi-agent-systems-failure-taxonomy/MAST +- **[Zhang-2025-DataSci]** Zhang, D., et al. (2025). DataSciBench: An LLM Agent + Benchmark for Data Science. *arXiv:2502.13897*. **[V]** + https://arxiv.org/abs/2502.13897 +- **[KramaBench-2025]** (2025). KramaBench: A Benchmark for AI Systems on + Data-to-Insight Pipelines over Data Lakes. *arXiv:2506.06541*. **[V]** + https://arxiv.org/abs/2506.06541 · https://kramabench.org/ +- **[Jing-2024]** Jing, L., et al. (2024). DSBench: How Far Are Data Science Agents from + Becoming Data Science Experts? *arXiv:2409.07703*. **[V]** + https://arxiv.org/abs/2409.07703 +- **[Hu-2024]** Hu, X., et al. (2024). InfiAgent-DABench: Evaluating Agents on Data + Analysis Tasks. *ICML*. **[S]** https://arxiv.org/abs/2401.05507 +- **[JudgeReliability-2026]** (2026). Judge Reliability Harness: Stress Testing the + Reliability of LLM Judges. *arXiv:2603.05399*. **[S — 2026 preprint, confirm]** + https://arxiv.org/abs/2603.05399 +- **[MOLE-2025]** Alyafeai, Z., Al-Shaibani, M., Ghanem, B. (2025). MOLE: Metadata + Extraction and Validation in Scientific Papers Using LLMs. *EMNLP Findings*. **[V]** + https://arxiv.org/abs/2505.19800 +- **[DCAT-LLM-2025]** (2025). Exploring LLM Capabilities in Extracting DCAT-Compatible + Metadata for Data Cataloging. *arXiv:2507.05282*. **[S]** + https://arxiv.org/abs/2507.05282 +- **[AgentSkills-2025]** Anthropic (2025). Agent Skills. **[org — cite primary + engineering blog / agentskills.io at submission]** https://www.anthropic.com/ + +## Benchmarks for cleaning, matching, and table tasks + +- **[Abedjan-2016]** Abedjan, Z., et al. (2016). Detecting Data Errors: Where are we and + what needs to be done? *PVLDB* 9(12):993–1004. **[S]** + https://doi.org/10.14778/2994509.2994518 +- **[CleanML-2021]** Li, P., Rao, X., Blase, J., Zhang, Y., Chu, X., Zhang, C. (2021). + CleanML: A Study for Evaluating the Impact of Data Cleaning on ML Classification + Tasks. *ICDE*. **[V]** https://arxiv.org/abs/1904.09483 +- **[Arocena-2015]** Arocena, P. C., Glavic, B., Mecca, G., Miller, R. J., Papotti, P., + Santoro, D. (2015). Messing Up with BART: Error Generation for Evaluating + Data-Cleaning Algorithms. *PVLDB* 9(2):36–47. **[S]** + http://www.vldb.org/pvldb/vol9/p36-arocena.pdf +- **[Mudgal-2018]** Mudgal, S., et al. (2018). Deep Learning for Entity Matching: A + Design Space Exploration (DeepMatcher). *SIGMOD*. **[S]** + https://doi.org/10.1145/3183713.3196926 +- **[Peeters-2024]** Peeters, R., Der, R. C., Bizer, C. (2024). WDC Products: A + Multi-Dimensional Entity Matching Benchmark. *EDBT*. **[S]** + http://webdatacommons.org/largescaleproductcorpus/v2/ +- **[Hulsebos-2019]** Hulsebos, M., et al. (2019). Sherlock: A Deep Learning Approach to + Semantic Data Type Detection. *KDD*. **[S]** https://doi.org/10.1145/3292500.3330993 +- **[SemTab]** Jiménez-Ruiz, E., Hassanzadeh, O., et al. (2019–). Semantic Web Challenge + on Tabular Data to Knowledge Graph Matching (SemTab). *CEUR-WS*. **[org]** + https://sem-tab-challenge.github.io/ +- **[OAEI]** Ontology Alignment Evaluation Initiative. **[org]** + http://oaei.ontologymatching.org/ diff --git a/docs/manuscript/research_notes.md b/docs/manuscript/research_notes.md new file mode 100644 index 0000000..6ed18e3 --- /dev/null +++ b/docs/manuscript/research_notes.md @@ -0,0 +1,150 @@ +# Research Notes — Literature Survey Provenance + +> Working notes from the deep literature research conducted for this manuscript +> (June 2026). This is **reference material**, not manuscript prose; it records what +> was found, how directly it relates, and how well it was verified, so that +> `01_introduction.md`, `02_background.md`, and `references.md` rest on a traceable +> basis. Verification: **[V]** fetched & confirmed; **[S]** snippet/secondary only; +> **[org]** official site. +> +> Two research streams were run: (A) the latest agentic-curation / LLM-for-data-work +> literature (2024–2026), and (B) environmental / Earth-science harmonization and +> general data-cleaning literature. Stream B fanned out over six sub-topics. + +## How the literatures map onto the manuscript + +A useful three-layer framing (any harmonizer reconciles all three): + +1. **Information models / schemas** — record structure incl. coordinates, depth, + timestamps: ODM2, OGC O&M / SensorThings, WaterML, Darwin Core, SoilML, and the + ESS-DIVE reporting formats (our concrete target). +2. **Variable vocabularies** — naming the observed property: CF Standard Names + (canonical units for soil-moisture variables), NERC P01, ENVO, AGROVOC. +3. **Unit systems** — normalizing values: UCUM, QUDT, OM, CF canonical units. + +FAIR [Wilkinson-2016] is the overarching motivation; CF Standard Names + UCUM/CF +canonical units are the most on-point for our specific target variables (volumetric / +gravimetric water content, water potential). + +## Most directly comparable / competing work (lead with these) + +1. **Harmonia** [Santos-2025] — *the* closest system: LLM-agent harmonization-to-target + -schema. Differentiate: ours is an **autonomous agent pair** (not interactive), + **environmental** (not clinical), and **rigorously evaluated** (grouped LOO CV + + blind prospective + cell-level output equivalence). **[V]** +2. **CleanAgent** [Qi-2024] — closest single-agent standardization analogue. **[V]** +3. **Dataforge** [Wang-2025] — closest on validation-loop architecture + CV-based + evaluation. **[V]** +4. **DataSciBench** [Zhang-2025-DataSci] & **KramaBench** [KramaBench-2025] — closest + evaluation-methodology peers; KramaBench covers environmental science and shows + best-system end-to-end accuracy ~55%, motivating our work. **[V]** +5. **MAST / "Why Do Multi-Agent LLM Systems Fail?"** [Cemri-2025] — the reference point + to contrast our two-agent error-propagation taxonomy and failure attribution. **[V]** + +The strongest *human-baseline / analogy* anchors: **ISMN** [Dorigo-2021] (real expert +harmonization of heterogeneous in-situ soil moisture; explicitly names our +heterogeneity axes — units, depth, time zones, sampling — and documents that ingest is +still "manually, mostly by email"); **Maelstrom/Fortier** [Fortier-2017] (gold-standard +manual retrospective harmonization process from epidemiology); **OMOP/OHDSI** +[OHDSI-OMOP] (common-data-model target schema). Strongest cost/labor anchors: +**Heidorn** [Heidorn-2008] (dark data / long tail) and **Blagoderov** [Blagoderov-2012] +(quantified curation labor). + +## Stream A — Agentic curation / LLM-for-data-work (2024–2026) + +- **Cleaning/wrangling/prep:** CleanAgent [Qi-2024] **[V]**, Dataforge [Wang-2025] + **[V]**, Jellyfish [Zhang-2024] **[S]**. +- **Schema/entity matching, ETL gen:** Harmonia [Santos-2025] **[V]**, KcMF + [KcMF-2024] **[S]**, TabulaX [TabulaX-2024] **[S]**, Prompt2DAG [Prompt2DAG-2025] + **[S]**, plus a large entity-matching line (Peeters & Bizer, AnyMatch, "Match, + Compare, or Select?") — context only. +- **Multi-agent / error propagation:** MAST [Cemri-2025] **[V]**; also PEAR (planner→ + executor robustness) and a 2026 failure-*attribution* benchmark — both directly + relevant to attributing curator-vs-harmonizer error (snippet-only; verify before + citing). +- **Code-gen / PBE lineage:** FlashFill [Gulwani-2011], Table-GPT [Li-2023] **[V]**, + "Can Foundation Models Wrangle Your Data?" [Narayan-2022] **[V]** — the key prior + work showing a single FM matches task-specific cleaning systems. +- **Evaluation methodology:** DataSciBench [Zhang-2025-DataSci] **[V]** (uncertain + ground truth; Task–Function–Code scoring of execution outcomes — informs our + output-equivalence design), KramaBench [KramaBench-2025] **[V]**, DSBench [Jing-2024] + **[V]**, InfiAgent-DABench [Hu-2024] **[S]**. LLM-as-judge reliability papers + (snippet-only) justify our preference for automated cell-level metrics over a model + judge. +- **Skills / structured prompting:** Anthropic Agent Skills [AgentSkills-2025] — cite + the primary Anthropic engineering blog / agentskills.io at submission (current + coverage is secondary tech-press). +- **Scientific curation with LLMs:** MOLE [MOLE-2025] **[V]** (extract-then-validate- + against-GT, mirrors our conformance checks), DCAT metadata extraction [DCAT-LLM-2025] + **[S]** (few-shot > zero-shot — supports exemplar value). + +**Caveat:** several 2026 arXiv IDs (26xx.xxxxx) are consistent with the current date +but were captured from search snippets only; confirm author/title/venue before final +citation. The single most directly comparable paper is **Harmonia** [Santos-2025]. + +## Stream B — Environmental harmonization & data cleaning + +### B1 Soil-moisture networks (how harmonization is done today) +ISMN [Dorigo-2011; Dorigo-2021; Dorigo-2013], AmeriFlux BASE [Chu-2023], FLUXNET/ +ONEFlux [Pastorello-2020], COSMOS-UK [Cooper-2021], COSMOS-Europe [Bogena-2022], NEON +[NEON-DP1]. Consistent theme: harmonization relies on agreed reporting formats + +substantial manual/semi-automated ingest, with automation mostly confined to +downstream QC — the gap our agent pair targets. ISMN converts all SM to volumetric +m³/m³, resamples to hourly UTC, flags rather than deletes; notably does **not** +harmonize depth. + +### B2 ESS-DIVE & WFSFA (our data source and target standards) +Repository [ESS-DIVE]; flagship reporting-formats paper [Crystal-Ornelas-2022] (11 +formats; reviewed 112 standards; 247 reps / 128 institutions); CSV format +[Velliquette-2021-CSV], FLMD, hydrologic monitoring [Goldman-2021], location metadata, +sample IDs/IGSN [Damerow-2021]; standards-development process [Crystal-Ornelas-2021]. +Ground truth from the East River testbed [Hubbard-2018; Varadharajan-2022]. **Key fact +for motivation: ESS-DIVE has no dedicated soil-moisture reporting format.** + +### B3 Ontologies / FAIR / units +FAIR [Wilkinson-2016]; ODM2 [Horsburgh-2016]; CUAHSI/WaterML [CUAHSI-HIS]; OGC O&M/ +SensorThings [OGC-SensorThings]; SOSA/SSN [Janowicz-2019]; ENVO [Buttigieg-2013] (note: +C. J. Mungall — the project lead — is a co-author); AGROVOC [AGROVOC]; CF Conventions +[CF-Conventions]; NERC NVS [NERC-NVS]; UCUM [UCUM]; QUDT [QUDT]; OM [Rijgersberg-2013]; +Darwin Core [Wieczorek-2012]. + +### B4 Classical data cleaning / matching (CS) +Potter's Wheel [Raman-2001], Wrangler/Trifacta [Kandel-2011], OpenRefine [OpenRefine]; +schema-matching surveys [Rahm-2001; Bernstein-2011]; data quality [Ilyas-2019]; +HoloClean [Rekatsinas-2017]; record linkage [Fellegi-1969]; Magellan [Konda-2016]. +LLM era: DITTO [Li-2020], "Can FMs Wrangle Your Data?" [Narayan-2022], Table-GPT +[Li-2023], LLM schema matching [Parciak-2024] (context balance matters). +**"80% of time on data prep" provenance:** traces to a 2014 NYT piece [Lohr-2014] +attributed only to "interviews and expert estimates" — no rigorous primary source; +cite [CrowdFlower-2016] (small n) and especially [Kandel-2012] (peer-reviewed +qualitative) instead, and note the critique [Dodds-2020]. + +### B5 Provenance, reproducibility, cost/labor +PROV-O [PROV-O], RO-Crate [RO-Crate-2022] (+ Workflow Run Crate 2024); reproducible- +pipeline provenance (Ursprung, PVLDB 2020). Cost/labor: long-tail dark data +[Heidorn-2008], digitization labor [Blagoderov-2012]. Cross-domain harmonization +analogies: Maelstrom [Fortier-2017] (retrospective vs prospective — mirrors our +Phase A/B split), OMOP/OHDSI [OHDSI-OMOP], ecology (trait data [Schneider-2020], +BioTIME [Dornelas-2018]). + +### B6 Heterogeneity hazards & benchmarks +Unit mismatch: Mars Climate Orbiter [Mars-Orbiter]; UCUM/Pint [UCUM; Pint]. CRS: +PROJ/EPSG [PROJ], CoordinateCleaner [Zizka-2019] (~3.6% of 91M GBIF records flagged). +Time: ISO 8601 [ISO-8601]. Reshaping: Tidy Data [Wickham-2014]. Depth: SoilGrids/ +GlobalSoilMap intervals [Poggio-2021]. Observational-vs-manipulated confound: +[Yue-2017] (experimental and observational responses can have *opposite* sign). +Benchmarks: error detection [Abedjan-2016], CleanML [CleanML-2021], BART [Arocena-2015], +DeepMatcher [Mudgal-2018], WDC Products [Peeters-2024], Sherlock [Hulsebos-2019], +SemTab [SemTab]; LLM-agent data benchmarks DSBench [Jing-2024], InfiAgent-DABench +[Hu-2024], KramaBench [KramaBench-2025]; expert-gold-standard / reference-alignment +evaluation paradigm [OAEI]. + +## Outstanding verification TODOs before submission + +- Confirm DOIs / author order for snippet-only **[S]** entries, especially older works + ([Fellegi-1969], [Dorigo-2011], [Dorigo-2013]) and all 2026 arXiv preprints + ([JudgeReliability-2026] and the failure-attribution benchmark). +- Replace the secondary Agent Skills citation with Anthropic's primary source. +- Verify ESS-DIVE reporting-format dataset author lists/years on the OSTI DOI landing + pages (several have v1/v2 releases). +- ODM2 [Horsburgh-2016] full text was publisher-blocked; DOI corroborated — confirm. From 9b86fa3161aef4fe57dbd3e7fddd98f70fdcaa1e Mon Sep 17 00:00:00 2001 From: Claude Date: Tue, 30 Jun 2026 23:18:58 +0000 Subject: [PATCH 2/4] Add Figure 1 workflow schematic and HTML slide deck - figures/figure1_workflow.mmd: Mermaid source for the agent-pair workflow and evaluation-design schematic (Figure 1) - figures/figure1_workflow.html: self-contained in-browser SVG renderer - slides/index.html: reveal.js deck (motivation, agent pair, two-phase / three-mode design, output-equivalence metrics, related-work positioning, results placeholders, limitations) with Figure 1 rendered live - update figures_and_tables.md and README.md to index the new assets All three Mermaid diagrams validated with mermaid@11 parse. Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_01FD9C8gFuJGBBYqemVFtwTx --- docs/manuscript/README.md | 2 + docs/manuscript/figures/figure1_workflow.html | 93 ++++ docs/manuscript/figures/figure1_workflow.mmd | 49 ++ docs/manuscript/figures_and_tables.md | 9 +- docs/manuscript/slides/index.html | 421 ++++++++++++++++++ 5 files changed, 573 insertions(+), 1 deletion(-) create mode 100644 docs/manuscript/figures/figure1_workflow.html create mode 100644 docs/manuscript/figures/figure1_workflow.mmd create mode 100644 docs/manuscript/slides/index.html diff --git a/docs/manuscript/README.md b/docs/manuscript/README.md index 38ad6aa..37e2083 100644 --- a/docs/manuscript/README.md +++ b/docs/manuscript/README.md @@ -23,6 +23,8 @@ Working drafts for the paper describing this study: | `figures_and_tables.md` | Display-item manifest | Mapped to analysis code | | `references.md` | Bibliography | Drafted (research-backed) | | `research_notes.md` | Raw literature research (provenance/verification) | Reference material | +| `figures/` | Figure 1 schematic (Mermaid source + HTML render) | Drafted | +| `slides/index.html` | HTML slide deck (reveal.js) | Drafted | Read in numeric order for a continuous draft. The intended journal/venue is TBD (candidates: *Scientific Data*, *Environmental Modelling & Software*, *PVLDB / diff --git a/docs/manuscript/figures/figure1_workflow.html b/docs/manuscript/figures/figure1_workflow.html new file mode 100644 index 0000000..02dc8e8 --- /dev/null +++ b/docs/manuscript/figures/figure1_workflow.html @@ -0,0 +1,93 @@ + + + + + +Figure 1 — Agent-pair workflow & evaluation design + + + +
+

Figure 1. Agent-pair workflow and evaluation design.

+

A curator (Skill 1) turns a dataset identifier into a structured bundle; a harmonizer + (Skill 2) turns that bundle into executable transform code and a documented change-mapping. The + agent's code and the expert's code are each executed on the same raw files and the resulting + tables are compared cell-by-cell (primary endpoint). Three evaluation modes localize error, and + the Mode 2 − Mode 3 gap quantifies how much curator error propagates downstream.

+
+ 
+%%{init: {"theme":"base","themeVariables":{
+  "fontFamily":"ui-sans-serif, system-ui, Inter, Segoe UI, Roboto, sans-serif",
+  "fontSize":"14px","lineColor":"#8a8a8a","primaryBorderColor":"#c9bfa6",
+  "clusterBkg":"#ffffff","clusterBorder":"#e0d8c4"}}}%%
+flowchart TB
+    classDef agent fill:#0e7c7b,stroke:#0a5a59,color:#ffffff,stroke-width:1.5px;
+    classDef artifact fill:#f3efe6,stroke:#c9bfa6,color:#2b2b2b;
+    classDef gold fill:#c9962f,stroke:#9c7421,color:#ffffff;
+    classDef score fill:#3b5b8c,stroke:#28406a,color:#ffffff;
+    classDef io fill:#ffffff,stroke:#8a8a8a,color:#2b2b2b,stroke-dasharray:3 3;
+
+    ID["ESS-DIVE dataset identifier
(DOI / package ID)"]:::io
+    EX["19 expert exemplars
(in-context references)"]:::gold
+
+    subgraph PIPE["Agent pair under evaluation"]
+        direction TB
+        C["Curator — Skill 1
retrieve · classify files · resolve location
infer time series · detect manipulation · pick exemplar"]:::agent
+        B["Curator bundle
decision · payload/location files · coords · interval · qc_flag"]:::artifact
+        H["Harmonizer — Skill 2
map variables · convert units · generate code + change-mapping"]:::agent
+        OUT["Python transform code
+ change-mapping JSON"]:::artifact
+        C --> B --> H --> OUT
+    end
+
+    RAW[("Raw package files")]:::io
+    GC["Expert transform code"]:::gold
+    AGT["Agent harmonized table"]:::artifact
+    GLD["Expert harmonized table
(gold standard)"]:::gold
+    CMP{{"PRIMARY ENDPOINT
output-data equivalence
cell-by-cell on the natural key"}}:::score
+
+    ID --> C
+    EX -. exemplars .-> C
+    EX -. exemplars .-> H
+    OUT -- execute --> AGT
+    RAW --> AGT
+    RAW --> GLD
+    GC --> GLD
+    AGT --> CMP
+    GLD --> CMP
+
+    subgraph MODES["Evaluation modes (error attribution)"]
+        direction TB
+        M1["Mode 1 · Skill-1 isolated — bundle vs expert labels"]:::score
+        M2["Mode 2 · Skill-2 oracle — gold bundle into harmonizer"]:::score
+        M3["Mode 3 · end-to-end — full pipeline, errors propagate"]:::score
+        GAP["Error-propagation gap = Mode 2 − Mode 3"]:::score
+        M2 --- GAP
+        M3 --- GAP
+    end
+
+    B -. scored .-> M1
+    OUT -. scored .-> M3
+
+
+
Source: figure1_workflow.mmd. Renders to SVG in-browser via Mermaid + (loaded from CDN). To export: open in a browser, right-click the diagram → save, or paste the + .mmd source into mermaid.live.
+
+ + + diff --git a/docs/manuscript/figures/figure1_workflow.mmd b/docs/manuscript/figures/figure1_workflow.mmd new file mode 100644 index 0000000..56f6d85 --- /dev/null +++ b/docs/manuscript/figures/figure1_workflow.mmd @@ -0,0 +1,49 @@ +%% Figure 1. Agent-pair workflow and evaluation design. +%% Render in browser via figure1_workflow.html, or paste into https://mermaid.live +flowchart TB + classDef agent fill:#0e7c7b,stroke:#0a5a59,color:#ffffff,stroke-width:1.5px; + classDef artifact fill:#f3efe6,stroke:#c9bfa6,color:#2b2b2b; + classDef gold fill:#c9962f,stroke:#9c7421,color:#ffffff; + classDef score fill:#3b5b8c,stroke:#28406a,color:#ffffff; + classDef io fill:#ffffff,stroke:#8a8a8a,color:#2b2b2b,stroke-dasharray:3 3; + + ID["ESS-DIVE dataset identifier
(DOI / package ID)"]:::io + EX["19 expert exemplars
(in-context references)"]:::gold + + subgraph PIPE["Agent pair under evaluation"] + direction TB + C["Curator — Skill 1
retrieve · classify files · resolve location
infer time series · detect manipulation · pick exemplar"]:::agent + B["Curator bundle
decision · payload/location files · coords · interval · qc_flag"]:::artifact + H["Harmonizer — Skill 2
map variables · convert units · generate code + change-mapping"]:::agent + OUT["Python transform code
+ change-mapping JSON"]:::artifact + C --> B --> H --> OUT + end + + RAW[("Raw package files")]:::io + GC["Expert transform code"]:::gold + AGT["Agent harmonized table"]:::artifact + GLD["Expert harmonized table
(gold standard)"]:::gold + CMP{{"PRIMARY ENDPOINT
output-data equivalence
cell-by-cell on the natural key"}}:::score + + ID --> C + EX -. exemplars .-> C + EX -. exemplars .-> H + OUT -- execute --> AGT + RAW --> AGT + RAW --> GLD + GC --> GLD + AGT --> CMP + GLD --> CMP + + subgraph MODES["Evaluation modes (error attribution)"] + direction TB + M1["Mode 1 · Skill-1 isolated — bundle vs expert labels"]:::score + M2["Mode 2 · Skill-2 oracle — gold bundle into harmonizer"]:::score + M3["Mode 3 · end-to-end — full pipeline, errors propagate"]:::score + GAP["Error-propagation gap = Mode 2 − Mode 3"]:::score + M2 --- GAP + M3 --- GAP + end + + B -. scored .-> M1 + OUT -. scored .-> M3 diff --git a/docs/manuscript/figures_and_tables.md b/docs/manuscript/figures_and_tables.md index 4bcd48b..8d2fd94 100644 --- a/docs/manuscript/figures_and_tables.md +++ b/docs/manuscript/figures_and_tables.md @@ -34,4 +34,11 @@ | # | Title | Notes | |---|---|---| -| 0 | Workflow + evaluation design schematic | Curator → bundle → Harmonizer → code+mapping → executed output vs. expert; three modes; two phases. Good candidate for **Figure 1 (main)**. Draft as a diagram (e.g., Mermaid/Figma) before results. | +| 1 | Workflow + evaluation design schematic | **Drafted** — `figures/figure1_workflow.mmd` (Mermaid source) + `figures/figure1_workflow.html` (in-browser SVG render). Curator → bundle → Harmonizer → code+mapping → executed output vs. expert; three modes; gap. This is the main **Figure 1**. | + +## Presentation + +- `slides/index.html` — self-contained reveal.js deck summarizing motivation, + the agent pair, design, metrics, positioning, and results placeholders. Renders + Figure 1 live via Mermaid. Open in a browser; press Esc for overview, + ? for shortcuts. diff --git a/docs/manuscript/slides/index.html b/docs/manuscript/slides/index.html new file mode 100644 index 0000000..78c965a --- /dev/null +++ b/docs/manuscript/slides/index.html @@ -0,0 +1,421 @@ + + + + + +Evaluating an LLM Agent Pair for Environmental Data Harmonization + + + + + +
+ + +
+
+

Evaluation framework · pre-results

+

Evaluating an LLM Agent Pair for
Environmental Data Harmonization

+
+

+ A curator + harmonizer + agent pair, benchmarked against an expert on DOE ESS-DIVE soil-moisture data.

+

Watershed Function SFA · Berkeley Lab  ·  authors TBD

+

SCAFFOLD — results pending

+

Press → / Space to advance · Esc for overview

+
+
+ + +
+

Motivation

+

Harmonization is the bottleneck in data reuse

+
+
+
+
    +
  • The same quantity arrives in incompatible forms — and reconciling it is expert, manual work.
    • +
  • The field's canonical effort (ISMN) still ingests provider data "manually, mostly by email." [Dorigo 2021]
    • +
  • Practitioners report 50–80% of time on data preparation. [Lohr 2014; CrowdFlower 2016 — provenance weak]
    • +
  • The heterogeneous "long tail" of small datasets is the costliest to integrate and the likeliest to be lost. [Heidorn 2008]
    • +
+
+
+

Why it matters here

+

ESS-DIVE hosts diverse Watershed Function SFA datasets but has no dedicated soil-moisture reporting format — every dataset is harmonized by hand into a project schema.

+

If an agent pair can draft that work reliably, it unlocks scalable, FAIR-aligned reuse.

+
+
+
+ + +
+

The task

+

One quantity, many encodings

+
+
+
+

Soil moisture must be reconciled across:

+
    +
  • Units — VWC % vs m³ m⁻³; GWC g g⁻¹; potential kPa / MPa / bar
    • +
  • Coordinates — UTM vs lat/lon, differing datums
    • +
  • Depth conventions; time zones → UTC
    • +
  • Wide ↔ long ("tidy") table layouts [Wickham 2014]
    • +
  • Observational vs manipulated (warming/irrigation) — can flip a result's sign [Yue 2017]
    • +
+
+
+

Target schema — 9 columns, fixed units:

+
+datetime_UTC ISO-8601 UTC
+site_id string
+depth_m m
+replicate int / NaN
+is_timeseries bool
+interval_min float / NaN
+volumetric_water… m³ m⁻³
+gravimetric_water… g g⁻¹
+water_potential_kPa kPa (neg.) +
+

qc_flag vocab: d1 (depth approx.) · g1 (coords from registry) · g2 (coords unresolved)

+
+
+
+ + +
+

The system under test

+

A curator → harmonizer agent pair

+
+
+
+

Skill 1  Curator

+

From a DOI/package ID: retrieves & classifies files, resolves location, infers time-series vs discrete, detects manipulations, picks the nearest exemplar.

+

→ a structured bundle with an INCLUDE / EXCLUDE / FLAG decision.

+
+
+

Skill 2  Harmonizer

+

From the bundle: maps variables, converts units, and emits executable Python plus a documented change-mapping JSON.

+

→ auditable code & provenance, not an opaque table.

+
+
+

Implemented as Anthropic Agent Skills — versioned instruction+tool bundles; the 19 expert harmonizations act as retrievable in-context exemplars. [Agent Skills 2025]

+
+ + +
+

Figure 1

+

Workflow & evaluation design

+
+ 
+%%{init: {"theme":"base","themeVariables":{"fontFamily":"ui-sans-serif, system-ui, Inter, sans-serif","fontSize":"15px","lineColor":"#8a8a8a","clusterBkg":"#ffffff","clusterBorder":"#e0d8c4"}}}%%
+flowchart LR
+    classDef agent fill:#0e7c7b,stroke:#0a5a59,color:#fff,stroke-width:1.5px;
+    classDef artifact fill:#f3efe6,stroke:#c9bfa6,color:#2b2b2b;
+    classDef gold fill:#c9962f,stroke:#9c7421,color:#fff;
+    classDef score fill:#3b5b8c,stroke:#28406a,color:#fff;
+    classDef io fill:#fff,stroke:#8a8a8a,color:#2b2b2b,stroke-dasharray:3 3;
+    ID["dataset identifier"]:::io
+    EX["19 expert exemplars"]:::gold
+    subgraph PIPE["agent pair"]
+      direction LR
+      C["Curator
Skill 1"]:::agent --> B["bundle"]:::artifact --> H["Harmonizer
Skill 2"]:::agent --> OUT["code +
change-mapping"]:::artifact
+    end
+    RAW[("raw files")]:::io
+    AGT["agent table"]:::artifact
+    GLD["expert table
(gold)"]:::gold
+    CMP{{"output equivalence
cell-by-cell"}}:::score
+    ID --> C
+    EX -. exemplars .-> C
+    EX -. exemplars .-> H
+    OUT -- execute --> AGT
+    RAW --> AGT
+    RAW --> GLD
+    AGT --> CMP
+    GLD --> CMP
+
+

Modes: Skill-1 isolated · Skill-2 oracle (gold bundle) · end-to-end. Gap = oracle − end-to-end.

+
+ + +
+

Evaluation design

+

Two phases · three modes · one primary endpoint

+
+

Honest generalization, decomposable error, output-grounded scoring.

+
+ + +
+

Phases

+

Retrospective, then prospective

+
+
+
+

Phase A · cross-validation

+

Grouped leave-one-out over 19 expert datasets. Whole source/instrument clusters held out together, so a near-duplicate sibling can't leak into the exemplar pool.

+

Also validates automated metrics against expert judgment.

+
+
+

Phase B · prospective blind

+

All 19 as exemplars; agent harmonizes new ESS-DIVE datasets. Expert harmonizes the same data blind → reference.

+

Estimates real deployment performance.

+
+
+

A similarity covariate (distance to nearest exemplar) turns leakage risk into a measured generalization curve.

+
+ + +
+

Modes

+

Where does the error come from?

+
+
+

1 · Skill-1 isolated

Curator bundle scored field-by-field vs expert labels.

+

2 · Skill-2 oracle

Harmonizer fed the gold bundle — transform quality, independent of curator error.

+

3 · End-to-end

Full pipeline; errors propagate.

+
+
+

Error-propagation gap = Mode 2 − Mode 3 isolates how much curator error degrades the final output.

+
+

Failures attributed as:

+
    +
  • Skill-1 error propagated
    • +
  • Skill-2 error (given correct input)
    • +
  • interface inconsistency
    • +
  • genuinely ambiguous (expert unsure too)
    • +
+
+
+
+ + +
+

Metrics

+

Output-data equivalence is the primary endpoint

+
+
+
+

Execute both the agent's and expert's code on the same raw data; compare tables cell-by-cell on the natural key. Two correct programs score the same regardless of style — and we avoid unreliable LLM-as-judge scoring. [Judge reliability 2026]

+

Harmonizer composite weighting:

+
output equivalence0.50
+
schema conformance0.15
+
semantic mapping F10.12
+
code executability0.10
+
documentation0.08
+
ontology validity0.05
+
+
+

Curator metrics

+

Decision P/R/F1 · file-set F1 · time-series accuracy & interval error · location-source & qc_flag accuracy · manipulation detection · exemplar match · deferral calibration.

+

Statistics

+

Nested data → cluster bootstrap CIs & mixed-effects models; effect sizes over p-values (N≈19); variance & pass@k across stochastic repeats.

+
+
+
+ + +
+

Controls

+

What the agent pair is measured against

+
+
+

Single-call LLM

No skill structure or tool loop — isolates the value of agentic decomposition.

+

No-exemplars

Agent with the 19 references removed — isolates the value of the exemplars.

+

Naïve heuristic

String-matching floor.

+
+

Reproducibility pinned per run: model + version, both skill versions, sampling params, seeds.

+
+ + +
+

Positioning

+

How this differs from prior work

+
+ + + + + + + +
WorkWhat it doesWe add
Harmonia Santos 2025LLM-agent harmonization to a schema; interactive; clinicalautonomous pair · environmental · CV + blind eval
CleanAgent Qi 2024single-agent data standardizationtwo-stage curate→harmonize, scored separately
MAST Cemri 2025taxonomy of multi-agent failuresoperationalized cross-agent error attribution
KramaBench 2025data pipelines benchmark (best ~55% E2E)real expert ground truth + output equivalence
"FMs wrangle data?" Narayan 2022FM beats task-specific cleanerscomposed, end-to-end, executed against expert
+

Lineage: Wrangler/Trifacta · HoloClean · FlashFill · schema-matching surveys (Rahm 2001). Domain anchors: ISMN, ESS-DIVE reporting formats, Maelstrom, OMOP.

+
+ + +
+

Deliverables

+

What the results section will report

+
+
+
+ + + + + + + + +
ResultValue
End-to-end output equivalence[[ • ]]
Skill-2 (oracle) equivalence[[ • ]]
Error-propagation gap[[ • ]]
Curator decision F1[[ • ]]
vs single-call / no-exemplar[[ • ]]
Similarity threshold[[ • ]]
+
+
+

Plus:

+
    +
  • error-attribution taxonomy (stacked / Sankey)
    • +
  • similarity → performance regression
    • +
  • automated-vs-expert metric agreement
    • +
  • Phase B agent-vs-expert on novel data
    • +
+

Every number traces to experiments/ → results/; placeholders are filled by the analysis pipeline, never by hand.

+
+
+
+ + +
+

Caveats

+

Limitations we hold ourselves to

+
+
+
    +
  • N ≈ 19 gold datasets — uncertain estimates; effect sizes, not p-values.
    • +
  • Single domain — WFSFA soil moisture on one repository.
    • +
  • One expert as gold — some cases genuinely ambiguous.
    • +
+
    +
  • Model/skill drift — pinned versions; results are snapshots.
    • +
  • Stochasticity — report variance & pass@k.
    • +
  • Automated metrics — credible only as far as Phase A validation shows.
    • +
+
+

Intended operating point: agent drafts, expert reviews, with automatic deferral above the similarity threshold and on low-confidence curator decisions.

+
+ + +
+

Takeaways

+

A rigorous yardstick for agentic harmonization

+
+
    +
  • Output-equivalence scoring — measures what the code does, not how it reads.
    • +
  • Decomposed across two agents — error is attributable, not opaque.
    • +
  • Honest generalization — grouped CV + similarity covariate + blind prospective test.
    • +
+

First such evaluation in the environmental-data-harmonization setting. Results to follow.

+
+ + +
+

Thank you

+

Evaluating an LLM Agent Pair for
Environmental Data Harmonization

+
+

Manuscript drafts: docs/manuscript/ · Figure source: figures/figure1_workflow.mmd

+

Framework: config/ · src/ · skills/ · gold standard: data/gold/

+

Key references — Dorigo 2021 (ISMN) · Crystal-Ornelas 2022 (ESS-DIVE formats) · Santos 2025 (Harmonia) · Cemri 2025 (MAST) · Narayan 2022 (FMs wrangle data) · Wickham 2014 (tidy data) · Wilkinson 2016 (FAIR). Full list in references.md.

+
+ +
+ +
+ LLM Agent Pair · Environmental Data Harmonization + Watershed Function SFA · pre-results scaffold +
+ + + + + From 80dec0d73aa59fbf7d7fa2c0b54a2ee3285dc7eb Mon Sep 17 00:00:00 2001 From: Claude Date: Tue, 30 Jun 2026 23:20:51 +0000 Subject: [PATCH 3/4] Add static SVG/PNG renders of Figure 1 Render figure1_workflow.mmd to a dependency-free SVG and a 2x PNG (via headless Chromium) so the schematic is viewable without a network/CDN. Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_01FD9C8gFuJGBBYqemVFtwTx --- docs/manuscript/figures/figure1_workflow.png | Bin 0 -> 340947 bytes docs/manuscript/figures/figure1_workflow.svg | 1 + docs/manuscript/figures_and_tables.md | 2 +- 3 files changed, 2 insertions(+), 1 deletion(-) create mode 100644 docs/manuscript/figures/figure1_workflow.png create mode 100644 docs/manuscript/figures/figure1_workflow.svg diff --git a/docs/manuscript/figures/figure1_workflow.png b/docs/manuscript/figures/figure1_workflow.png new file mode 100644 index 0000000000000000000000000000000000000000..e1986002c6c4929b9da3ddb8335269fc44bd517c GIT binary patch literal 340947 zcmcG$WmKEZ*Dp-HMM_)T3xyUbQnYw+cPko-G`I(f1lnRPTHIZNg;3l>ae{jxxD(tV z*a_`@|DWfq^SmF!VLic;lN8hZl)QO~=cQqux_b=(%dbg@OVG7Nn`jKQIx5N7y8dvLVd|D2 zH_YSGOICA@3!GJnB@^?u(>b}v!aakA0MfV#}4%aEQ_Fa5VNj{c7??ZJ%Z2%DM0?OJiA?24-FV zDT?%aH|zYj7=NYBxb<%Z^H=dLJ@ z)k-DOF4(|nckUOg5zXw>HI$`f%9St5a-w$(VV(#2gY1Fmwq%z_NmwY87eA!*WV zx+TWHE>C7@YL=F|qOLKo!wmoRfTpLrJJ`UaLT}dR`m&Q@Vq(JQ80bS>vsBhE;J<& zPPV5hYRnMPO#LF~HcQXe-^sHOaZ(4QUsBclXP*1qGm&l*}|AbHG0)3LxK(H zDs-*fmm;qM6;XBEw<8n+oI9@$Rjv;k&$lt1R5@aVPmzUvn>Z*cY zZu#SoNV7$_(r>XjbvZoNm6mRgQV}7mF?$=RD6pEtCF&$+VzLnKbFtXTF!bfeIt_uc(m!9PV)OYCO>d&|9lp$PBT?=i& zq0*+9Y0Kl3uR0l8_|1m)BE$5SBIhYfxqy9?!XDgC-$zLE1HLMBBQe(J`I>@Sr9Y(F{yO&$3w^%j$4&{TbA_ts6Jq#88`6CZ)S7S& zW%jNO;FH86?a}h-h&zJnT~=X_U(5e}Gyoez?zk^sym+uq)e5g9KVs(-(i@=C@u zzb0Lpq`etO%6G$a9Pu4AJ_diDyZM+4c=3lZo6JVW#$5Ln+D=bTk@j%Ie+I;VVobkb zC1cKgj|TgOqp^q=GI#&>eSY-)X4?Oh|C7rU`EOx1fz6;=RbQXwKjR!sBmQj8jQeUz z)r?2?&I!(c`ljC9JaaRhysT_zXQyhu=08%T$xreR%YW{Yg=&kQEk*s?E-CeYFA6^Y zDgwZb{xj>pd+=|=zrra#M2Nbd!oEIZ&_Wxk$M8?LqR{_?=pyY2qZ0i3Yi_c_Iz~R{ z--)oALdE~b)pL+pR!lA|nEWF$jWg-~@1e;1^E`g)`C*2{vTGV58lV{uB_{zZAFZj7nRlsdkEp}Cc1|qbC%uwY_zW9CYpxDD61K6&Di4vna)`Z_k_VNG`|>yMV|d4~%9 zHN`W3{|SLNnmElLjNzrP9t-8tQoTtl+dKlx2#ar&y{5$#78g0s0G~61#drEyNRQ{v zrpN9Gp78rM7j?6DYOA!;(i8!V&NF)3D8dC^5U=0aAI^?(-6_O<=u7&`80!ZUp`Nq+ zWCN1`xabh>8n_zO(l{n|r#Yx6-#K-FY+Fg5hrU2-XhXiR4uzO-V;3m6y>jaY@lO3 zuDOHiY#K<5wtM-~kreFeT$FV!xa(OWY*+Z3BjdJ2Q+)iFY|jm!Xx%i{ez z4-|yeRdrr7Vl!|v`jy$tiUqoK6mahVcqI{HhRfS=Zn51P>|Oem+~sY9kJtJ7tVK)Q z4`?9+%g~8h5&PdAb4kk;+rO)`GL^GQ8cWP4T8rX1X4;3$VW5+MRGa($W5xqaenZBZ za;q6CSfpf7)C_1nH<>=AaeTH%ENRyJ_qf2E)tmfqGqy>uq?ZdvfP4i-0R5#RU)t&Z z?_8Mhg~($@JG6$bd$@|)X`**ke!JQ*(_eDEp%-JP0rC)a2QC5gTBiLsO#2OXXS~*o z3z-~C#7nCQcmgSioze+q zbqa4%3xmOW;~&HsSbhgoF1)WoO;8hq#Esu?hj*eFN z1puB^E-RlGo>x6KycDzS1VlC43518`*NbA6YUCOf;as>v3xhj$#bQ`lQ@Cu$XHKS~e(jGUdb z0SW4%3cXT4GoP4WKCyHzFmv8J9B}@6WbdU`2?2f#g4x=8IJx!IbdVP6CVvn?6$-VD z5d1y(mOuBdW?0-EtK#Bn)ym2Ur8u^Moh=;Lx|i=Jt^CIo1%AnNWfWuBx%5+~sA z5NkJO{AQDxlDXquLQGX_W~ny9@`-sEPwW>?@{7Lxp<$;q!j-SLhqL3iu=sf3vs7p* z%{&*)4_NI*4J#X2;o#|m*|(-cKlmShZ%5*`zo%7E*wbpeO0@|X7&8bR)0sw35AzI_ zr4D&%FWY6y{?&||O8%(Obf?p)4+CFE!=@h)XdQIyV2+`Yj&10|9!I3xFw2WAFtomFh*Ej$s{EkT;9H9 ziHL9si0{o>QZ;E+ky^>046>{0pi*7K%e5(+u=#0rB7ApdI8Yv8!&|6-72fY{ancR( z#5QNdbC>dW7iMA{l4gNIR0NZY1>*W&W#^KLQ}lK#)0DjNK3U(FFeh@28^N_CO3{WW z@R9aT#M(G(kFd$IXlnkp;jPn{{O-~xyH&ZgvTA;0Rf$9&e_4E7C#?CR#g=8_(n;!s zo?1SI30}hCnYZ|wsyVWBSmZ;ADz0V56$g!qQ*vC@V7~$z*R-RqBvMHp%0(ld{V= z;mgVNY8|)P1;R7)mieDl+wwK4q@WXFHdfp*_<2mrU~#7=8Sw`0N-Mq80E-U9;;m<9 z4d}-q&M4t!eLfk#968t|vLLGsRFQ$$5E^vo8m{tn)Qj0;f+~;B3xc{_7p&^a7rdsE z-R1HF?&iSG4wG0D-MbQ#2R5SA`K_gLho;N?jW6e6l)wP}F*;=!XODs$7bF`+knQcB zqRw7ClDvDAy8c4>ymRYhNfz+Ug*&C);iz>Df@>~e&;EF6$$cRrBS$-Dxu*lMw;2FApNWHSiSNMcbfNtNrr*^xwQZs$-hqyS5-P-Uh@I2iJ zqnUCm@vOTAbeCHa2`Q-mO1m#Co=u6%wzZt`3laRd@+_xOBTGWZty~0Qc%E9Y3KuIusf; zWfiudO2@fBXwM-%$JY~w|9E@c2-+iROVdMakDuVeX!)z>e2 zM#(0EKTwD5K9PD1yN);v)Xs~NY<Nq=y%F>^vzK-Un0QLitDSe<5L%hQqvqCz@ zs2HP13=IK$$5VJ!hMs)fg+Mu9cb>?08!7)g(7X)$9Eet#+K?FMU zzf9byq+-UGimIsbALmc|WiL7q)tg@R8qwd+#gsg@zh8<#q_o)uvsrlw&z<-qSo3g+ zgWGK0Bpu4b>-vmzjkN6!s>2!&72nnPc%E!h$fwk-)j*UgP(Q6abdO1N&xjA_7Tc= zq?)!0ocV1lgB%*L1~<-<(sT`g%Oc1}D`A{+Gx&fAu%f z>o}=KOq9n?>rZJ311_2kz>+n$gY9r1NLjhM@%kd-^;Lh^kKNlOGB=m8&7lZ3{CNG& z8n12-e9G)5FY}v7Ku@|sc^y`!4}Glit{PEgJ3=-yMW~D>1;uFveAQwBIr#@jEI(d< zXq{vFVN)aH*D51RL+mFGwo51E2@y@Hz#V4a2^(&|F(($y+f6?fMiU-6X4<~L1=QLH z3x+r^i{GYOvvEk?t+k#+YCW>#7g)TI1lhF%>P41+CnlE7qi=nA*IUTLyO{FPv}ilO zr>1h&XLn<}`A!g%a}bxr1TNK){fL)K!S#%(1_`i_?^XNxCFm8oRpABp;)w2n21p=- zyZ?Q87W7aCaP-)wBagbv%LQqhI1sk6taqE0wUJcnBO)c(USXIai8o1el09tu8a_Il&dV?I$4(>%HU|=g$+0v;;fw);&1(? zLVO%pv9F2sRjqcpL@y@A!B~Du!${HxbhR((MXQX(cF1nU7VDW$|2$3r^QxR}gU9Nj zhHclQ6aIyz2Q0ig(SkX7wfPRc!D_+pWgYi;^ka__KBlKTOeDs1R6&h&7F=Ss*_cX; z`n6BW8*BgxLW*bx*m0N3L*dQhVuGjo`{OYU_56fy$Jr-Kf!6aCZ`j}KVQfQCp_?8! zMvyn|{V|NQD~?1zwz#Ug+rT`AEo~qnc&z@dvIf~F7FT5RJ)QFhMst@jaYH<#+REz-2N05<_%5Fr_z3wsgAKk8+ z(SiozAM{FB+L}cIv$lVf_2`k2_Isb%Xs0#rPbI;M5sewcJy--Op12jp#|}RJG?p1kUU5}9 z1rl*n>6~zGt6esk+(2{Z3~74FG%RkeIynd{pSKoGNjU{Y!}q(R&XcnIA*MVllJghn z-{{m&UYGW6`P)cG2q8S253(xky{vN%(=zzhJ+eX^SV}g9N3>DcZ;*t)u+cx;Aat@U z*t>Fi2rAnaG`+TktKX704H}v?)4!IQtq-~m876^{;yM{zUr_MCg}F*~`85xK-h9!$ zb1)|=a7N##hWf&^2U4(Vb*C>rN1oR{S+}Z3cQ|Hdbe-R;k@J<~JYCLE^3?z;7hro* z5g@Poo>%q`qj1Ji+16^xcqxIj_gfEq*zAe`+Qd-hQDCck*_BBLZVIOecq_#caT?34 zbC#PbWBP1$8l(QZgsS%Giu7vh?yS4@pz#*RfyX1}MRgj=>#8|=V^)S6(-Jyj4-G(# zXE|Hr+j=r)2-yZz#W!}mE9~kod*Rc`&da_UYS=-zE|iO$8ap`VT&IPN7o%u&<2`D^ zM|!3PILl-=421RK3C6Y9#L8|Aw%ohp!~Fw3KjTntc^ZW=8IA>$}Qvv4ZDCYCu*f zu9P6|P}Kd0OyOp-4^}Fugx)jxjKTn=AdL^Oa_s4^tCgZaZApNNEwe#oxeBkuC)RGn zk<>mzb#agcWZFb>WWv82`v7gPnb2t5>m0n~yTQl~GT4ZlHHXE%c!sXOtq^_pQ4Sjm zi^*<%<()c-HcDKTldXdK4RcXl(+B~Il9=e<#j7<7Fg8(DVtHUm(QIv_PV?uJFQl$DoR^A#q@hdp)&&@QDq{&WY%^v$nxvQGHce|Zal}upn+}vi)H(8 zL~!Ja5+A9F2ld0`Gt?3r?4mJuxM%t;xRzO(cAP@CLEGLqeVo$g;FW&=Ndep3!9-=U z8^ZqOPfL~Tby&l@mu5w>_`Mo)Rrb)c3q@GDpX)f$#&PhG6HesnYi$PU35^mvozSC^ zT`NtD+PADtwmupO(>t$(BUBO zZy(Y-UnYflw6v(4xSWPVRx(t$){Rppv;d21I#@3pjy8*H2QpN8-bW{sg16Q>)G}1G zyHlF%3ftNOy7YIky5}) z2#Jnx&RIY&of+|3R~_Bq87QRbW}FC|ai?9G?T(@il-JF6jX`+E>wIdtp!M7M&RcMS zd5ZGz-)Nj7}3D5G_GBrpZE1usr+@#t)7WmIW3c*Oqnd^Kq%5mw)pz7+}Ols6J@ zay$#>a(ZtFhUpFaFq5Ce^s9s%Frcxs+Kr=kxG&I|T=F!66e=}_`=srV6?&;Ka2;f@ z_Q+%ns$@C&gWt{Y<2GH5^%Y_=s$r#@^k{xm3%9C~Q9q{_Axs5%2aoU^D3M^Q z3=KMwQ6UD)*Q_^0fR&D;p^Pyh$w-4L>s2{aH`Q|Zj_^Jfl*MhK`{~>&w#s|$%Ct2# zH3fvG!|zctv<0(-+0Vi&c8F!F#@?>8^`vV)-a$H>Ya4wqvPi@Qx+j>LSLRsbUVvh9 zw9n4n{cB2^xE(}VWCOYNqQJ_)6fYU;@@$#lR}kPTx!>rUIrR4Mav-8exeB@ESXY;| za4U_A8)cMc&pl7)lTuN%Sv|VFf9Ca+0wj1R!*={En_=Fl5YOEm?aNU7*>-im1M1%$ z6okF6c&F!h-VM2x@%8C?`>%#|@t zPx7S@Sd}(Tu^NB)v$hRAU|(Qh-{fIZlKHTai*;~@_iHFb?;@B`KH`U!>^DIM00K4u zSrhL9u=21D5)jU>_00a>CvN@Pi9KSZALJP6n$CVEg@<=IfoW{Z9pkH@yH*V;ZJb%(Ct|OpuG$?V(+(4PVL*Pkax=18oJ_2T{jrG>xQV#hE zU{1074YzKhA6yqyJO4$2@Mlra2H1o8Y)@9FT(`t@T3*SRDzd0YE$xCl9XdJ6=tR^g z(9VYWeV)ijzBT8Z&u(dvSX;#b;gtsRNMT_E;%0!B{?FNQf#1AeMSvqNfy!{fCkV4( zr6C1Jbm5Tmg3F_MtSXP6d-lY^VxF{89j4i#hgL~-w)F%x*>hlVixR>B+u19hDm=7- zR{aGE>Dl~E!G6pu+SG%Q@8VLn5NvHbt4?exNGa}3we@X(ex_5sk9Ze zT}2H5^aSHYaW0y9XTlBDpGr2KFa3v5qHf;+AaGZ{ze-)(+|*>wFDzE>H!6AUDJC|! z(S}jLbDk7+cy#LN{74l+Rd{U+#B~e<@2O1ZA+S%AArd){mxdnJexP}slH>5@%Hq31 zHS^;X*HzYmt~ZwwIT=9@3+fr%b;r+2&MyszO>wJf%nt*nh>&AMaemB6EQc#X!4o!| zYi8>2nsi?C}&R}h1pkcNyZ~gG}vZYUb7vpTYKRAus_F0UY>F$Wz zprqx+Ckj(|G~u%cXq>xNZkJan1!OfFL!2Q&+we+XyqGwVNuycLYvr*>5t0rag$o-{aN&b+?#yvlE;eg=`K zWgRrcRt9@7-)%p4IDAU@f*99DlB&zq=fqL7z5PgOdhNx_Bgd9zkW*x+i=Kz=j7Nu- zijsA3-86d&ZjhfY%XW1Ksi4E1`yScDZVMN%SdSfj;g%m!nU8woGDaXhdao~gI`iuf zk%k~3l^M^UV+;tt;rIJo&MS&v9@cJV@+ip1(YPSw?z&d#WEaZR#3Jes+tS|GgH!WI z%r|6GrROT+kj?o|J-Q!0BlnxEr=L}&sgm|;i2L|d z)lgk!wA5Vn5d$oOumLSUa{$Sp?@zyXk1W}3GWS1MTQL(VuV2=wyWd3?qE)bvB@JKH zG$dK8I&V(EOZoMVgYvT*d{1}1c*WjgGFg^t%W#6!`#e!aC7Il^L`WGf;B05tANS(> z31NSMo3y$q?>K1jHR-ucOora0$*)I+`{;_H9Z-#s zN%qC|a7>2cV~Ig$h0=4X+c4xdIm(y?lw04d#g2))`7~ayIL8kBiiPYv?1w=|$XmK7 zK#-K#5nLQjEe{KF&w1w*SI;b=?Gr7*^WKLdYUH9bcB9^*e&ZY94QN#uB|qKg4OV4X zY7_x>LS4gZptHNj^WUh)Ie2-5df;Tzr&2>zL0mf{3MXGd_C`YN@S?U{XlJq@`-eG@ z2|=hDknx+`(5ryrIsD2ImwWX{LmhpIh<$pAg;vI~p}_@Ql1MpKvgj9-Me(ifWV415 zbTkwYOcc9!KYgRZIT$qASsc3TU_c{3O{i6_BIJn9-d)mXWxJl!@N8Yz7+Q*w{DeO~ zQGo8fn;}22-Z77p<*MV{TSp@D{`xG*2R$L9mGNdIKZGCg=*^#d^>fml8*ynYS#Jl_ zvC+@b!4q^hcCp5@?uOvoOleaGbrd+oEhkA=I_A*AIS9S*li66Sw4-AKoo;3OuRLAV z3i*n#c+(|#qT6i?x}-T42-80Y+?~{@6ocXfzV;AB6}>JCwy4d8{eX~tb85@&UyTV5 z7muM{-u;wfKJOAZuLdL`UIv+4RkzT-H1}{V7HeGm8LK^+}0$m{LWUe|r6?&oD<{N&`Mm9;e-jbgcF0Cb#evqTTzWSz7w zqd(-NRMePZW^J8a~!9 zkR9n)iL6?NqnupY>hECgmw`GB^*-*)CJIVnjOO3@(RS@HWse8pOO$lnIiflGaI9V% z(L8(DHI`Vg#WSDosct8|tY303IRFyPgGe#)4tfNyiXlhW0VXi&jI0(897x#h{fsOJv_m=J+@Jq>3;1>LT2yI}Gr)q-XQ&f9Et zg7SF=IqHG?0;r;}#td6{g9}iu+$$i%bXz{Av~U-svhrxLj*8F9Qqi# z?OZdpNH9B|fO8;zi@YP_&RC-yy^&|v7X)Jkw)FVvAlWq={-*RXPW*@o##fShB3YVD zJZF6|T{koK0Xn)~=oA>l^00T|&~uxLSb(|y6CnuLmqgcmHi2i zPvf#zo~DWy8l~7>kw}~+NI4$G>E5SNAmU2{NwYGd6d6wJ)5d#KCRf)_lNu#%uDI>_ zM2q>rYI;osyDfi#Py$y>rfFJa$X)EOYC?K*ri!<3RkeJiYOQFho&^e)+fFYYQ0q0$ zH|-6NHi#Eo32c0U>o;s$tRd9>-E||RgmWtb`hlW`gQFusi{|z;pDvg2d`^1suQ%u0 zHjcFdV9ukMNSj-K!KRwmNpTN?1=pNgmUgL_mzWq=c&BYQ&}BD;{k<+}5De)`Y|mD> zt=q9v{M&F!>r*Cs3{yX${I8O2FUylX)b{%_3`uRYZScn8JIuq>fv5r{)l&CWTI(HO zdf~(r`^cS72*AumZAIOlcU6N;$0V>)H)~1ubmK7;d~xn`;xVx<^2H67tffZFXZWFy zr_iIe)~uH^i*mt%+TgOK0;Y3F_T0y??F;<;$`>ET*1(G`4&@81vqCo5%1o>a?EIFv zuyP(OEq;F09tx8oe7QS`Z_75b`mJ^-k6aXS9O@JI!1d1gy(uq~XQ3KgtB?R6Tz@nDa=nRUndwU&HPoe5rqQ)euWTon(MhuQahv~g zR@Qjl`tIKS1t1ZyHwCu#ee~5#((%B`z?k3cSuSL6QeUiY)gh&2f_7=`$}rfiC)sq9 zYEFltSJ!hqs$=Ioc=`K_E`>E4$4Wm9LcERbPIr2K(COKeFNJ-VaVQCU;egOw1`3y9 zPnyK)o{@8-##NY)Y3b5oIivEm=0`gnrr?8^opQhPaihh2-0^bNmr23HTy+YnAqx%$ z_9&@@S+M+QROVAiN@5xRLEk`+Jf6%(|2VE=Zbe-p%4$gvF?G08th`lhLkn7}+}Vx? z!`)qGiQDp4fWmu~>vbwhqsJm%K|PkSqlp8D3|D9_?|*P`EyNfoT~pHx_+;9c``keG1aMzV0Ec~sIIZr> zxFR}Kpe}+b^R0fF;^w;WfN$?o7T5X~vILjEOX8v(%wEm8zHJfI(ZvK0W#PPct-2E% z>MpFusmt#fEDP$y?QH5Y%{hm*3) ztV>ujqq}g%mf2%Gy~a|yX=X_5{Q8$2&X#!DyFhW&TYzvA?ogVPje%vcscP?Ad?}}g zqh%z0J0}my<}2%Mt3cPUcH3~$(oR9?g4`4YFxi1+E8y~st4}l@It2-56!ikpJ5IUw zC-s?PWAyQSxdJndz3JQvF&K-`jzp0Lqcs>4NPfr`Y}BblUUBsqz0Ug!FD@cwby=ttyWofC@3K4(aRFULl#)ELgm zjj)C%bq>f7?ijkQ??pSJ5fclj0(wG-dI4yD*Gt7!J1@EowXIHEM>ovV%)=T!RneP}Sc)0cuzXLR;zHHe!(x>NkGF!CESDmv+6qk+P|D5(~Sa6JIF|Ogdj5P*? z`8P-4Ve9z*S*k(+)i95EMB_p+P0`wJ-K3tR!D;mas?;*?P)lx0yf^Z`=v0$+nY(rU z*!k|v4RkKrF5LAR)@2d>e?&K_2=gBR@#K%Nu*YxY@|3KP8|zkIvE-zCpX40f3h!Qb za9w4v=v6TKeb9Z$z9J=AAX78O$fT*)@XJP5vwZ&*@`Og@XrA2-N+ZGRs|aK=7!Xz| zCmihSCO;|JOZV*8hr9L5E3A3+eePKQ@z(H`Kkc>urRUhHRhFSlAup>Z==wNw8}&dW zj*EMUc`-7ro(Zp%&Eq_$!gcFXPObP2iZ^HA7~OF0Q`$=Ett#T3VRIi^W6|C31}#5K ziZsYaAbPnNJrd_0;sCiUSXJz2uTiBUF>KMb^&}_PPT?Z~1MWj*kz9mtbu92=8Tz3; zTc9vm=w#Vy4Fhq@Tce!&215U8~%n~Ovcsf`^ZAjgUw?V@3 z$OUa(zgnnwuEV`2C`?o+B6|E$;a!MklJzwS{QdI9rev7`#XB+>FTK@Ln!{;*R@J^dB)P_#{+s^xwPM-*%-racGT;NYFt-_af8kPAwFSc|H9n$H&IfOlaX@L ztNH5fuWbDfL<_EZUN)QuhU@6?N+_6b4IWigq|V!9VGP>t7ll?SeY+OjmSh==)Gjox z)VkMN4ffnJzjh*LwjS6O-&{Of`R(G&#*qrS^O7z zc$pq|?QD8fT@`8PR_iYh>E|TF4S1chk1o1rpECVH&=!WnID|Jr9Bld4bOkDCTQQj| zB2_zMR%r$aXM$w~yIj~nY+c>60&L3KxFs2pJ5cAzY*(bA-I@6N9JXm=L~lO@XItw| zB?b%zOwnJ%eHqndA*(aH7A1+wH{7<&dlVaiJA2pS0_0eDY;A07YfHjz6cQc|zdT0X z3vjBv@X?O39aqwtDo&x}`1TfD`l0b9E0@vf#Z2OuN79aZ?)3WC`c~22N|lyntIUsk*So%#SVQKZ*Y%RIwq^Dnolmr%35ex-QL>j5ce|X3!O$z zznnZ5{loE551HVO!G$_nvCA)QdD|%ke6Mm*i+ahdp}<5`%9p0o`-8CVc~U)tkCR#U zv$PF7p7_`EBuV7@6277=ucTOx%5&ueCZtDkEvrX7%9yo~E6(NB$~-*icC@fsva5U| zzm%taq%=`&$jc`-C5HL@^GV$KxU4$9s|Kw9dy>IlUqlT7CKG2(T)c%f-$(q;%kry7 zc=12}>g%5!3I7Xkj*%EIN&hDhxv%@e_g4aff(iEjNf4U5f0Kmp`BzUe|4kCnRFHtL zolirS7Q=;z8>Fd~!~(N5$=O7qeU+|%K*>#J@BT>!=-fDJJel!ijzJB|&=55(jd-9g zA~bIE1-p(f<7*od|B+jYbbk^1@gaYWBp&k6KkXQwJ?;O9Isc~@>z_O$U-y$~H{V}# zC-Cb_Kfk+wmT)uVGll=3J<-tDr@*~0C?N1pW(20o?)l%Jt|)C)f$n;h781T{Pxtq= z#$Ok@1#iAasfmk=OG^j*ZN~clz`y^~`(JDMyLPgx`b&&L7$1(TEa+7-TD?IBl5{b8sfAiNO)dw#RN^gni=adQES=h+S~^-Db7%yU23!(P*`V9cGva_%h3fWc(8a94b%{yFw8O6W{Pfu^+ zG)|uIL~ns}9;c}$#1*~n(GHH-uY(3Y)~xmuvDeLza^sRN%b||OAlwL=^AIeZjZAeM zLKtj4n1MGwM^2?F62l^bAzz26r5B146@(5<*%E~jm0PZ)f=>eL6n8L898l# ze<;oJY>2qG`qeA#Y7^9BK3kavFpSN}VqeWB>K7A`r@BUI-kg!b&j=wLv{@~%5&c&=Ni;%Rj@m*|*# zxxcvZFutP)y!JD?>p_^{V@N@Lf@|z@y9d6*)=YAY0oM5_0m3D7x)P5{0SDbBG%+3t37nvv%gEM#k z8<%poZZakN*VWnjH9G5zs@CL$vo`V=3+Ju;Vxgy}AFc2ib4mF=ChUY-!)i2t=3hms zC}O>tY||YtCG+xz)nyr$^HqPny_jFN*D6&f;S5&;Upp!F=5~t|{LKY4`fV)oX&t8i>@ZiM|!?$kcZParv{+nR9NGkVg$G?=7ABQ)N&Z-&Zx@n z6PsoTxmvN!7bM-4G#Pr`>7uLG?26v|Z1<3TOsKZ= zQUcEI&6KI7aQn1VF2>Bh8ToVUWPkilbo-MEk2^`)3ipHdAK)G^V*)1AW!(E8;8hr8 zd>R`I|9kay+`E)H6miHD&l6bHVJO4-A9uRC`x?r?bntkKolw#y*Es*eZ&7fJn8xL< z&@#JlncCUGmjC)9iSFL`1Y%3QNin9ONl~Ol$sTt#gVRvp*U@qgn{}5MGEymauQ%zJ z>WJ0RO=ccV|Ey$`w86&$_lS$SNrz^`Rg~d?$dk8qqu$zB;I{l{Adj=K>z-xu2eii< z1!a393~~Cl136wBk1(wNLh5gxrnr~y_)QKOZmTs5le6r#6~-eoBvX>Ao??eEDmiRx zu)n%^b(XO!(e%^4Wo2=DM&)2U#un+t9o?G_*hdu`P2f)J%L8^7W(=k7pEW|fkSnED za8f=y{RE}HF^<(FnIXZc*8%R-2$z*+p)<0+(khne+kD!8Bx?Is)hrtaf2-xH%6MV6 zU~eCQuL9fn6QQGH2Az;MfJ4^nQo!2nY!kAiVt_gu<~n@y_a?g0Za+Vg#Xi+v&VQ~s zk(Cl5W{c7Hbd0C>ZMD-JsHVZj3Fem-LD$|u-MkCmCzDd_MF0-t^?Qahi6lYH+Huu> zu>E)WTt1g?m4wuQm~cE7;roC=L5#Vd@vKmTJmhQU?WoC}N*zje7Fc`_A@*)O-p3zJP2HvhlnXgU zaz^t7ydZ27$_BPNOos;$idBe3c{5;mp}@)^b(aU-YSEUnqSzZBm&v85W=XJc$Ch>b z^G?XPd*-iE$Y;OT(jeGN)0;FCxvWq;kBiX0m*e>X2U=gQ#9aN=+-4V<R401-}vSK z2=`|$O?p${KHPgiKvY1lv}XHaeP-R6FFU|GwGK-sal|1P41qZ*pXT$`9$8 z4V!#et-5pbIK2HGJjyD_yV0(%K3h}b>(IE(^r5*@{~*oSTbT{d=WL)E?#0ZJ@}9f% zmPeX~WOLcIV|fwDK`+*9`VT_GHx8w9^oU4a30yYvrWKBYbMqS!Ze63R^$$Im_4SBZ zqBNQAmkp^o$0U;8h$@ctKbq1@N?4gtbbmU>GP#yX`A9B`~d7JTfs#3;P6RX9su>3|c4j>baLoHf8}1qYRT` zrThDLQ@uS8+T+3m&R>h@d8j^Y89tw-Y=2%26$~2REd3r1KqI83SJx4qfm6|?> z8&3lb7Z3BD`GASegkYNY+|veA7hTNf_Jzf^oXSZT(1oI!;AbzczN*Nj!5V;VR5OT4 z{7R#uEh12u8adC25Utd_uN;tEZzada&Csgf+9;9#sP7f$ia_Jr15NLLeF4Q&E;EBs zzcVxvQWIx}%jnC-%o;2CVWopt8->CYI=u>MaNRWOGb?j#3SuR=k6{_Ou58~yr9iT- ze>AmLHjS+zsy}V$DvMpj?NHh66P@*kpL!nC_W3T}FRMhg%Pi2s2X%FMipdpk_a=ug zl^R>@;$(286Vd@fi_}?9`oaVP7X&G`tA__PqB-X@+FIpopY6Z791EDJX}GVEy=7^IogZ<$z#(@`=4H4_WQ@16$~EOXIcx%_B5~a}_q4_hP1X$SK=#tRLhdGU;-4 zz-8N@9B%zeOQXT6eQIIdQkcf|VoNS4BFuTVF|CSWN=Ime+F9i~TVt8m`R5Hpl4i5! z8N)9VddR{zK^9@3nh)NsWWSYOr2Y`$_!Pz}>{dDcVavvmoA;DkuHdv>MqN_|#;Z>B zC%NjS;Ei@q+T)8a@_S~cSaYTXrwbhC_j)_$p_Qpv@qD^Q5PhA^n1FpM z+o5gkENGHAY>Z-mA5X%sVl?2Z(4t77M9`D>Ds>mcN556>RS{Ml`|Xpxeo;X91asL` zj#(Wr9manBDn;x&l9^?0uC4RjSg`t-Z29W40d|Rh3eF(I)x00vW>Oiiq{6^F8b?N2DlA`#q!GxasYIX$F`Zd#a&v&HF^{}E8Lr@sZ*#3r^OiMx`t4V8g zJK0%y9uMLVVtjOkF=tb>JJ<;u+WG zIoIk!Coge3m+BG0LU7rvA$`+e2!C`atrrv&qm!(4MSp<7v8!Wn?0vnb!Q&80eJy8R zM9ksO{9rDU6?&G-TUA+Yt5zhnOnO=6;C#UbDotCt52W7#IQ*Gnf(1R;=Zlxd%6XTpeTqg%#d<8fTR%uo|IR zA(nK4COkrmViE$NK8Ro!^kz)g{8mdQG%dbaTDH(8wzWC9X~;BHj!@Dfl|~ zYcUZ{T(Pc3}bvHakKs4rvt@366=B^{+6^HuNV} zRp``Uo$OIjG*93LFdzZbngdb6swFNlKVCV!-69vtYO(h5T;yVp>AsBWmM-cKmX%d2 zz{9vi_r<-%U}QAdv26Nf`>K2Uq07`VUE4#M!rgSvuTcoMogY;%LsWKV&rRHTLcca^ zscZ1D19TR?{V(?3`Yoz1{2N8F0YL#l5GiR11wmRtx?38h8|fMbq?DBIk_G{#hwko_ z25FG)80M^*@%g^*x!!ZG^ADW;g95W=uf5j3*Bzg`W&?(Sbr(_cm;jZ%Td(Sqp8C_I z^K);0ppoe()5e#Q=RmY1Zmd~c6VcDw)acq%qU zAMl$g*erAEqVwENg}+h5_hX;vVpHo^q0BntPeT*hz0I}ZT2Y4u!6ibJOCf&Eqhb|W z+jY0#oBbz&^4q>VgTsUHzpvo|adY8{$g%;)(;K*_9TEVSV@-7^w8QNElDj>!V z!tm%ekG-w;GP1>Uj=wCFEM=VPM6{e--_^LB^K)-Y(JiSHPx7nvS=1}CLmjIU2kWA%4-3?{` zG>J%Na>LU`W$Gtc)LwlV&eGsq8(vx*JfAo|=qoHQbMt^Kc;k1T7|P19x8DoxFR#sV zv>$d^3wNM;f0DQO`)XxO2lv^Un(>N{QKL=0-SOJ7IrWP3!H{vNG4o`O!lre%_Nra| zycM?nx$LM~a=2^K@!DnNvrH3iyY=*f;94%?^mxI8>NIdZyBrREtIQ)l)@WM0tsY;_ zJT6i0TnT31jOI2!kBX5eQHs2xojPYd&){t_+z@@;7Y31&!1q`^E}@(IyY3cryvGgw zr4aEf(vOa#9RDae`M>}il@8u~@t1`IIl^QNx(;K5TTU9u9oZk@MpKKmBCEh>4roOL5nml_h^wxmK^JVRPYk^4G_21=~w||5Y73InJ zc1Is~8r>OGS1rUX_>=eAp2rra=;$aZ({@T3Ciu0-sIt8C$KOUtBcOh%vf>)Do7n08 zu3v4f5$GRXX4Y1Xoe>Naj#Js4KG%xEyS51u+gtG5wrGmTk9^MIf8qt~yw@K6a7hy< zFL-=awLf!VaScYAP=2T_GdX@PBeitIw@lNX-|oTZPuJ!0J72|B|5O=!nRmX;)Y7ux z$B)aU2)P?pkboP(0+D??2V9lBEheX1Q~CK+nEhU!cKvFr%X5>zM-mA}ZRUP0<=)uF zRk>rrt$_bPXdA>3+AS+O!F*rWwq1&!em-sMl9`#n{veQIIef+{#JaR0W@JF}@%!K2 z@l%C&oZfeqS=@`;7o&CZoL27|QOEopkI7qcg3(Z@ISQYsw}f=6eMuu&Mf9 zGaG+;{Zv=8E?@3Sta^1frk`!?WlIxuCmW4Le4Ey;r#FLa9LPVK>C_RqHOM< z--=7_`6OYck3@F=;KPEhcVdC*%)}%C$`Gy6PbghSnh}Z0c;0PWz27TMK_HoA)-#^R zWUpNo4sp0D^tqMNO}ge`K|G}<*?Y&5YToXm?Fq&SPZ!ft;x+wb5y3xwk7EepFKFo3 zdZhM)_sK%brUhzjr_H=ubE3v=-fz}Ckm1#N8Xq96`}4UZXLIT$NJGWxj@cPAzo2U!R*MlI*h!4l*dlN zq}FrBb+P#Nd5V8nq@=^C@`+2$ENPU%H3=!H?p&p8ag*35k)If~*NSwZZVg^$e@N>ojDP-LEbs~q!IDyIK7|d zk)id!5WH_FP_gypImp<0V%aC!A4Z6UtQ&Fa>(+uE{j}y=cI@O0kn5EQcsaDql#*hy zKGicAVnp-C8n9v8lY?`w<+ez7*biXb#)(VQgY8*Ba^Us?7^<@1zAEy z<_}c3nrA%@DBIKW=fthq1=CibIToMfy2s!VPOYL3SC;az`^65v2c_Sv@7!Dz24&my z)1pP^N&#V=x(_WSrfVf7=`?RalK;dxn$s2O7R_ZPQPmQ)Yh?9KHGy`ZsD@j z+B);FXZ2H7STy)1%x76)WxbFpgRSkPs3BUF=v7BcUBkx~RJW}iMAS`T1+r$}`p5C> za4Ut`wL6ZmTtmM0n_~4Q6ugX0m+ea%GbB~MeL+ies#U)6#<}2{-}_v4V39uFt<)OY zcHlDj?0>cTVTU%iiIC*_>?Xk^X6hFuAK@Qf$Ag(TWtAG=Y0Ap{*)e^nXsu%7u64C= z=&$YA#PqKd7jwRqRy6uZzx2c+$HMgM{=El>;@w0{5$|d?W_>D~AuC_kb|)?sRPU8m zh(i>jf$fbSD{?tyKh>ETfmo zCSg-bFnIVX&vFz@Q|lGHo%3x(`ZxdC9jEHg3jPBBEB>8Q_J$u1-_W4MJxJCaj>$`f zl-*v zrg?7JO=(w=non)%x5NAD79A5!zBrnkTj%jpKy?X8vm=-|*j7;<3~4L$kvDs?e|sra zWSj`jsO|LBTPQR7yX^k>-?{|?j@=6_?Lo$)SmTCW9wx?RsylwuHDRmlz$rJGgZX`= z>zb?Dn@a7vFW5Lxhl%pw1St#ecY)XI%yG~8Wu544!mWn#Tf|@AuOnd0NG-1Ui5NBX zekNR5P-H#(vt;^JNBQNp`Qt~kg#pBjWjyB%`O0vn8jymzAiSMFecC~8N#K1qX1emh zU7W$bX7PK5CoqAYXQUX^1mjAH19O7{c`a$FMbyL&B2clM{e`j~&z3~>n+e&O$9;q* zC))hg%Yj`=mY?!&sH{0qteG9*koASCZ2S>|mQ`?7eV6SWSa<6DswZJ<7WqWv{-K$H z7zLi8M{q}C$inzR=rE0suH9&0Fto&ZAiHaw`i0^Y)rre`Z+xh&H14ury|+o3bpgL7 zd!FLs<^^-}b;0t0)0W-kD@ljhspy)wTzJ_5WNjTSw2vPOX(|(!LoN_AoNIO}d-pQ& zYUR=#oiq{d5x;^D-WWUkuW?aO^Am~fGib=(U5`i`*@$NrH4*VP1)Jm-LxWLD4kMCoZC4GeNnl2GGy`w zq|Bk?0w$6i(%dv^b!Jr3-X;}}-=!28odFP}JLSrJf{hw)vkNXG+->i=awYLIddmFTWS!#chFu8ZvcQ(=ZMS{$&B(94v%&Q&qUcR?wUa?4@ri#)CSO$=Gxg~Z*`8dzV zrAw~`BJN6s65e;bxNdjtWAfzl2n^UU#)Yd*b^`VptEFCP3FcM4gVfI%MtwIAqsrn{ zoTc=8M4G)O8$>)`R=__n7`0{t(G~dkhj6bMaYezRp--iBW3ttu+vuAJsivWHxNRjN z!ymt+o-gU_k*6lP)Psr!Z32DG+(km75XSp7#U)#>V&pDIRmL|C-BxmkuxUYLZwa4zkBE1Ayp&Vz}m~vuxyDq|^^eQuu!*Mb}8}U9|RD$eDYaRD3tDQ8DJw9K_ z%=Rzj64(~}~YLR>C zjniD@ot|!9nm@A9Dx3sGhj#Kc2Da{K+dW;w_Q`Vl;@asUNVZj{w?@M)nideO;i{V! zfGZYy`v>I<0UmEJz0>b5C!{G++h$J}3jZADv#+ikyI%`)mfQQ`YFw7SQkC`UaAn70 z&$r9;lq7^1{^y5O5O6H2UBj-2^k&n^26wmI(SNcUM!cT;dZX`LcXJ!mGsX@BJ7&gpD2p=oP?Sxd7Mzyg)S<^sA51D{ zUS)Nb!H@Sb3DcPAi;2{S+^F8KN4wk6=lv~tT0Dl6%z;b?J@nBqWMB(r=Gk^qjr8AS zUrG>-dc06RE<%Q@o5W8Lt+0ljxBV5XifRoRk9~go2(A;;J?xaGG5%`Qr+fR!!q5Dn z1KsIrr%b+_YweAr-4$dzTzMX9t){TBdk1-|_lhGGss^CvxKWhuRHupMIbxlg0%Vz) zj7cn|(e0*Ab2<00y-C;jYT7m(tAxCsMF~I*h9PIL3ChemqU}pZN_};@krf06Y2SWP zFp`zH$Im{Gtu;PiUA<{$p2HyF_ini(PpjowU-#*D00DPWO^N%n!%0{x&sK1v zJO!SGgMZ((D4~4vD`iseHIq+%fuuA!*j=9{z7r3>4z&>jT&#UoeNJV(XcJn!j=OVt zw&QUmy2gRmr5>mPFNXe!c-cIxS$gqbWj$0l0e+C8F`LOUuT{1kHl?XkYX}Z+(|D`- zL5dFE<9BlXOkKEJjcPErlu`a&&KklCP})>zaB}*4&5sw_Bb!Af(Nmc#CcS4_gNCBc z7Rw4vS51(rX%e&aMvIMr=nMFjjRRJh?5N-y@Ft@tF093j``9}vZ>E%$es^tIEG+pk zSF*|-Si@!9Sv8u_?Yc@~A44H8%KVnXp>d40Stw`t;5l%ErGKB^-a7d?uAnh%9e&T0 zSt*&p-p109uP9*pWj{9%|!HGRA&acCLvysG$;sZWyNwACqg*2w(+s6x+1L! z_T5E7x2HG^S?1w(i1|&9_sgA*neoj)9)W z(iWCsB6hsmAdH=3Rx0vt2t{-=+#d?kH^3+kV-_)*4P9fH|7hgXF zdtP;9JKyYn?QbT<;4bTzX>Uu%$*t~WZnpMod~$1(f3OL3d-VjZ4Z&Hy=EQ`7)xnm; z^_5+|`+c{>R1%2wyQdfO7xeMj@xQlRK$vxX0-16YcQjw9ypk97q+)roZwm!6*RKy` z&kVVpJ=A)4@+PMsy(!k*}Hn<;kp1 zjzF=E?D+6#2B;dLz1Al;tNDy)NX^0XZU8czg>4~2^WJmY_tUF<{A;g{k#X8M`8lsW z*0wNtb!LAa$L$~{nnQ`mOMR6n4+A}loi2tvVp^F>?r~W%zwC12k79+}ocUE0jE?8M zJ*^!b6HbT`Ec?Y5$)HC+S)6NQvHaL)&Jt^xJ*T++rJYEst>#|Hm$N6A4(%0|zm0HY zSbmwP$O(A6=R%Vi5)*zNWX`nbSKl7ggj^o|anPRG5Cie9Mw#j%e9j?{ie|LNc>9|n zRezW3HLII>Oa?K_Oec=vf_Ik#D#6FIHaWa$Z*u#~iI;c_0a8{Esp8FU+!27REsvU? zaDTU!HD%O5l)6W{FO!Z-wCCin&Q=9rwwmh}-z(sp)ZppvdX+|d%BbsVcD5o6IZ&BC z7G`CQkqx&T`@}fUc7DLtvQzAzz4-HamlXMw=`B6H0%N{c3r}9@hi{1X*$>DyeCj;k z9N3JzD);5xW+R~%ZvK>eV4Z?sb%M}kyM0J)t!tm5m{W@9B;M|Zsp?Se&6fIe_J3m{ zzk?uhV;KuIjiyzv&h8|0g{ZP7M_quR7l@WtN;~d2ccN{%m(kw80Bd+9qiNq{buypF zobGeEzaA8*SZVs@CbvWPWnSh5`>JCSx#h8b~1}eC(JRPQMr&#iDDVtFvLN8qHnv zsM}NJGzk@Ry&F4LV=@tc)O`#4!R9{a)=UwrPWA8x)@w;!4W09Ft&{gf_*OBilU$pO z-Cf#K^cZcRer2doney$(!Zmv=8C3&x-KfjY_7E}Uqj3?OUam}mH%+Hk~phUjIkkutmme) zGL-t+m(7vQ{^(MS+~HT=SkvVavQ)2!1ZvL49;``qht}C|J%~D6FA%9CBW*4tb0sq{ zR9%R!y$UwhWpp~Tpdlq^otPfQa2&Hnz?6Mb%6-#K*!7CQT!{d=9p7==u_ z-9&Wp^@idh^=kQCTKli-U`+LxUh!L_%EhzALC6mpbT=NZzrjKUlVw14{LKiM^|STrN`8XYEY* zJJ0Jj`~DOz&o)C(TeI3!xrb)o!|F6Fx~o7j#lBJ2p3(jtGuRg;+Gp#F%{Uvf8i5jt zF?%zk4EsW>nO)k+ZRSP$Np<){0?yRy3Z}Y#>oD^K5-wOZc~sNcQCVOKHcwb2J2m3y zv9|eXg1~wH9Kz~{Si#Ie)TQ2+jhIq;8@^UZyUKW_XiM zTDXw%%Cv|`KYVGF3kum->PMK4;$8a)JPlX6buh!PFy3;q-O56MoAye1?60t%gn|%h zz9V6m*C`*KA3au%j&eIe7=$Sh?!G1q2mUvPV#LOk%uPsD9y%_X%5MDmIH12scjcC4UQVcZ<%zCz|dpKHM2y}DPI zB7uxgL=pE${AVn?Xg-6A#$6p!iN=4}!+1wY;%&@hu9i`@X zZmoZDc0@){XR=l%rqvV`()kxZzU&j7%9=x%xuvsGv++p9dh6^-gAS0OiFiP9O1S^Np(uA z(6Ya{ggYFZ9p-g1e&j{vEs^k0X7kM_<@Tgy0etMb@U{~tDKdQ^Fg6cQ56d7Fo zcuKe4z(1@Ojp^DPNfaq(_Qa(;hfgUk7dpI{Lm@G59u7}*>VhOiy+3lsU)!E3P0NDPiB0!_Ia;^h|6u|4-klblnfbS;zE^CQj+)s?%8e!T zL@%6lZ5BFCl7YGt!yu%{^i9u1*7KRy%6PlgTb!GwZrX36tX``3WGPq=TSMXgp*!|> ze9wHHPFP8u-`9?27jRa0lYMLNHhF2VUMNM?yjID?d1>R2D@CkU2ZKV6ZhUSrJV&8?yn1TYhMZGca?8uF%ZBNhgJAhzE=lSdXOD6 zq&3T=FZ$lRW0J?enD11ILy)7orgHAWVCq-tS&~_|AqkPW;BvlO{eMPN!)WiwKQpn* z%6gn_qP&F@@~9%clYMM2NMVLo#adQ+w7u-LCP_cXC@-ehi`Dq))K5=T-|;MHDn0)uU=^ zd;a>+1uRcCN(Pym=J#-auzXT^%BPdW_2;=~m|}Lq{*cU{bDInGaU+{zueVV}^#S5`rIV7#wJ&yi6;5$z!!9-z4X~U6b~v6UiwlCNc>A zc*Dwv$6Nv?b!Y3liiiQbJEn$)hHMf~=Aa3{y*aL5TO~a@I(Ru0q@gbhfqyNxu1kfa z7W~RnIOpc=$*y*PwR3LP;)9p7H##C6*%l-UO5JpAanrQPW$#F*FG{Bqt8tkmia9fe zq?UMjW=ozwZeu6x3Tjn~$WUADDzZ@FvgSlN4fNb!ECT|LO|_4XGT&EeGM=gE=ed9I z4r+HPWXYffvt7nAzc*B5r_)x@zEhr3cr2=}^~>$ROJ2Ymmeb)Lv2xv6E`}^;_pYzu zXuNw?0)|yk!{dr0n|!=3?(X4I#%p?i#5v^Z>9}-G6D-_FNZO4b5u(5^W=~{IdcE(8 zSwz5@)a2B#T==v4V88I96n{KW(OA98<1gV|TDrM)O!M)md)?tpA}-MbkwhL&Wo3Wo zy=;x?M?TPND4A4A48NF0Gh8GK%ci*$c03L-s?;u$l;&#F&O2W*=RlU) zt~yilDLK3Rs&Vny;{_@-oF8ev>ameS4i((o+>pY6>#Zr{Wj5URdPEXP50g7zFF{xI z#Nx8f3Y;u98}C=Ui^282dUOAFJ_{BGz9fZwH5|7L6_w~I@b7jWrytmedpbK67-64D z=CCun^f*aXgp3Ddnj+*Y^oHx}HEk~CI~B~e`+M*u%Ve#4bgS|PUOx6mF%qyjv^c1c zbXtx~({(|(dv0uLB#+YU5YYkc-ktd>2}$5yvex|gy`T~^zZ*?layH7>tNwpl{RVR<*NU_ z1ghS*A)5<$Hr5>U@q!Ky(Q`XQ|MOX;7xFvE=_0><`xX|KrkMMWoIv-7e{U{KqrhXo zlkWfD4nW9$<>upsaqPxd^S1@RXua3fMU=Df1zrHdV5DB*+tz!2x+ay^c*iwTR zCNddGtJr>hNWH{Jgyrl6ia6QW*yx~vp2a{-U&c|zL`FqH&eyZUq#iwgZv)Pg7>f(| zw?Nf=Dct2?WqD- zH9)vWe*YOJbxvLky3vpZw{_5fzBRMPDG=}>XdxrFQ@sP)tbn#H$fh?A`%9guUljE- zXhm4Kowlc0J@x{@BV}h4g5_@H7u|o#M+|$HKU(xu>2MruI&b}ri9ZZaDPIH^w9S`=7XM1QcRLgdy zCoWwq6bvo`Pu@SIP6o^P^=l6bF9Lmd-ksNe^Wde5xpCb~O&6kaIEMHJKM+LN0S=Fx z=;eBa&iPEtcFyRAz*hS=mW%7bGS{wQ5WhD@=XV4Vp_uPLW?`W$NF(OqdFTJ{kXe0; zmECUx91%}ZN9SsQX=?V_-2o zRN#DO$AfML1~SsHu&{1g{iv+;VPN1vu~uSIWGFkh{Jy11hk4v!`_z)pyQnO&@wijT zn&^vBBXwxp)mkVnxj>>v`%Dl@L7AuhuaQb+XK6)+RK?z&d%=s?!^1;DCRO@>jJ%=u zYyBl^iUye#x4Dbi7XO+NkXBbq5R^csQQL(e#Z zI`$UZ1<%*g;TI>mFk2DSC>wTE2|e)h{1}(@Y|Z*`cJ<6Q^r8Z42h&1Jz`Lk2IADaQ zW{PB{Dv_E8{u5p0d}|Af&rDCkN|+A?x(xpWB~shxMKB?0s?%&Mo3sp z3=oXo$KlPR>x=J5dZL)rXrDbpbADRlEYRl%4Ez25teFy3U4b)f0vmahj#3b3DUEva z$h%?e%AnU><%3TG^6$XAiH+3I)Rar&0b8J|rgrjMKQuEd3oQ)mxc@!uZX`ZZ?$Oo( zvOhxZ4l~;<5P&hPx&n}M`-&;JHR9OX>&+YV6uszHFWN}4VWVVXZXR)c>uIRiJWni- z!v-$&RJc|L(SV6dPjC-SMO^+!DUC5eqm||bV^#!<9M@A3h8S2*}X< zH!=X#ZD{Rb;Xp$qJ^(92eEdi$UM4gy?p^*Ns=tGaa=>A5J%nBY6*N_0QuN#do;ihah$Te&eif= zBrr=l&UT^67j!DhFDQ^ftYH|-_A^@Kn<)+M$^r| z+kynYA!a*rccssb^VumLlLG_tii!{K+(1k6aU3MYfQ(LGu>CTWi332;Q|!uo2@Tb) zvNyH0-5=M5!UJ_Lg;2M)nQ1}lW)wc)Wq2kuRS4gx;|2r_*~rZXdrJ;a`q===!|gt&ICs|sosw%Pwh_|6|~ z!Qc;hMMY5f(Kz&^dJcW$+X(+XH<)NXYd2pPZgz3Jl>)tjHM9);{{0)+lwcpTko*>T zv3dO46C@wzx>-NRI>Q(x>GieIgu`Q%ckV%&7QTQuYGj*24f1~>f_uJ5&cwFhi%-&I ztr<&EcQLc$%9s3@-N@&F8nsMfB9uy)D)svM1Jr)v6BClqX0<Gi5kXs3oPW<^0Lo=SODP0@qp`SY9;`#BlNHL z7icy|$EL7u&lA`>G%=yBqr-ywGBGBK|1`Wp60@|8PEiF=FN?;pr1fAhSfk!9=wL^H z8W?m9^~x%d?aDmw7G`FORLMw76T~+BSzIJ~{+=+V;RAlkg5q_`N#&8%)&Ra~6zErK~g$G3kI%A?GODuyV@_42YZKwNAJ zxuY8(3hdrA_Fu<~vx8N+#MhMQdAZT^%GlZ#uNAIXljjuKo!_`X6ffK9XB2hJ8FdQ2Kh8rfsfaj*e1Sg(nZiK*IlS@P}MYbCPN&$sIl zS0f`M)!S84sNrlvC@PvKL2?1w$`a##om$sdpuEUD?E`XhxqlY5{lC?Nu?@`W1qB7Q zwc67fVCC6zDNlh6CM9({+EDq#q{tYhcpDryaG?cjSkqmc-riook0Z=2EpcvK{{w1l z4ellFxQhjxaJ_o;SPlADekB|f`VxSvXtC>@6ZK8m&y&HN=HS_iB*9VoR z+BMF5f>(=zKUDKk1VwclCFoWVJ5|-FgTc_0$*x zSV%JuUrAb8+MRDsfBF!-u?*0%ak9L;{J-qzk*ge%bP35M}G|oS& zFXD07*VnX!yrJZeJs9vlFp|T^rkmGfv&irDQkm?(O*cd+DeB z(aQAAb^c#4^^N`>?BN;3nC}ksl9()rx{akyT#(?7)7!a{ouV7YUf@=us4FK?YieS0 zpFaFmys4?FdPy&B$@+F_*sEPyPM~3f7Kp%xL+e4e&|2|%JdZ;jK@G5g0(6Rqfaud5 zhEW&6L;2}gOGTbvoG$$mrlK@om|RU!e_hQRUMMaxJd)zRWl)PJ*7r5#YB9aUt#x}- zZfoXp|HXRa((PpRQ~M(6FOw%Hm!adc`Bov*&Hh7A-%Wb=*PULbJDmiK-rmTUl)OJ0 z=yQ9h$MjgZ=pAS8@m}YNBnPg*#W2*FL-^Cg@o_f-QeSX3=}lgzR!o<_OJ4D#Z#`*b zKnwC$ue06qqftX5;b~}X(Uam69n+$b5Se~K9ye!h%ow@aVvmyP!8?vq{&aH@S5a+$ z&!Eo73j&=_!f@DbhWv#Tw2TeF<%_c@M`|0E^SE@b$o(yP;ZEX6s~))AOCG;(WT`$F5rCJ&U%px<|4sK`zQmkb ziR}7n$mK~rbW5LGHYM?Bpvl)rDPD9^*a6TKUH+z^yCvLil@o5iRdZ!XIJqIWJ!4VA z^(-UqDVt%317Y8UK%o`h_l0KuqD!$c_5e8P@zi97Snqy1b0Mq1Mg6Pb?*-=rOl~r~ zh=qkx`dZ5z7MEpyB#q$;91iduly;XZ(aCwdpLw3)XyMN6vU-4eBDY(=PI^ia`Y3K&v@Si32vvU4+*#Dkrqd z$=qo`p0vw+oD5?3cbVi}Tvn~h-3}&F-HV@%I>noMeSM?a3^*S(`U1*Cj_X--tx95Q z_WjPX&4tVGmZ)&61tJ6QXTZ-&`bCXya=@-K^J4euQw<9G%ZJI1|R#*+w@sK0+ z;*ByH$U}J+n=%1g|K$-y6~AGg@G1j{P;d=sq6ELIl#C1^&`LZt-_B2Wa)oGp+Xjh2 z$7k*5EptjntXj2y-ol(4d?UmxVyK_85%D8}xZirr*Nz|8>5}8gl!vPw`vqVb5c1bk zy?j9_k637Y$o+xp%&V+l&~iJeIJLvs`K$zM`O?czNlf4e&Y^W`%~%;|%EpMxe=Prc zF%_nMFu$?7Y;kK{5U74bTago30YahWNf!gEFJ@gid0shRWvYD$!j`1=NNxC))*y8 z`UhvGS%OfTtV3(fvVwBxbx+#W!Q1DBQQ@CXdX@%iOB2P&|Ay?hx3$Rv`#fzxJhUtK zktHbZichj6b1a1;Z@^bM@MLnW;VWJOqS0pE^BxVL}hR^YwsyV4% zt%T@#tb(pCbbW{h$S$BFH=v0dr7gp3_echix^Thd$z2Xf#RLXHOo>}WVfr_7gwzAnW%8NZwJ z{;1JbdgD`C!P`mUH`1Gg>vH1kBI-=))Pniy9)}tq-ymge%_WVfkga6E^dHY>UE2>| z+8T+Oi+9a%m*5MOmX%51SUr5KP|SOdH8Lez2klbTo*NmN9vB!Yt9~P#({3qqv>)yf zht;W$`);aANWTllyP&VB+1wd3J_3DzxR!qKlmnN|Afm~#Vup6cmG9}lHKXkZ-Cw_c znUzg$9&vJV$|i9Bhy^`_uixX)w)&P&_r*^?0qShAyUNLR^W0~Z&=gnj;K$bV$_b0l zRYLjh#3MH8sto|-a*(z*qON}ZAIFe zj8)Yv-M1ER4O$4obkDeejcO=w1shy7-~0T~km_kLP03sDPRFVSCe zNx}Bx9cZe_S-^LHW4Z4EI3&RcEOp$_j8}}ix9Xk z%!aphD$mJ@3LkyX9RoE>AkG^jxuL8af#;!8I^`+mQ)i)YGrj4qEV-0_Ru=&~6#zOY z0!m9cDBCSIZUQSfQ(Qs{O9nb1@0RA`bM#<|o9Nh`-v4O$sfNy>XHA?R?#W+n#VI%5 z&UhP|6u(J39O|M5HZ8!L8ufnG)vywnJY zBlD0oWWGb$oT-53mfl^Z$K#zqc zQ;$=dlt(?`du1ajoSHk%hXXXEJxK|N1isa#fYlVv>^rTcF5M}tk?Gr($MB*jKdRif zV8cD3D?azIm1Al?6+2eC5vSQ$`Yq6Klum?P`fu5$_{LcJ-=A21ep9ATcqNUt?b|Ho z^7C)yu2V2S0oh6eB9Cs$nDG%gQiUQ@OJBeKf4KN)cVy$VILJq$ez z)ZwA1LU!er&j9De)|6XL&@x(y1hDJN={7FequlR!l@`4FU3P+}S9Ub3Rn1#o=lVsp z$H`Ah$N4IcC&Y}Me1B)hrFb~PB<>&TzCN71QW~Fqk`))2>>AdI0FjH7G<0{r{Fw|{s8!S9eK;5d;xF2nik?Jh zCVngX$!476e#^J*aNgofw%Om!sdh$bo6_cN`ojA1B-qU|J;C~WXWc4H1zNu;u2N6E zdqI6WRHfT#Sen~3f-#Mf82jA4Ze_&?&HOO@PEc*gX;T0uJU%j_Rb{{4$GbxSv~T6E zzwJ8H51m~WJv8mS3k(hrO5v$-acOs3dogj!h)9|i&f@sju~$QCidThUY;Px~8B{U#5T% zOZr?Z^HXQU9OCHehKz=Wtxz+tmC}MQK~h`iiADOICNABdwI}h1)KvH1+HLus7YgsL z_II?h(MGhiy!!`LfOpS^fx_`As!BOvKCgkdq$B$7o_+DNpvO z&qzGk)FWqY-DX2*cn7y`UzA=oMdkF8{os7;rM$s&E%;xdvzauc^7~|+VS_@^Nbo$A z_iDNIyT5e2&#c?Gil{^**Pk}6DaJG3Odvv&AcwgsckNoRNnuRFU{de5<9WcViVwFJ z$|zfFK6cB2&5Xi32R%#SzMz0$+u>^e&U7h<^4`Y`0+N!6k*LuJFWvwa=M#s^e=wX< zn9QZT^+i(#KqmG`kmBF#1(pAhAT-8+oUm~7W>I}|c=&z=0RDjiTC(R_>-8qJzo}ta zqk6`p@VHo9qPrtL=@qa3q_dpufbxY~LT%NR?VAUa!lG@Zq~RU%P>bd~H32Jpb}VR> zU76N8g*?J7p`hy6mV!pVywiY0MXcICUt%K=G>82AtN+fRJ*-?~_M;`l+unlUo=kja zVuYR3k{>|i9+C1g6bgp*^`D=+`U*YhOGCjSse)*b-JT?vRzRRGH+OwwG#?3#0BTJ6 zsNv_LvukpeR_s8dL&`QNZ#I;HndV+wU zAngB}T7wz?Lf_B+VpL5Rz?*fHTa+k}6hm47b^7(*Abr3{PPW^-zP1K>Is|?F8uzS( z4ap%eJjLC~kpErt#d@29vh82u%%mJ$~yo+Xw4OxtW>X7X*eDxB~-*Uj}|5s+PL{{;=hV4^uGrXoGTaJF2% z{d#!!ay+W!+`J#Dj*FiucYehR+9hElweS*&$79nbwm19$@C676NkM(2Cys)=8}wej zd?|3Y6frre0rr8-1Gw9h-S;I{N>S^^Mv4vO*Qu$6s|xb-H`mrEmH_A-cHQd@0HPFg zK|Q_AKux7nedI6<49o>4Q`GnuFBI}eLF57SXcWNAo}~1|*p7~lZ7nV6=VD_(Rk87; zL{=KCG90@L5GmQB!McA_Mm6Ar@e0 zb?aX356cOn9r@Uv677&!oht_hApx+t-GR60{he=u(EBX@*8dk`=&}0G7X5l$*;yjD2r$!$8J_USqh+Tg{zT@Beb}Ky(Dn(`< zQK9Ds;5PVyt&~A2t0(qFdPzwMDvrR0A5f99nwpeTsYU^6Bt`_P$TtS@Bkx8EBnoH) zhiMh+vE?Y{a{O0&@xLL_y9DF1*-8NW=D+FP2)cmpUfj1*Q7hEqwqF~thF<)0|1|tT z5+NHl@T^f%rlozblQjVy)X;7V zFH~d`hM)QI--#jMTCt?Slcy&Do4w}<{sNvK_y_Jx4;7b`0ANcz49#ZjX#8qwb7B%a77s*{(16__5ZV3V~gk?Mvoon(6xw9>qYrs~&(1XyD|u z07tqAU|1MIOzX^|>4V<}Y`QNFhvh<_a5rJ2!b1Kdr7+lLVfBp?SqzLqZ@<8!gjQ<7$P@C87*y3#%a7~>8k+4&b@X*&! zZz!+={$$`tf*t`cFZyKvGc81e)B&=)*u7x$zF=(ba)s-9#2XL-cUtoA#{af6Ix32g zlyuV;79SrEKG9#Xd7Ux!KY%=j{eQ!NFDx#OVTRl77~lzd7T3KRGNoeBhz#&CR#YO6 z=dMuZwkTjU$)Gw(Z7n?6Z5#M|gWhUsO=IHn*<5T@fZ(OO(?j5$xUCJO0n1?B2Ia@W zYrsPj#o!$|2;}PW426F;{07^z%Zp`YIYf6mw7oeJcr5^L|A5}V^sw02*sw5wvYY`j z2TS+-aJt{knmGj`LC{R)9~c)DK4=nYMRwmGJa=h)Lc&d2|9rI~P7V%nr*pvW$`vc4}&B zZf-MN=**wtrgZ{^taT!yr z^_RuK@qtL6s;~xdutaEuw?rXeh`=dRg-=8TXb@b)oM`U2nrFHYbS>$CM6XNn%uB*7!td_ zzSi#ZkJy9JdpbJfGzSO!KDl*_r1b?F{#Kf0JwA2Fc59EZ!f{Ux4xQz zzzT3rt$9@mbr896KFm!MwWYo+|i+AVCz;k79s1Y%&yB>-Vb`z$cp+D#$Bxhmq7jc4sL$6nS(_^EbY4bzx zv*;H=S|l&~2txEu57&oV*)}+kyc933!n=`up{eXKQfKkflzN>;DUi%9hhoU-Z&9vq zQRy`KOArQDxfsIh_hXptv}PU9$;aKtUoxEV8Rp?2*kb+2)k`yw$_Mo_F&fs;Yn4c;m;m26o_xkK-?@qGPP>8Eq>@<3 z#0KWQ;Po^y@?s7Sx%Nfk(-5ZHUu!ox}*E`T3I_Gs6BmJWAZ{ z0s#j*I0w92Yo+{)JiOVCC-tL{zaji?ZNKV&DYYlFmmUWlzDRD3Q~Bg5&pXTfmpF{O zg_+Zolz#is ze9xb7a|odoA?JR(1a)2hS?)e&b7;~_r?$!$Tk%n(4X!wBmM52(Zyil?JpY-7Df)S3 zozl25&Z=zkk6foRonRSPgSWCeIv#EC^jFf)3=8Rd#@j_5^8{Rq%ddvV3Oh0-iQ+5t zd>p4U>Nlw2M-q1VC+tNPytE9~_N9BUU%4t91!+e@tD@T+a6W zH7gKw`ST8*J*!$&A=NWzvOWHcACG&RIu3Uu3R%1dklOkkB!aE?2V{y~4MWf62nh*g zv+bO--WI3=ekvO0MZa&21Jja?UofiDqVuSOjEBh;-9PK3Z66V}V&MHo*H=}=$R_5g&O|v@v{jIx0wyO^5 z?WDf&rCo(FC!};dA$YsCKc-e!g7}{Gun8OobiDvP+@4P7)K|reVRF^CIW3^-D|XO} zaXBM;%oOGA(**5T1(+c&_K3nr!lKI=DJlPFFSO`Rb_CZ6Np>*fevKOP&gLH7uNvWx z?^(Ul;K?8kd;{+XQ9Q<1Ox4H@5~)^V)7xCrj0;p z)NVRUy7z!SN4hW3FD{tzSoClQakVp8UoDV3U4BNFU1*Q>V%3YrcXP}=e^-!2Yqz83 zy|sSC_>}vx(`X9+&)v=Q%H-NVEBqh-<{{>F`Dc=$@bLy}ta<^~!X3{Pl^s`0xvkNT z4=)kjx)7+sJ_N#4r6KqCbYbqv83%FhO!iU%gU5H9UZ~bDii2y*t08jyDb)z)!=B~H z%c5+i3#Zejiz2bm%R%aQe`5;hnEE)Nxh@oR=5JN~t`PX7ht7o_3{{$l(<1FqV*Z`0 zBRNE-b`B1=dh|7OK}2a^RBU|*0d(#tf6G3J+qRg2zOHiT0Lp(_ja$?w(W$O+nUd_x z$HZ`KW*$LD|G5ClXQxo2_jUd1nwx`5AAER+#e-!#A*Oz1@7e=j&rO5Y+hLn=5#cbs zh47|V`B*XT{KK*Kx{ith0i^L5QACxr$2fB7gJXr2<7+nl}d$F$WF+f zC0of>cG>rkrR=-x>r4?+ktO?9_H`85#v0kT?2H+*?>jRXGxI)1zQ6bRgNJA4x%ZxX z@44rE?m3@pkj;<&YzXUMZrqJ4#jDz;`ImSt2rmTen+o7i5vhPfAoy7_W1nP1U>03X z@WwytSq}7c>6~%xZ@kN5MtQvzjP1vXad9b|d%O)b+d8iKBJi~4l;74h=4!Gi|BT_QS#nF)az!gZ*gP17hJS7tL zokqMk{Cg06iNyf~UCaK-ZZ%?tGrW>A*?K2AB=vl~Z~+V;1PG)s>n!RiRh~1}jFO>^Dkz`1y{U z7q@rrK&4A?!i#TRFdNe#@2C9-#m{$sSB>d8<7ObRB7?Q=u1M55|QqQZ8 za@dm7t#wW5eXE{+A%pm2y1FKRLj?(r#ofu3yJOVPP=A<2Z9$l6Ftq+bU1X;zi0$*I zsNnbYW*NejD`$wwoZQ#bwK~=UaM`AhN2?lu$SGu9AP+U*_zOG%G(|M{sT|g7C$2PL z0SI2M!(q{vo%4CLy-FKt4V8>`<))P8Zo_dMg%Sl4(-=FzNL*j01{ee#XX%SC!;$Bxxl|pJ;EHeP`jP1=&e?_lbnc zwJd+(N*Y-7l`e2$*m>f9C%|DPxCQ;#TD_luq$DHYv0Dq?Cee9^_a`^&7C-kRKij)j zbqoBKKI8x+0VhK4I(3Fg!*&u72NDw(Z7IqwhuyfF&@IcLZV;x^bHD4bCb{Oy1K*1! zR3?b%#-(h_WknE18VSN+9%R-`PQv^uRN&HIZ)E)359vL~I2nmv-2B)$L7qm&&Au1X zJIyHAE&H_sOU|`&GYf3_Q^+<(*!^8K3~}6PV2$H1YrX?*j^+fjP{2RZLLnZmHK4P1krfVRJzprM@0yvSwN-N= zo1^`i!B=0DAa;7d@YgL)gabjCugAw2;%5&_YY$s6845V*Q1-2bgzI4uzy2&7JoCLW z5cs+uHM(9iLjx@}Wk4D_ZhgNjE46VXz^)HT44e)`ks5M2Q(sJs{gO@dI(YnEy~6Md zN`v|gd5G`|yK)fUSa`ep|saJfF1{S~{b(pWn`#K2)J@9o~};x&B2 z?iQl_8Hmc@GtsKPClqcASb{*19OcE+RdT2`L*l#nsH62IAY77lq2S&qsPA!^7F0i%1}n;XXg+*r*}(B=g3uR@Vk(d_n7+^Owf5lE#z02~ z#ww}kSq2&zL!xjRb42il>&fXHG>2Ae+!1oP^A-T|HU-OVra9Qn9E;u5KT#yEBbU)sNC_Y`DG_Ou;szEX#=}tuFdFde^br zUPf`j(J)Wlfuh&R_E7)S;UotW{g5S|`})o6zm17_VG^i{y9=Wfi*`)MIHq1lM*bP` z^B99N4y?qH4@uH(o)#z%DuKk zYF@M*s7DWB#JlsjvQ{w#gGIh_eb~?OdXSFC~<{Q2QI2z zdAhv1*+3g1*E7&V(_odb2Jh7PZ3&ew|1KcJ4r|lja!UMym&c~qrBi1bxNG}iWW?Ooh?sbUEbs@YUJ*%`R=?&e1uYIKSC}pF-8||_9}X9h&?bTxwJsY z6E8{ZF#>_6}m$} z`s0DS@W5T!wN{O=8^}KkkB_=B12c4OrE9qiu0x$?TAutVUduNLFW{EFw>Duv91P?S z7NQbg%On)>cx{aTSVK$)&QnoGzBGZT1neqMr%ar!h%u4QU*t zgRT``V<_<9P*w&>IDG)Jb07n^os`#5mDd1nwE?VJB$$>9)@~W^jPoWVf`sqL=;lhQ4mu92$Y4?YcW`JXqtzb=z0QNc{iD%TQ5FfEdDeDemA zRU7$9;{oI*87IeKxPFxP3G_Jj8)mPw(`>W!v-<0ujv$I0d>^Tb0!kjRm-VA*_h&vN zxAEOdiAp!qO^xLg2suHvY;BfYyJ}(pdWw~(Ni>AQIi?W)O}t-JR4?-|>0cuuhrU2> zMrOQNzP)Fbnt|GG8QQ*fwPKQcYKo^qGtR(VJ=LzCGxzOt%!gj^WW}6|*T_%PdLPSn zROE?nZ04Ap!?DTKsP)#!vVL+yZ>`biO_*o4;W6i_qGbBS;FHY;7+$!QX?mx6qrD-M zII@!s@_L8Z-9|Kw2Qg_X4TED=227+SpbqNV&j#u3OvXk=7e~{DdX#Tl2sw-vKcBdi zMKb6sXux~8-qGm|dcND8O%a$)X<#h9;S}Wa*LWI(3_`p3y&uo68+5TT1q{FgdknVm zh#$@Bq%rVYj<1!c^@etQfQlI|<=Ssrh?j_&1a!$Ia`V=M@>ZY)-;Tv zU=aM+yo$P)Fv*uEz2ndO$-DaHUp@8)i9pyPvr%C(#}IaR#G*l4GPS1E;cqiseXipe zgq14i=hJz=%x}SBWW}6;D>ZdRjqrWC2c|!g6$OQ2Uqnyc$~1l!L-L#|_RHRHVt%l= zboezq(q+zl^K>00XzOy3*0tI~=dLx+aHChlwgv^BbPPTB|Hb(JMU3>l-yZd$~cxPy| zANneGpG@NLjVTz{qKs`g@RDW;S|3Z316%`B== zN!tHQNW&^Bo6K)9-0kbS((Q7e*U2gB#_jiN&;0JP3T2_F`L{5631J(qU~D`JXZMkM z1amZvL7!}S1uh=X>#TrFrXY|GX&T|d*Ho6PN%nZ(KKBM*Wt)GJs0dSs=7d~ID?>2c zL>9^-ZFlV9Ti59g*a@>{`otKy!3V@*o6Q4TxF5zYk4ny-I+M zMfTX%Hg%axyhtn&)%`$MmJRB=ZV!%l*vo5QZH?K;Vedabi$t&Lc03)g=J)x=7z^z_ zpOp0YgZ9eBv}@V|ccXe)0n=wGa0jpfK}S}P=>Hv-#i(%Bm-B8LGcen9<~&aRrJ3l*IcUF3l^V=ogQ_eb7%KXXW^rmjBqJO#g8P7Y9EJ zZssZJ{;+->c$T#mSoemUi+ubQ)uxq>e_X34s$s?{HWuYW|L_g&efYS$)X8e_T<@Tk zcI91{u#UPq2n13;?r70}C&4r_DtDYY5IBLD+N8R+!d~v%UYt}t5gQPS z7=5E;rl*&d4{W`mirmt{8#yO*rcF^;S}*m6-C@UglL>j;P zEVOXXxD-piX7L;N0qqc3ig0r|j-;7@8u?v?za4)k>^{{M6qM@g_F5+zQ|RDCm7RBY zumaEt~=AU$i^J6a80`a(})6F|8S-ty1N%v3M}vV;bi z+Zp@`Pa+yWrv5f{H$2YmHaXZ#*~Id`Vh!UnrQH}#kx<7-+B<+$uno6VXkWU^4EF4k(97iP&>Br8`Dwresj2fg;FL#b&V6%A zsnQcU0BaiSc$#l{>1bs)!w<13{u?Y_pyt7g<@Vc~jFYth-B-Dd*T=WwSj+a4yh`kH zD>Efr)8DBz#h?G`OYMb37z%5-h92-jZXmHKGF9{a;Ga3!rQ@`28JyE-kXZ^eS)I|A z)>Sx`yR<3QXKoeP2|wJZ7cpVs{bI4|2W}rl@sm87=~NQ;RH;1iaoA~*<>?GF<(=>n z6j*K{uHYb30rMzA%a)U-5XR(yhj0?~_HF1IY2virPV9z%mU1kom5hHv^*1H3e66Wm zEoeL^H2Zz~6tPM3z3bQAZlfXJx=r>rr%(tMQ5REFX#J~6Bq0*JCkL6qWGHzbRP+`x zG3viYf^~l!5KR)ndttG9L{nd!7vk&aM9O!(#Z4npt<1qAa_aY3s@%IjSKc90|3zYJ z+E+)s<$}-}8TPENI*DFLA);*qDbyS;|2AFuQE-QSPla{i_3wq%D)8=Nr`N5Na;pQK zK#s%wlaM8l+sh72!HVz2+qMCEG+5euUO#9%=Wwn@+8syo=S_ zcnfP0*$kSEHhqIi9mft<7TPk4zr@C{sqDGmmZ7p&*EbkroI={qooKCYrmr&9pX^?! zeV^{&w4+_l_mmAr=Ku`r{MTkaI3l(PaKn-!8|#QLM#KgCx7VmjGk0NLqFCkqhTEir z!$$OR1A1Pgl0CZ-Z4dI`sByOI6q;p4hj;9}d#`3J_?Hbmeo05YnR2#!i||K?0U4;)3maO|V`CWpqU;8LagvX=%x)Kg}v9 z?bUThgfFugAPIli^e?9FP&lxX`W(1A=_Cj`H5|?C+bULc=T*WcJ&%5Cm_iC7B0&$N z-MXHGV=*$>3GxY;?MkO>EB;r;P_hYRPJzDYPTMJ#Uq3CCe`&kR)T*uK5^psg7fK?e zcMWBj^iKYUYz(}D4(fh~a@t_)F3mJLmWsX5viiKg>D~|8pOY-*Q=|Xpt0!(+SbuZb z&JOJ4Ae zR&1j8y$FTTZB#^ad>uuY?k%c-xu%YG%5|PYn9K-u$*3gJpZZL7^w_KrR4!_J8FDZ0CnvGrx)}plZlL`7C+Wn+x(Ea922mW(@Si!p(pL@9)6_AvvlCXYmi%t-5V3-=cN{gVY_BymNF#wSBWRpD}gh# zrF2h*V)h>paUauetAh!hgL9K%&xc@4IG??P=nKaxV@P0G(MTeiCjlnmUT>^q2u6DP zU@m4#9}*8$_Pb)g!J6Le;txjOD7eiij!QC;2CW=_y})AtAZ2BkoJNUt=f~eM97lU# zVTnN?_K|t_HwKZ;%d-?w&(t2M5LaMcVYijyHQ+JpUAg4G8@3Nxc2m>)8IuvE^Ti!= z?Y-Ut1@7o1H7F-|iHZw^JhJIB3!LE)c?Gv~^;{0lX!6Hnw z{r>0$yU?pgbvgagJR7VRk+f*emJ9HUR#Wj~>9iE=?|-_jElOBf7Syj|3O~yb>a#6U zB@~O2m6mr85u5sH!E%a_+u?0d)MN`W-$)bPx>9?kVG_3KguZ?QANv(tZ2g^_h7L?` z1nM4OuW|K7eYf0mXqAPo98EdXvU?o;p4@u&Izeg4IS|xAPqMoo$M4J)!&r+Xr{qvv zFYYKW_kCd(8lI|^-tg15VgE&qhW?0OC}V%%tft4K*mtQTgu5hz>TGO0OpH?B!dGWz zsWRejQ00rg@HbRVnj9l&tk0C7wRzurv^Qo-PeC2iZ(L*L(BBV6hcipF1hxTM`&70B zJs$sN*!D2vW12Lv=AS>CJnHY42$sG`X6F^wW1zi#2XNig$1{a=?{js~UenRu; zB@(BxwYJVIqIuaq*sJn>1s~|uII8)_`N&Fx+ZHLP$ci?)$yN6}2ZF22ZhYdu@9KW- zsMWXS+0yPlo4nY7c0ER}#9uY<%bKe1%3z2|{FPBABE;Y=T7in&yBS7;>+%>S{uDpI zqq01l@>P-1^zVvm$G$}_!)n1H@aSZy?)5QBi3*SGTNiHkwhpvIx=kk^)qUrcytLO& zwVn8fW&wIKV!aYyF-A`tV|!I39WVm-wo$76f>(BZ#jH)XxFlw8B*L$};~N`QAlK}w z%gkwM*VgLg=T79$v=R4x`Hm_2hqOhL@U0tRr6#p_1dt%{qN>=b(BR_>Fxdgrx&^D1 z(omCNR+sMg;eD)Wt?L+lnXlWGx-i6ut;3gR&>wp}E3Ca&_%xI>G;UJCgQZ9+z{5Iy zV@aylSPfS6u0by`!!r?6j0JW#CKM9VrDWK|_2a(AgnBOTdaHS*GmA?03u|w@u<8Rp zLRvZ*B%o!o38b) zF+f{!WeQqWUqA$uJEvaz9G<2H>E`Lu8BHhoImQUsb8enEB@zIi7`bf6?6mT1@9G%C zZh)?io}N=e`Ws{iW^>g+UiiXIdzY=-I%}fBP1Cj?Blm8iT_eR_zB|wkU{QNvd)p*I z8VVQLSG1(>rxsuc0Gx_n=B@y`Zg$jln$-wgv1dNBE90`{sKtg_wz_0fq?c-Ck^G{4RA|xw@b5R-o>w&LZ6hcl~iXgl1 z#-p3bWu_*Bl))w9PZI=YM>_SO>2UdMsw<1+5*1s{)5nX-DAF4fBU0(ZA>7P6m3^YO zB$qn8g}t7?6{E1-^1XW7V^6^`N8)MT$6{)Ud-D0+h;oCk)|1JXCx1sMA;h!4Go(Cn zsgo2O5b9c5pq%GOEw(2c)*lhRLxI9N9&YGytPeK-ol*5u_pCYIaGFGpPMK3gLLd;X z;?e2cEAN0%#r-9n92J8|pL=>icIgziqWy&3Dw#3Di#8;Qcc8VktY%h%STvq}!~EGc z@u@5AO33W(sgeHlcc?wjlp{U#*eyZ$yWY3#i^nUEdXsGpfMZE^-52}8k0gN<;+zlk_dJfg0d7It+jCNliG4ElGoQ4K zxtN&W4vJ{&TT!q&QqzfVswZvJkh_1_yt%6mj#ulD0$R``|15^va-qsy{;B%n!j zDc0P@L-2aEc~^IL=L2Q3C;jAqU5dwl)OYdb4mvqL$(K$Lf+5~h4uD)yZsKEHfTks0JX1;oM|2wWn`snxUB9hjh_jr5T(TE)n_x9beaMIVmqKr7xtahJ z!c9mw-(tbzX6%VHE=-bbWeDfA4r#jTlAt1kO{W?rsC&_}ds3SgCM;etKMA}XhYoIv2RH9qbsYa1|bK12d31CNI|2R)wI1w&x6 z9JFZefhoU#=kW!*EGrw(L4+VaEJ(w4kczq**PBE1^6SoS{ zsJsb%5y)(`oQ~^Jx!h7t3#!6gDJM`B^mXiSx@uy=Dq@;^D64v1FGzSNl;VRO$){1Z zx3`eb-fxaf*)omo(~)}<+vp}4@TtY#S=ZC^YWzZzUl{GjFPdjt;H+f+wE)ImqRr=v z1tOCa1$WFCHL6ggP7ij+Z)0L@RZR$;Gq*;G;9jWrCaaB(-ij%ZVKz9@Q0Gmjus^c! z`K_*VEWpwC3%>d0_5%3m7lGu4ojrM(fl=>DK8A$_&;D?OG2GWelNXP4t@uIX_eVGu zt}Ui|y#B=p7ts^d!2TJVl8xZFVMZ+_6YLB!h2~d$M8XP<`Um55&|aXQSc>Y3TU#3< z6`OgCjL90OYlrQZJ702Be@@vu88k8RK4@pZ_ep>)XsP+yq!Y35{%55>&fe(O7FnY? zK(nT0`x5D{h_NqR@c!BjDjPLPm8wgz_k-GXnC438G{5woe_dE;cOP^(Q^4;%-9@99 zo0*e~Lcu~XfHht4u3K+ppy!FHR^OkFC)2JJZX-&RC*2hsju#$yaqD>q#6pnW8_dy< zEUP~wVfK^Td455p?>88%E#?oq0W!~rc`5)m&^5b&DT#e%au7Nkmfq3dV$I-b{%H}Q zbU2M=1D2n?)2&A!WGo~P>FRM3@U5FCQQEKW?rYFf)BV8p+(OyPsfR6kxg)KKjrfMtV;d+tVLR?e|yPz8Gdq8TfZ7o?;0e zYAEkol^jtUix;3OH)!D0MOE;sS=8JQbm!_fZ_t|1dL}GK+`Jj97kCip>-Xzpze?N` zgjo9M*Qj@qzeE1F?#d4v(&PFI8cOdOKzDtQSC=vb<-1$98c*Ph-bMUmkh@`g1__Nu zWgZJEVYDt4i6w(cYLaX-r(XG%V*2A9X*o%~zhG=)yem_FFyfpxE_Y zOkIXw+9aF=YrkCM*$D;UFmb?X8{>0k)$;>s7{iK80}m9#9Mh*%P!Abl*x-#h?&4{= zSJ?7acDrvY7=x-vxEkw5+}`@0O18bvS8H@R|H<;7D2Y$o0UoyP#6<@2>-lNL;mUx> zclS>Xqi9biY-h@VL@-~e?YqOz+`Q9p6Iv-P`y}d)n?Exe3HB z4StwBo?VHN2J|W!GLl+3={ezi;r`J;#s;#b7Nl*8iMII^m|ne~w3KBRk8ClYXknlTI zL-7#d>FV`FThp{WK=JazH0TTK>pZOX4?IR%TBVP*=2S`5Z3W^nB!VruUhF5varA}_BYX0sEO zBKH-EX$;uJG<;U$-}GOa2x}L;-r(E`+b1SccQ(U~Lu`3uvE6E$=@L)YPnN1A#J`W} ziuKp&ddw0qUaWm#pHF72=@fUH#A0IOQIHY%3dJ zqE~uPkew_X6M_p=l~gpnYEs~XCS$z5UGPKDNFq^tRKM}@NqLw9D+7-oNIkP(rwKuE zbS*kl7@CX<^IZQ;`&F<6BwG5)cz!pz5VY$yqTwg4&UsiHC=DGb7T68@Sm_7}v@?zq z*F-N8<90UEZ;%2j(gF5nUJ0ViYDHo(r3tYHkWc_Yu|TLhd#v(lzY+*o{nJ)YJBBI;pBWVFI!p&|!2loU(j=ZOPvhb12W#t30Edli{JA?c|-qaR$N`l_^XM-xnn5Th~g z^eN{U$Z;Ak54nPPEZ~GOY^|efZLTk>q{uz)ouFl1KRXPd0w6M}K_KYpKED$CW5w`z`I{>y_(AD<6Fe zJ00Hq97`WdHxe{^*w@ZdHs|1vlmo!oF>&E!Yt3AFEZ25lxO?zICDKzs(I zeH_Ea8vgk%&&nZZevGT4NZ@{dbZJ3uond)L$`qYU0*@(HW745k4kol!haSvtJ&9V% z0wDUVT9bR((vE^jgZ@_>s)zWP{8pr$io%7l1(I9NeOmGygHf;vM(rGF$8>3FOlSzK zL*696emhz`dr;D`D+#xqR=j%554Hs83>>KneREUk_v#8%^bRGNCiust?f1JqSJ>U> zbuAk0!C57v){St+7_Y$bFX`!zMrJ;9No5VMRZy*iq6PR7sY242c*fPJyq)ID#y>5X z`Z{Ym%O6sbQWJwMoSd>F2K#^b+GhyCAAmB@<)a#LtMl)&iCQK90!Ql<7Z(QroNlMQ zclhzB6MniK*OaFCb;gw3zSyTmu1n}1#fi`K=XmxHwcX!=r|`+miYKz~0CvC=09{l7 zwmyI>@!1KLJMi25gO5~M;QTdG+&B`8G^rT$xC0HEO|9ka1+2b93^U#$>5{>6Y&RPk=IK5iL{~2an$cQDr3eAD(T`1CvD%AAr%MdwskJ z>0AmE-?uRN70={iD5e2K!&e(WqL1vC#=#v$ft>j>h$NTs41>4N@8RD(3d<1*lFC!!v71La7BbMfqV{DM}UkY34j~%Uk$>}YXEU@aQGO2 zIUZ}2@)vcF!$01jZ$&4~T!xtg;d`OjoEL~ypK17TvRQh|?m57x|Z$AKsW zegl{w@yfX@;A?PaZxxk~FZgi(Q&f!qd+PBR0rgx1a=L*C=YMY|Jg?n*AeSAFU^RKj z;6pdgDhUKJiQ|Vc@aLupunt9+8kOKPl?ViSd(H+4LA3#cH0T6Ki1;6SWAa%{>8VAc ztD_U!y>Q%R&;X=w{418^U$Jle0j{Pz7Be98;gP>+mJePSxeVtbM&(;|#|HvrX=a#J zgEJWqoo$kxzIZr8?kw*nm+kxu_)43@UV`NCcW?ar0Ak}ac_&W&CkSI8>WGZ~k>)u! zabnw9ehH9m$hZsSbYM@}MJ<`x*yQnV5r{Y+=hb1u0AuCFx4`u4=}|osm$1S=as6d{ zB}|Lkn8xz5vRaH58=iHn-KT#Y+axJhq7Pm(p2mX?1gidcu?ifK(1~6JZ0D*?>3JN= z1qVKBX45jhw&uGgM|HA*B7^qhFw>WvYk@>QP0lkRdWTQoiy$BnG6C{3`oY+}{}_ok zZ+`-EC4fLLj|@}(-}f&NJQ+@S+a$7ObJ1JI(-YWstQuQ8OUi+>ofX90X<2T4+RpuC z&*^oCi{4KZ72BC`hwUBBy9?d_#`m>zISaioFffpj2~x-c$g9s9^zTJLsXeYnCCCYz z4ALvXXELZI`mei#{^MViJnxRPp0Ycibxof<54UH4)FT&wqw_Qi68{{ocdiF|Y;xlJK}X&`rUq;@+F5kT2Mf;(5?zxmg>4dBznUqbCn*zGWLfkYq)@sBLo*}&ekKA(Hm zkpRV5#~8r*iibD|Hr@VvB1r7oevSo10{eylgPkX!QU5ib{=eph2;)}?bL+dOL*u_I zH|=~FdDncT;>QA)YIG^(Pw~;3-I4I&o#%>Yy#t~);KUQ z5a3DGE;#w0=HWW;VK@S4nzUR4pJACL((%e1UWz^R=PlTD){s0H1_%L9Y1_{x70I~{ z`n#E{T@ZiYkOg02qT%@~@qfOt3kwHmnNF|GXN;eh0Au;A-ZUR>{fmv;e0HW~ziHYI zWS(Z40hOisZ>S;K7u>|s)RR=>V`KlBCE@2Kzyq;I891@sgMRFKf5UmOJb`uQf8t7U z0qv#4pQa!K%O^?e)zer{;PTF>uC4}h70(O`YrgaT$3Fyg(|j+GGq{}~f zNB*lfA=e#CUek=sWwfYmiMZYrJ$(x(+=2J#<8a_~D1tI?t?yE^16&NLvW8RUi@$T!&U7- zoLS?x&%cf(4X(f;jOA#CS-$hdGuD!cteKV=~L{ncm^jGl^a(c(y<~u93 zBdHI6g>P|0} z=rozM9|d=(hZ|p~tiO;o`^r4n0k@jGlDhM}>2)#NqbbD)rfFqOU(LQ0M=))tILBuB zq2$RlM65pn8QkLZk*7+bGluP4R`{{$>JnAJ)Ng!noT;exT2}w}Jv)zR={s{-YgAyN zWzTqqf^okAZv%=qb2__IMzId7=!)EyK>L=`HTsut8FxAjC&wU$r5|ZdUVmcY+f|Iq z9HvG|r|*@@!gaD=F9<1FSXj&#wZ78gKC=qm zBt@RphAfILzSy);+}MAJlrov(qdphk@o&Tv$e(0S*iYQXU;00YrfRrAopH&xWB;nrgk5qurGrdO^<_iy~+5WM;TmkDqAvCT=v4 z?=FEV6rfZ&8-M_5!|D7MDCg>lpG|8SBq`Y><8h1>;>)XGqt3_|T8rW33SFSdeMzqU z7jn;lZj?@sP-+?C@;!c%UcYK`k+_{FuE1U`MzltN)xAcGgDUpkvY6}f-wUtNo~FB% zo<4BVyuF{3Ssg~T;_i!NXR9Az`JQQ$9?R%vWT-^2-ya)niz3pg*lu$fjeS<$GO+&0 zT(W*4%I&E_d`|6Fgdz8qk)10EOiT1( zYxRrLs(VE`BFQyn+i`DD+Q!E9-M%Mtiz>mD^d6a5m7k+&+%7k|;ySx~Ct4Lu3bnZc zY#uV?|G3=xGgEUdg0PJ1=@=0mS?Z?CJf+$8hkjOChPct8Xa~bY>W15l?Z4)9%0#VJ z=1RdgD0EkU+Q44pgq>!qrL4IUK} z5|X+(DJVNzgR}20@5YabmF@?oWcTy+j-YHV7*R%ci6U}69mX6i{G?a!xaW*~xcoVA z>F^ce`{=SDy4x$wseM&y{Mg^vMcFi0Ac;4p?PZqo@0AZ_MD*=Eks~6Hj%R6hIu{M} zDIJALe7&dd3TA@QCRU0(Rg=j?amcXre4&b?I=vc46f8gmXCvrTm*_Pw81m{5(aQ(V zAa^Nj#iY{1nP+wGZ15P?*fuSbl|*(0_BKRCL0{Fdc?W`tylazJSNjCkE(y>82dIp| zVD=S8*Us4GHo3}A*VQ*k&vH!@qa|#8Le-ayMaG1Ps`OG{uR%FV^JvXnGF+Yb}&lG!D7i#}&CrfoNo zNZ_!kZ+I;3)RBk4j()wbERqi%6dYaeaWDq^eYtgC?h~81F`V?V1Ye5u!wtXPyKP@; zcAqRh%NTRo^C|U_%wsr+W2*Y(o*Q0Q7Sp^yI7D(l+1dzkhbIwTuMQIy5s_~IoJ6w` zTj4X6>3sqOjo6ucdm*b zEtaCmkU)orT2flBR%iq)rS)JhrhC;G;{39*Sl4;LTU%WriReGc7X9<48r;NW3<`;QXuhu`k zm0+t(;_=>&W8PuhWXY)ExwNI~l?sI_CDz_9t<>5>(gDP(^SJN4lA*<2k8;Acy!j2r zFV@SI8Pb|=9SX@8s!844+yL_oAX&w)vVR!?B@Nsq)_mzYe2PnGtY!TZ$P-HlpZ$HY zgcSEYpvits`a=oIC9K4L?Y;QBKpV%tTvau&C9_^aBDec^n~q!p&w+gCQ@y6+Rl<}^ zyU_Ph?Uf5(`%d;}d$3Bu&&=`-^nC)dh7ZLx3^o`Zy=%GfK=N|`NyDpPH_WEv#`2!( z?|UnPX1K^HH}mOGXgr-I?R;lVt-(B*s?cx-q7hp~nxG^^P^AL!@|@Lp>b%Cyve)|_ zxP`p`Qsr{|82e?}YxCr5pQt^%jOfFO=c#co2RLUb3w=WFo8G}V$>-8P(t9V98`J%} z`pO5*{DZ}}zM;;j*D3yg6R$Jy9#Hc7CLLqtWK+fDi}LB%{``gbBVZB@%rur=A@QxP z;jz2j2IEtjCdeNys@?+YV%uwyl~=msTaQ|g>FUNyxJ*sptBBg12LfGJUVq%(BMO*w zD&gZ7TdkH!J~O)rbk5ux0r$(n4AF(|JqKj7aojc4LP5Mpp{$;{n;DK%P;0#c5J^@+ zH^ub_Qf?uFAWd;tal@~I59Un+cWxf92@Ub+e(*_JrY`v5WfSZKpVQ@}vNU|=`QVk{ zFwt6fJGE`pB?c;u1XGFnr%s%emXvfP#QU@d+a<3&T94Wgy`B~{uYZB!g$Z3#P5{-3 z22`i?eZ#VQyu507(Yk7!g}wA>e*4-lby$KHuKy)GCD#wQJo7ybyu;P!`CVpu`WNX5-3qmkP5sZfJ4=XaRv9UqqXEKJmo249u-c4A4oVs2kiRW&v?)}JimFi~c8UOn8cQ)4yp*G18kmh?0SC^psM zQT?86LfeE+FF^{c2ewuBM{K!9Z@`;u+-hd`Y_)N z|Mc^1#-%_51r6c%Z`KZW3DK}C0hzs1Ce4pXn<$LFoBslsj4({t7Vu5{IEA+y_G
    GKP@pJ5kwFH8`IAP1cKC7=bpzVMMd6d&%DjOMGW>Kt4mhL((>4kjiqHoIaL z%yl-C%eD&wds##zvxK|9++lUrdo+H}iv5@@+Xy>e7E{T(z#Hww1~Jj96WUWKWhWwZ zDnT?4H&(INo>&D}3|OIvWLX@m^b$mlWq!p6>Z9EqR972HVu01in0Hcp$eft7r|o2; zNQlUqJ)p1Ls`u#w)_W^L59P6PkGs~S=G@Fy!|!)|u?`-{r198GexZ&g<$t_f?gZoi z$j`;MTFZRh3AyPxqj^NCfUZ^p*50_1u1&In-$g`{LUx7@ku=9f)XqHOKISOZn#RPP zQ+^=4P*9@(44<(9z&e^xwyyQrVgrB?>(!bx2TXcKi+uYzD z8UOX^wDk8NU3Q!)(yP(zn>NCuEL^k6V(+~?^a4|>-F+I`k3WY^n2Pxw_T~F#^2N(W z^8Hz}q=#3LStJd6&K$ufovL_ncKHMMQ8<>7MqInHySkj(7w z()3w>7UJXH4~nPRMe+Q!PV*XWcKOui5qRlFlJYN?U-POcp~%w)#2Yg zA~wY!o?KLwfr_kyOxBmeHp#0v9jnp`ZPyrg3!o2s|4IfRTXN~v89d_Bq+F;TN`~j1 z{KcmArI>*7)=DsxLe@T^KUM*0>1vG%=E}(^y7z`+@`?!`3nqW6jYOPd0d~*z^4Zx4 zR8Qk#0NShsFLW>N7akdJs(*9lt!P~JrLC;1x0u1|gdwm4l{>zOFb#%rOQ#zLUQ{>V%EmTr^5Yy8H2 zoMEeRWF?&ckyE@(XlKla++fX*p0Y>M4I_Wc z&@MOUmWy5TL7N+rDUNg1t2_}*Ul<=5OWzH;-78EXp`Lpokpws`#<>c8?&1?%eXs@~ z5~m8xTX#qiM}L%T-<2Zb5IV26RLs9(Slw?9L_o$jd9dmv4;@%Xz_Ld>LZ+t%XKfwI zeFGS>8guv!*p0MLSCSd`4EM)!Wx!=7)D9?uf2D~A`cOX)u}93?S@k3ra!1a$b^l}R7u z)yon#d?GC?lC?CkF z1#+tUQXca2@!>av3Eps?gAhYO03|GbWPUvdQqRVzz{Y=Qt}GBA1yHhJrvW@B`s``& zIS^KDVQ+5_usUHE%&_wKXK#M}2cGsF0_O$*CtX>-Z2drqUt0)B{zFImZ0+rj0lp+5 zLBYk|lr!kM$^1VIh@^a?cxYwbt>&!2mj4=K0i1Yq2BlG-nXhLb06g9Q)i?sEF=}Jd zduLYu%q~%AKSyJ#Mow`x06z$8nNei)l{>4m)_<=L1hzc@!X#E=9{u7oJ#2m!!3liO zIU_eW7vS}poe5(~f9R5aR&o~WKm5!UeFBIdIH$!;RrS-^vxjHN?EsKm`AkXa?0tW& z&mpC=vh8dt&pzBlfd(Lc&{4M;zzqdpz`!LSdgfA?N7Zn}M;Jm6zoQZ>atF!pR9_3( zx#j-Nlr;eI%s7~pGxuA{)Xl%dX0+GeeI59(#rdi>#Y&g%aGji?U3+Z?-rEd{yxg_`90(LVxdACC=l$udn{cy@A@ccNZ_Jdi$AxHRC<+{_ggNzCL5%y>`^umEenU= zdMDx|X^T2Vszb+AVi9>}N!GVbzppcV%g@c}0Z})ebT`V~(Jy|^jBjt#_*q9--^_Wr z2Ga|%A6PvxpL3z()4JHE)}bw=Z=xaOah!7TSJYLQL|Uxm-Jts;b*p~KFRyD{Qq@V| zGdTt%_(>t_lxO7hGI#uko4i+zPA(QI_zi!f&Bvb4`p4RnCsr;+^2yy-q0|TIUBe;q zBwUt%Heg3n4I2?}3lgm}QF4X@nCe#K&JWmil%Lk`QVP3!ccmO!36I z6%kn1ou3Xt&=!nWy#(;Ac|pT~{e)(7fkz>Bp_4u=_y!R4+&jbN#AV`-qpV~@^=~dPpr8J5s8TCq^9e^H0cmO)c(k^^OtF+4gS_R10*M%?2HeGIEv|2EDZj8r}97faLcKUi+s z_J8~kQPx@f(ZIHs+m6uF4LLe;oKD2Pd~_HL)_5LQc&nbw#%Eq-&G(J`_RoXixA9tA zU8L&5!x8EaSwbEls^PW*4kPSVF9iO2{uI-AD6AX#c=%oOp{9Ppg!!P{X2^YtniOmk zWd{1^kZm$1jORvPX+=kR>|`XMuJb|cz57$WK^$+ zrAyDqwShYxmD!Nzdw#k4167cbJWV-lDOC5RULvnUS3|Vj2b%9iQ@UT$8hahU8-ITx zn9#@eRbBFiM>)_2C*9*&N6djqU73ZasA4v@i#6^vqoU=()xUex1{GGOKf6d zR@J0zGXz0X_>^gN|D&MX%Bx$BMc{>|U*5@Kiv)|g{IGhj1-IV68mt_YH8O`hNL=yo z=yWXWfwu0C4yv@it;Tv(+SDX7otK7{GPFl~C02@7_D9>FS~#*bNyz=WyHM=A=Qq95 z#9o}DTPDnpQG@vOwK-;-(Uc>KpVB~dkG2SQ3%44db~cmq4&PY3r#S# zqiuS0BIqHa0hPz*CZ61feKo*r6Gp)P3wOwbW8JNg9z6}723Re@#8-l8XlPF9KJjzY zT;fc*c*nn*Df6WBr3f==ET1F&ngvC%t?c(>4q@{DuVL~3G5(T zJ!Wgap_u7Aj9zV`#!d!ZZ77YJ3*~9Kkzn6pO3h(nY^j1i*xjp!_!ai3@pJMUIz}f` z)jS`2khiB=ZjgUKOzNjnCaQdkqw#1+Jsz(~{)$q4`D{4X{(xAUBG~9q^XWvgs?PqUgUHZ1 z^#lY}JWrfmxKUQ$cTGfKOJ#^L;=syga+VSnk`(&0gD9GA)p3pP+b?IdoUWgciU^Ak ztN&BAy`7h@zZ}l@$kczywsjk3sMWeOqrkDcm0xrGiNz$9TBrtAx3|JE=8&_XZxARi zo9zC61KI$oL|I%!7wovALtX z0{3_>2_ylrJjm0dJWpeMrcezi;3PQyXboVQdU^(Qyt#G;bHfb_RnxY`zy@`={U|Xi zU?5K(CF+imLHpz5nF`kV>BdItEGp{sc6(VOtT7L-_f==f@3{W;`9JdB`Yp<~?Hblg z1te7(1SF-qQyS?;y1TnUML@c{ySrmZX&Aa0x(9|FI=)f7?)Q1_?fw3MZ~M+)aJE6t z^E`4t*1p!-N1`_NLz(?B%#PiGwgN`Zp*m)&W>AXpdfUe~{eiFJKH@N>nq~=4hGlM4 zE%vS+ax;ZClUyBQIhfJUUL;*nG01(X6GrE;OfynHSicme5oN-vITb!gY;3D;ps!jh zva(ldr)ArHRqkQYKaHz7bDoQP*H?UDTWEhVBCZFbtJBh#9e-uZza;_-6V4=^y|4S| zUVtxK{D%|k>J6+-lK01D+ye{)Z+5Vvl4`K9CC%X;*Iu+!mf?VM!enOcc)Z$}hs2^yMgc-6Z{WXaMjI}t&g|DaU=T30E_{(Vc*TlN_(qEG7t z;Q3%ox%eIMbU@r9HXAm%E>rVI|L|V3ivG%-GPCKa`hM;HhupmflDtt6wBJHp9?hXcYB~CYmzJ=NC@3Q`TG> zD|qUPxKb?dwhaDh2FnpFF>l+WZU*5-?R&DC$^*>xk_F$n%oXbtS%%P3AvclVzj*x; zgo$EfE*zKjEA#zCKWvDu8|`5}&WJ5S)!``ai;V~S&SqMnzZM!)u}fv%8>VaOSJBc$ z?Qemg)3+R{E~s>KfqwO31snOVjw_GcWTqh@bOgfSdqr2k@@fQaQm9GDsfUtMon+C` z9nfJ1;gWpQvI^?@Qz)h=p8C8E=ATa2o${`2V&b4N@16IU1&^-=-j-ku?_wf%BVny%&BTSoo{WOgHU43`PP1r_gVjx{+OBIpDRG-vx)CT2~t3X&*za z;rK!mh{FCsjuoRo0u!7cu#Y&3fC-iYHXAPY3xz6FNF{glTc(W|JC}!7+0aQ%r(jEp zZUR}@Ov^}>2VcO>@X(3x3)x_Xb2#gIj}&DsRXSFDIP4xwlNtp?dW6TM=WN?ljTU)a zG?v~j7dKjGB+xkuEcsnT7AHPo@mO_jwO<$r@K(UxEh^{u2UY*>ETEYtQhZ78D6Qa75`Gaq|HG>N5{x8ubK3l5gv`9%j&+bP`(v)x5J zrGcOIK2=uy$}?nZX@yTnD=~qmnCIaOL%wYLehS~f-(P%c82eQ?z0kyLZ=O2I@Y0)J zQ}5umUs0*n!o1Ve68~EJLND`BGWU)En}tI??+qQ_qDb)iZ8Ekdvi*MSfqxLYxYP=6 zT(d!rti~}nb!yv#pEOh@!e1-NEL_K;vd6yHb?9Rx)i9dzP>Jb(uz)zDT_u%bGJg`C z@mrcit+y!5zqB4j27SE_EJkL?dUo0Q@4wHx)VG#rRgbec?2_&BY+EfBP9w^=T%Roj zH=o^Fp_uk^4|mM{@a_Y0uo@c`^yHOtP$V?4UJ`00(}>`e2s?2P5u^|BZhWDf!b)T1 z{DDv#cKR$(6YwBWK=p%}?UYP1#_5NqcDQ3&1bsw^y7FNtJVz(?rn{o!0980C-sa_l?I#XlDmy z&^HCysN6kw0_}!9|Mmr^Zfi^!I4XTk$a}z}v~!9pZBay-{$+Ur-zhN#mL6v<@T=sp zGA)-=XJk_YO-4hibjpN^WTJ&Iv!}sL4wh_3O?x3O<=U#bLAc)N>UYo*1YL`()wr`} z(7c9dsENYGU_a6LVM9Y>DM~#^b#b-o%|E8HYJfb=1_dHgSUlDfC9r~#rej%i7J7J~ zS-(vaqfpj9@kMPu9-x44$YiQfI%1lQvYSDZJD|_E-Ex@qEXl%VG^TKX|9GX^w$;2< zhyt60?3JaJW&~cCC7}#D6NMoJ93_z5G7YkF_Wq)(%rVQgfY%ply)`gIw)_PJ=@eE1MsJhqkGf zji*!!-<#dt(DPQSa{Ci-Q*#zJ{>A$%D53a9wHj@&|IKbB4ik0E1I&L48WK;yuU$92 zu)0gr7U+edxqLgHBs#=yghQ)gPW#W>f+`$R-)atC69yg)XmgqIuu^a+$!>9>V;t5l zW^?Fztdc1|Dngc4WZc^{n%qOTH1MGHD2oV0_HNEMBn}1VzN^{SU2`VBrzxUfcOq9E zCYx{y$nMWrGM+gjvU&tSuM)d(J6YQAeK&k*zP^Sv*FtF_?9dM@d9404)|j5QcWL1@ z`42d;u(DUYt&gT)YkvDcf1dC9RK;V5cMqDLD<`2j;RJ}BLut#|(qYONa{SKaOq6PP zlE@RK&jk}U?1nlLHVS%F0>5{rzyB(f6+3hT=Z3@K-xEzF!{VeV2T(ip+*w^{A1)mr zFz(ANgY@5U?;-R$Z8#X7N63<|*NVOZj|APu6InF+6lNBj?mi6WL~Rsd^BU1I<6+~4 zVabFF87?%+CQ#psnuF+MU%#zqOBJY*Xy=txN;ICEH*3iqtKg6{wEVmD*CY9N1N?j& z<@j_|oo(J5()_4k)!<;$$a9j-ZW~Li;_bd_6fD8Ak3Z;gNXu_L=ubKcoq853C6`QG z4YS{t*U+H;(>xt>PuNg?8VrIWqATgkqp)1>Ny9?bITbINWwMZ3|I#PFyU3we=E(Y4 zYGKR7yEQTouNvz@`NkURSD=uHrAC0<$Z(zKPgX|n-1XWn$>W<2p1zy?CtR5tehS6Tv-FT_&-Egju%w(D6kjw7xy+g0jKuE3o5cBxp7RQqOQh{!m zId|FNJKv_`fujT{3WKXoBMR@D+E}5v%vl%$I(&Dr@M648o+23oy`|MIuuombF}L>M zk?nA#>WSag&HPah@3iBrwN*49Ou~^J3%)^Ee zZ+?g}YVfG_8*K~|EPWY+*wslQCHv#vG;NfU>YCT9hz<2pl9O*z{02g-lqF_xxA98S!CVviN+Gaw63sqwt@4aTjZ7ZjuziH6yYjC zzNi6JIE3r>Enuzj+Z+j%0h<%E4m&=D0GKoRNE**2?A5Bi&|ZREVO2DLUxrVhTw*vZ zp*5OvB9sCRCO4j1Lo^y}dO@?K?>qYS4l&OKk_-3Gjqq#pZR3;)aqk4}a`|2~hHUsh zfsddeZIw}lTu=z@4c;ixB?>{K-gh;=rmnUPXP}vfux9lzbElSQyhpy^M{d}aJiC4Y>96}mt5Lg;>F0E$LEvoa z6=S~}gs@<>le3}ni%z%Jt;wTSHnN8X^Zxm(CL#NpN}(UD{}m}MY@ zhLX~4JINGQ=hf#43GDsDPk+WpCpx%aAV&}rd}|+qn9g-4vfR~Kr`fbnxfXpnheE$F&WM`#p_xpC)Mc6>F(z8-P44Qt_(qO( zce%bl7R!#lLe+?I~_2{%Pn{)hE0( zzk(*LKx)~BUOiSV`NS-rTC8lm7`_zc24wnvXGqI$nfA!+F4M_(C;slhpK2i&hpC7} z^ybA7C_K!)$&%&VoXOrix$(JWi{@hT=#)yv>riDJRL}7FRDQprf_oX%6zB8gtC!^- z6aKvQGRDsrIaC4BMM@!A*2i^?46kezA5pLIbli`yWlCnO2F9pKrq4KK-t)N$CEc=IA}!rcW7O2LYPTS_kPFbkgG3n`lH{3+wE92*uOZ0py2gH67E8* zEbLV^z_$TW2o3B!WjGvHb<&DgFD`2z^l@r64q?geM8iL4mam!| z|M#@fFao=&_vg_rZ@3p5R&wv=myEg{ysN30fIIe!OAtr8^l+Xl$tG!iX(sw9uyn9& zOW%fFQ)&7q3F&4Y-)RT`h16^obLh5+ztGFW-48~?Gw175Q%LY!P zx9W=zD8@idyo1TrCs|^I#&-BOtT<mDsXVD)%1gvC(A^1dSealr*v7~Q558I@J5?bJPhOYV;;I&4rfe*T;P zVdREE5k9IWp>IGChY~;57mRo@10sL%lL65umONvuR#GdYXwIAJsn8xx7Ud{8uQ1a~ zwg|)WzwFsntiP!y=cMj-K1oQ$AY~`ofnpE(P|V6Z|ulP}=Fpmb#A2H_J_s zfQ`1ViS+r_-4r7EjL9{UEXvtR=5-5Joh^JBPmKuUT|Xr;^l*}z?L<###PW{WthnTz zx9R0{pZ@f!&EHHIW2#j-JC_pvM+`_)b-!(RH=G_qAkV@QG)QpmtS$gI{cpd-zy{s# z|El-TKYbBxu6KohA2I(4Wcd3Dmh^jC02#TH9Q0Lv?5jz39b>0iIlB-IJ$gNA#!IJx zDlbR9biY z50Fc{zT?;%qc@Nl5#>f2mTe&hBh=1om$QTi7}9+k!zmXoXN{$0HXRfZKf6wI9w8vL z$(4XD!mHL-;P=aS2Zl^pmdc4Wh|4!(HYQ7Bm_1D) z4cs&d?P`;YVa(10&Dbucg>crnAzJ>F+O*%Tp%w2ruVft`UmF!gwGGakdzYxE7vWZF zYF1{+BC_^lh&ig7ZziTmmQ0%<1=y}R7&)TM<2yP}bP|;3iuy0Y^Gc1w7gY|}TTFJM zh2M&OhAA$7UwRZIe?SU7WQ^u8%Njbox981e>Ge_%l|Pk&mkr!c9l`1Dq#Cqn9%-2D zEz{@-1F_rZ!_m^((ZubotS7fehrR`Q?cdA3Tzxr#Xr+-<&&|6uciNZEdsER$DOxC^ zqmOv_x!a9sq?=LE+?y!WG83M+gtKvu);i_qry#N*!#47yrk=pM!S&P09h+2Ygk~DN zr96!^A!Ay0N*Ztg0K%ZeFw9ONiq#B?V|RW0`O2ZreP>dxd{%O(to$YzE`jb%EKMzz zO;0XaU_t}|AeSQZWr z1uPcNeF%4IteZEE>+N`R5~LBb{KagHx;Du+wqMBjcF5E5jWKiDcw&9mGJY4^!sdy z<>^R<ZAoTI?fb~S0c>*OzQ6U7%RgVlbv?a!X6{TrfDtqrC`coD_zCdJM@MzE9p9Q0$~Gmi|(6qGm4zNwrl~G086?1jU+iy7b$VH&<<} z@5vB-->W)Ie<~1k;9RAur1%JsjeHoPeM9x5B=7B;Vh$8(7KtuY&1o8E>nL)6;s!F& z@v|bKA{NOIP<;Se**E=y*YE@*=?`|Fyd*CI0<-H*>T6O^HQ<3tZ%*gx-}9=h5U7W$ z8x80dVsK!ca$c~xieXJrB@Lkm8nqY~1suGZ_=N&eP3c~Cjc-Du^C3^3HskiSi>Y&$ zHzGdZdAJ<}(eIrnZI`B`-@{QjgjH`w8^;eP_D8#}WU>lpWbO?2(aqYb=GL*Pv@oDb zvYeOxT*{fnzl$P|8MkWkFKNsh!tK`v*ndhETqwzdK=Bfs5*s)Dx%0H1X6mZWUq|XM zR}q@{R>C_7k$L+gNO|kdJ4wkkGjC1aAf(o?O-UhHgvf%Ff>8tR$M-#@i*-^wVwg5a z`ZyAEs{w_!oW3?+ZnAQk+FI4ABap&;+_Tqv`+_l`qf<>@QT~mw28QdI&5gA7?yA8W z&8VjS-ILg;@*2#@=27h}+Qw|cG4ID|%!?}7;B!oM-6m@yEJ$W4|DfjAA(>z+^&AE= z8$nJzM~~bR`F+~`*SezZnytnlnA`^~r0Ity2r`rI_fb}C`w4%4=KA`HTgC{NmOcqt z>!$h*BP*hQy0WtV_I7B$^V4FSEb+R*Hbn`BP_n^?O-BNoz4~3lXy)Juz6=GMG8$XE zxqnE3q?r;jkfJKjQ7%5x;k`E(>e__*W1__;Fn5||oK}5v0j~jd*Irv{#h%2GBooV_ zrnZ-2Ue#5c>eO8g>9nF3QCHAvXn4=Vp zk9`I0IfdTWGW%a=mBMV#izVZ7<$uw5_!7m^WTX6SdKjQWtLiaT)&HfO^=j8N+1LqW-s}dOw-mgOWKNDV?5Bljjx=+Io z&6a89zOw-3;LnOYAY{yX1Gjjs%IJwWm%mCganRrJNHtU}b2#L#2KRx{x$9sHnGJvQ zr{Bj8w~%T!L~Swq(;~+(KJ{n*T%?mETQBZl1aYbRov5Pii}#8#-mC4#E5{vCHA9=| zG_`V@c*Z9dDy_9+?nUHdO!a5OGP<>nht(u9!zjt`e8KG>miOt9;JZl=FfpFgzA*ha z;S(XN_XDqiJ-6aNL}y1BFR1s_@Z_-DL5eqDbb`hsi=tHv*_yYa1d^^bQ%xIuj|JTg zS1fP%0?{H|#_V(lNyN(E7`nL%=+$v0S36GrfYI{XYsT%#Pm_ey60>4dJqR#mbszH` z=G_qQsx_}b7E7~~F*HWww4XRxZ){!+pwqR$T8Ul;mE^n5EN;T$@`MT6OKxsbaTqOu z-iJ@k6}448bSY5S?!W(Fqpmttj0SvT1%m342ra%Qf(ywW?u&UEaLv ze0+-njkH*e=pO9&-$b=7R8~6(sZ@s_E`Ey`$m2oZ@4bIgUF?XSBEP;(VHfrXqRnrP zgyV_~=@=1uh8RLFkPH(8qz%z*YAI-=!X=zTk0iv*H zzY6Js(7n0^U_ulfSC7fZYte1Foa3D|^k-iJ4C&0F`JV7k zClg?a$VrX0G4EAa)m&ygjAC2EBy-btm^k8Wyc(}x!8b-nM~N;IOf~c!HXyoTmK|ji zLa9um=)RLm{ZmNiui`7*iG9DmMp_^!H#63* zVH6g)#0q)njDIoB?&-v&Qxq-b8_2-cAo}r}Jl)UtzE|{;+EId}bWWb7*3X$J1jU5` zDu~pDSji|wzjJ@wLSxhN8xLzgh`%3WBZ%k!JQ&~K2MY=Tu704)O-itqR12?9(#mS(fBsGS zt#8dcFUmO%jBh376Zm{SVm|h#kszbqP*KaA8qq5lFo=Kl1Dk*@JlKS4JTf*Nz)xeT z??exmQnt)qJ4aMI51f_4@BIzzT^BYzS8r9cN3stL(lYGdS(<$V5|25NOn;#8WBtLy z@G3`GrZ5H@8GzW+)oKhYPKVn?kTy;jRU|!FWvY_D8UJ^KTYfrpbkHGiu5fFwV5t8b zx2C-Ywu$ym36fjQFU6Lo<)tv(U&mP9O71xlX*eHjVJ~YKv7^z)3{CgS3{^`{WY5Bg z2tR#rU;{d%Udr~a54XNtvxdFNxAIxiNtBB53gc%bDyJ3~pRrXOPkY`2E)&O#_l}e6 zyKlKuF1dKS3j}t}iaY`3GR6CrHA7NT^~o{t;_w(Ij)FZyk0TG~29tE5var|{Iko6| zwUSjrSlYyApbZnZ_gfAnW%O+M7uQCz(WS}a{{;-f=BOHex+b8_s>&z!InMVBlZ>fi z9FSuPV!9`#dnl=$^uWJkiO+nH;`?vDq@5fB&pv8-LXfv5^P~YRwJ?f->BR2-VtgL) zKS0{bS8bycy!)H)A7`|$<*9o-;aqN4h7mK-@jK&f$2hkS2Ci-%SYZLJEK=(#pQu7;@UwC zg1k*UY6WA7>}4$09$hP&A>o9?F(WD#Q+zUZA(|{vmdNf_DX6>AGh3SL>sWL;zBcx{ z4VRa>E$yRz>|l6fT(DFf$v|b}T{}Ihi-+b55}lGD6Qm`~dOQ7#<1fOw3B@#k04QG8 z%ud^hGz1@z(Exyk@M+@t#dvZ(?fo5gPguZGwglj=!=b^@PzSPr?5!aUv*fhtSb)sb zZ?ia0oBlIfAjP@Cjs?!O#-I-S%x0%~RFzCxap!Z~a;qaS`QszKO2w4*RWKQCx;NcVO%TJU`*6Tzw~HMvHUuYSQze-5Qz6eB zzP7QYrl?2VysX_R#1BOpwos_?^DEw^dxDFJJM}-WsWi>qAeo;8{Sj^Wg#; z?88@{;=LT-a4#9QEf_p5Z1OUtrY+ZWjo(lIw>`lnYEoCCfmY#c5=NR!MBkc&I^Vaq zQnL=C*<7NgOuI6{>50x`%LlwOt9DDf0PjDLm~i!H+}VA5*fH9%ZIQ?P8;QMsb7$!+ z#p?>tw1K7lqE`(6aEC}aOvY;^JCjt$9iu_fwd`SfmL!%y%Gnj};5;bg=cABaJiSxT zMW8BY7wq@S^(J2b=KAz(wEbMdF_RZyGyvJreZ%s7W#^$3s+7mgMZ72Y^kPr`=zriT zm{G&OZYiMz|C(d`>Fobt6Hi_LEYaoZ=Z8*Gj z^f3Ihni!FWsx2>3N4%lXGWk`|~r2QlgWEdDf)L z_?Uy~B(mX~zf9t<@>1G+i_U)o{KlOpFcgAV?WTVI2!Qt^(C_7a`kT)f&z`=_51_4L zWN>ZLm>DOYT zRUg#wrOue85&gNC#BOwo7__TQ+(+T`U3>PLSzc^AMTW=gYbBPln6sT!183MH>c$o3 zSM889nmDyAaTra7`I&*52dB>D+1EH{1-wH>#JCYjx`u3_VeU}M!1 zSWO&=qv~;=AFA7Ll2mD>U3z=0dvdd{nYxoITk;X7IZg@f=-YXy1X^O!$fyqLq7wDXy_np&+w zW+JJ*(PJ#MT9xd*0}{fw&v?LaazN&+Kr;UE8Gv=dtG(W1EX zDTFocifz8GX9#u2O?&P7xB-QeMgyRI9hc1rGjJiDT`PQReeYL(v6e*@x;H%(Y2WjB*{6vVVYt6`x?IChsiBVb3tgf(i zy(m1MxK6Gpq@n@C%E_?Y%z1fdZF!${G;Tvj02T7CMiT62l_r265 z?fTamR3odFj%jDeUrLw(Z6--mEPkXYuwYk3Av%LfC6)(_Z`GG(<0=2^<9hrYsE7&W>-3&k$k#)iY#mbNp5iBZ)%(#LVg9p zWHV(jN9n=U@PEU985mdoEr+xgCh(A6l^u8|mF{NZY9d$Kt9#^>?#K!D3m$gr{s+LX zgpSRy<pvVa%KYkyVSchMQGPgbYCBzS z(-#)(-=_H3hnBj!p(;xU{nSv`+-me?Nsvhr-0Pjt=gt))W5SAU6_$_ty7YNt9StiI+Xx*(W^H$@z5aAVEfoOUs2I3pt+^L?yT}O8o0Jf1wo(>z%;ENS zYVTIRJ!BAo_jZ|-u&7@sW`vjhogj%{{}|wO5R4x#_~?B~CMKD?1HwDIn142wo$ZM@ z;jG?VH&*BCStx@qF30|w!>zGj|GvyN4%M7?MB@Ek+7;wR3BDneZiF!?8A#NcZt@f1 zHe7R5BS#pHa^2v-6ir7C2JV&ZspFdtjFD@jnb6S28-w$*YaWknAMHP7aZ``AwiNF> znHqgoO0<*)n5SxH4Jb>Yf4WKrN7Nz9ng6J$W##!E<6@CfV=;L|V3>wv_>Zf*SAwJP zVC+s?)kWFI8HiStv1+I_wtw4{LA03qZf(0mCBw)Y%u_BMfh)mU$6sVx+c?S0-+0K? z@O9NO2*|r0*Ydxa2Y}Vay;|qR?fwQ2l7Dw+B+p~ z@~I@#Whr>YG`yr!Wei;4aCcH9Cb9bt%1;@oTl@iMVzSrcd zdu*)hWZ$l$qi{&*bLQgUZxK>UjRWxWw6M~27vfB2!D1<-9MIgw(@c$qnQS=sJC zM7qoPhe?zB)1t|7hmmp|jpy)7;|M9jp5lwfznHsgZ(e^$Po z#=zis!1KmwC07sG7F+gLz0c5_hBlf~2Pi^jDe@(eKrW`5;A6Nz@10cAmsmC@`h?>@ zAhE-V*&&ckzRA_LLwlm9w>fXln|1pAPibXW@ejORx7TZxIJ(WBuqu>A#c6-4*Y5Yw zl-RVE8Mx{5;o#&s@aA=O%DcH`5+5LmTTC-W)sbF=%8AUPUeSqVO;LP?+D65mT=j5H7B2D^kl7cQk4JyEaW zpY6_r81{(Qg`P6o-Mm>mn5fY345s*Rd{6gg{ObefTJ@WXST^4Khl&Uw$1!V}qM{=+ zf2Z`3uZ-X#t;G!FQRsCo_i($R0yJ~hu45b;bIdN{gI4BCcWJk+ZAsVLl{B<>4o2b8 zhG_Lyexu3X)KBkELxexHer`@IW<5^X%iZyW&po5SeL~^OkA!z9(CH#N9a(**ZyjrY^MefbWGN<&cD-{h<`4qLr1JEm0{ zRte9dqys1Eu8MsZ^c8nQl=Wu$J%yT62e8cTL~CMY442o~MO+3aOnY^=+m71`cIuH^o^8N7LlmbB>)czkZAn2HeJ zHx#ZWb<1tiDWg#r4o(N4$hGSELOprHB$>Fn2(0N6a!;*?vqe{yl8qUukCb0?9GG$5 zhf7hXel+%~sdke29PA)}s2=~N(G0Xl^w2Qahs=@I1@q@{v)s5utpIxRy4y(#M>Sql z8T(V4pG22jE4rP^c{^^j4{fx<-s$3fNxdjM*I21;&GzlyeHy8hx7oaA2}_LFvLy+c zBjsw9Xu+7oAm!XcqPyUFWaP*Xh;n7#!3#|#x!c(?Up^Z5Rhdw^;j0I9IknQpw)>HV zr%w19v^w&e9=cjSJFZxx-)?o;95}f3X+RcUs`y=NXukXELJ_uQVn-XNU8Q}AiD=Z zbo(ANWY9%LGvZ$?0G#`)R|KMbHPJ(gUlMElDnA!Tdv5_-O6~GO-+cgeCmUxI8nF76zMjdUv&3`6*-+1);kt z_wUS|E~m4=q>ET*4%Vqcc}zucrI7O(4xT|PNQ#aZtSZj)NICr zN9}ea121L)VU9g$>$Mip2X<7rg?{~a8AY6Vyw=O;eXSMyK*4VE?NdQ~_5DwKM+JXzno(N<*r?Z1#P z`-tTKg_fZoIJz73!i6x&V3jwdc@!{xc)AIr(qG~d3MH$5{2dhUtyBGBiKEWjhK&Es zy(%gTp@|KMzFU$e-uHuUUD{*{{?X&pK%0Kc@ssp2`aYV(ZkrYaK_=7Bw!*%7vJ*VP zMT}=pxA^)QD2>?hNPFTuh^CTAH z*n2ZBZQIpRVF(5?yvMjK5C4=hT_*j``l2H((fIGknS9y}pG|(Mybe3f01!hkv*Md4 z8o<8*^j-rlmHmHVy034a!i2w#{A}`X4c6YsJbt73$Sjo~Fp=z+h}&OFMuztB?oG1j z4B*yH7sB$A5-Ui?n4rG&=oee|&yuX_f9Ge;FvCsz?@6en3EYnZXjR7mDb9Jy)m@qowzQU3*& z**-k~ul`aiz(}qML2gzSZLAxywzh)h<8|=Z)jMHSXo?!yw-Gqt!Yd9j^q*E=OdQnS>xVE6r zZ1l@Tjr>7kPYkjsmci-XiB{woS{MU_{&tSTDjr3Ba~Ts41>$+!aNNxgq(<|+wO?sjj8 z<&blY=PHcE`B%KWOK2^hL_vuQYparxAxTL|uNyB7!4~rPP=*rj=^lXgr?45cQ!mXK!tpLVz z|FZS9ZV!2K^nKO=dUAkgJfoz21!mt@?ppaROrN6Kwt!-#NfH~-hRj-7bLkCH&zpqe z*Q+Ysh1t%uT(&V$W_}1!6G56xvz{@5IXYCoBEfX}Z=p9SK$8KZ8O3rWq?SprVsZaO zCrUQAW|tN5t?7X7S#zJZ0W|cG+>;o^Ir2ANR@4HRi696hu!*vQiI(V+as25$T|YPG z9Q2=bD2EAsNT6EJ`}xT=re%KHAq^&we?W5nDgB53?@X%a+?Ahbp1C<}sR&czC zo{>)_Ryp^E`Iu_M)bVEEMtc@asuc!zlN4fE4ah}0Y)P2eRr!dIWr9OjoJc0@<6d<0 zhf-&JrQp#lkwM0NNS^@0$FRqPk?}A-Vl0U3>TcA;5pc2OMFq4WeQC6^&ObkBir{+R zm&~IyDdUeDP>T~4e?8(&^5xueF|TTgQr%zxee?i0CDMK}%agYA9(a>U!UC%NR53dB z3q~5b87egpTy8#9QT&Hu&c$c)0cNxH>tEUd^AqTwByKz&)XDH`XgEH)+fv#hOa1ZJ zB#GE>q4z_*&F%sY>=Z4(ykxg^lLbq!B&!vJhXibJzXYw&0dCv=r#Ab&T0jGdYy&0G z_eR-$?})%#iLhUB#FBc(I7MVhNUDAwCh{64-)05Kx(O`F`(#$oT6+se-7PJcF89k* z)i!Bj&jasgxbg=`dl@k72spR6P3qEcd0T$mT?WvRlc${9V+~c;)dK*;exJ^G#8c36 zE2@pmWU_CgH>UD)Mht=0{ln@s4KQi>0#HD0S>RM8jB~-WN(j0eZ@kTN3x3~i@Z8+P zJJ+FIUvH~TfG~^AmQhs`>}6;(@S}*?xohW6+x~iQImMC>zy`B}+#G;vM8&Hg2anY4 zt;?ijF>1Co@g4%dp^#3_gcPybYgK zK5UzVh2`HWyH%kcegf4-%*k|vQ7Pk2Flz{~Ih=Wg6bjLSlap9HMi5+xht_k;qc&1_ zcz_X+6_tR4U51**FzXH`g{+EJOHQ5*c)XdVn|eL#tD~ukwzm-#ecoIoC3b_f`CO!F zB;2oEmqT7VqcL8*s1QE4Oj?hKXu?6=jBFhJOPkq?7k+)j2<+`=f@#_sioZSS46Ka) zr*J&&Q^k$Il0&75o-+Anl{McNZ5YRH3S-AK(Z)rhY2j*+#E;vfP-aVb8$F0{Y}TvQ z{apD6Nz-PK>0pj|Kz_jB`ek^e$DXEC?-b!0CJ_3Y8~XvMFj{ETn>)|9CF`^I2YY2oLKJu=2O;Y?Z-RFhg( zpEQ4cXVj@|ja?X?uOF?QF!d~&aVber9HO()Y;K*-^preYaHAS~R~sn7mOv*IGA94o z{x#?8P%`^bL`hYIpIkdj8Wj5S+ zywR4Qopbd0=6w>d3zV%PHncXdC2I^euNH{D2nWg?EMM%r61T2% zb(Ipnsn5v3=WnkdMMkSAY%Hrjt zbhiz)lk^PTm4-`jIAq}%60p1sx|V@0e>P(kx(vogKKFrWMU8c(g`6Khw);Mwp^*_* z-xdsk)B|<6gv4)qm-%rYp%4h7WBY;6$CIO|{Pnq-7#{>UsL0NObt2>#N z_5$^f-<21?3UVl;WVo$;J-f~DRCs)V2R5#V?jdwt-*>Mx2@AJ}(6_E;PYJ0xO4T+c z&?NBA$LF8;lMAZaclz&KTHc@bJ14i_>=%mR3GTE3Kqb(l*~i?P*9Ge-y#(VrkgjDA zSknbw6dP^_wGeOY@v>#TZ&|o~f(Sc*f#v0)?*blQ5&nYjr@f6yUl(zGRR^4s_*U<} zf(+jxw%-s-$Gr4!X4%8(!D35Af{=~rWr+4Iw`a*Ip<9)1gfF=ktKE|0qz^kA2MuSw z55V`_cV!vq4VZ~(xuCEd-K*dS-#Dm=SJWK>-JDEkf8Ld@-q``o6VJ2M?&%ZRuLE{` z6j@i@z~{FR`){xYNLe0e+n#Z&IHU3+c-$W<&m^Pi=Wlqt7Chd1YwO}oure#PN$QRy zn^m|1_&iHx)zr{rXlv`Y(%`7(WoxJZnCa^y+rF8`z5V0u{msr!K7;TqBmnerTAR4U z;$T{($Uoee=wXN;qu$4N?@xsc;Gc3AS4dO1sf6>`$^fs)Y4pka$IJSLyCU+1u^Sa( z-UlehEw`T!c3k*lK?X*8)aX5YC)41m=jEZm{KNfCUNc4NgWP6dMeF|d%404g;7+D{ zC!-<+VlfALFg_)5$}{n8G(3F(pQS;ioE0SK$SMw*WDQpy`h^~SR(1roF6#Inbu--f zIoWP62*FqfVASf=N)~Yk|9h#AeCi_m;bgue-&|qz;5W}ciwG)jYH@r0I+kg9z?l7` z`B*)m_|0>FFZEblT(xK6J=+}TXnS8M{^HT6+tK%3DzW-bpb`iv9M$6dD8tdG+5OkD zw`V)V^Bz{W;@x0drjwW`_sCe=7NqH@6#(ux{4rE4r#bGWYwBg|I`DYsb6?}PfKLV+Ch@{;`$ z5b5@U-3wEW&IqM&@NlKTTm;p{>8CIB=DTlG%SFl5NpNqQH#m0xxzHW#k)N3|{l*$s zac9Gw6Sjp{M ze9YDf2H#6b!DGUEPgOiNf!Ls%n4Ha2LhV_8^te8)RwAm8KI-11US5EfyeyhJMko3C z;~Wk!$Ld4~{8mAHoB37teQtpvr*L&Z0wyL=8VLfns;dAU!MMO7|HIrRyU+4Oqz30@ z0cKv$M|_F77VzQY1%vPHPK6?X9?|#o=A3F8yZ;YoKeg<4@blGh?o#G#+3B12r_MY+ zfxYF&0z)4caMBRVOF`NB(6v`PMAr+~uey(LqLbf|UKLm6L*d{ANFF_G@v#_l{ea=Y(v=5qinPilKRyGP^dV{>)KKRe0e z%GL(hNST`O?>?e$$rQ&Rb+2X_ocOrd20a|zNaEfSVbmL&KZGyOfSK6QgAhH)qPyUj*wCvwB;1b&0%S?RBE=$&}HA-qY2RU5M0TZooguB)>SK) zLg6*t-^3?ucl&hovBd&f4Z`JK6It+QDR0`g_4PX5+g0;8iRFq3%#SA0rmLv9y&FP@ z=-Gn7kIh9?b+Mj{G7a}iJ#Pi$&lo6Mj&AmjwAps=HV*VxAjx@;4&cFfP?5Co)=`06 zi^|`5xA_RmvV1k1Z!rh8x9PYn5@5q!`}MYaA6)@fYL7zUcXSg|=w-!M7Hcb-Pltcu zHM84T9|&wlFWNL81q>XHcwWA`zro(wb|UiFnL&GOvzdKMfCOuq=KVzy5r!E&FL zgM9$O;mGX`+6W&Y{y7 zzvHK<2#APCM-fpGu+Xariii}YNsWl~-U*OIK|rYj0veihDWUfk0qMQh5ITgCK!5}i zlH4TUIroe=-WzwE_wG33p7-9(AFv@iD{HU4_L{%>`^`1yBD6SlrvZ=S){Tpw0oz*b z)HB}A4EhNg?l{s>H^ZKYivt@;h2r=#(x14ATfc zyAtgO^oMWPaGW$t>e+76)`OqYaS@ay$%94qSy%HPT@tbCRb=n1$cs5XnBYII0SRDE(PnW~(M(zzRQ1RrZxOdLE@v zXMnD2ksijA@8hctL<4j(TUzF|(wiT7x3rjBudZ8sFUa?;5;r|IJmlYO-Ub;FrC`aB z*|4hKOkQE&7o3X;Ui;7#5?Gi6a>bYG2`9gJv`yOFpHvRq38nZPblZFb^>58LEV5K_ z!_S{xvG~q&LY7c9B4)BsR;vvFfrv@cHfgk=rVwe|RG>&QHPr47QQvk3IReC-PILlfQXQ@h*eG}F?Rc3c?;oQ&w%8Jd$ln@PA?>Kx8e?Au}RRv;kIrbHn6tQ5m^w z?`S)O$>dMczIT_4i=3N(7;%_h;}p(c(}*mApkf#dgeEb_;oal)UY|-Zu*d$xSn&&p ziT<0*UJ`-h*gR)|1Z1!!Xhkwz=GU*ypcKSjN!;X#4s7qDOWzP>q!3&d7-)GZ@NmwN za&i>iqypdF{g|UH+IU~t4tXuAn37oq#N?ZofH}-{^$q+~>{~*FNW>ikxB280Hcz~{ z4YU%5gHt9c1Z2Z&&*_)x_ZL?wC%p^ePdd?dPV1C&EVsLQ`FUtlB`@XBLapwSg_jx@`_OuEhoFkl=A&F^@8s$; z$mwQrUywz`*8W1DxgV{?hyT5@+^Mhb|H>g#zajG9GN4icExo>wfTwo{eC7(N;L&I4 z4*Z$-#h}y%xC!0>zK?&FjH}n9wUt8O^6+?#K!`A?={IKoio!XY;0OMA)5Y|ZXa6vh z)Ds_f{5x;>@Sk)xw9o$!J3u?(l)&=8hkA1UvsZ!k?=EVBG1_nH>;Fm9A1cf#BN1Un z9IIy=OdX`0o#(h|P5lO(qHmV2FoToZL?AddW9{Ak#I=r&jwT|Vj+G()Ysm`g$^X&b zxEZ3P7xJuglE}5cHe2UT$SspQ0B6F|zO`u$*&g5AW^}j7ua7sUjuZ^iih_WTE6h_B zeJ_*k+o!lM5ExqKy}d<{*yBuCvV+gI(!_5Z+9i-Q3X4*qCW;F~$~Af&g~+zd<2p9hd!?MsC%7;d3B`cD z7|jwM2|Klx4@)U25pkzyNe0VC!v;}Ny}sLo`lRs~p5M*noYTE7noJ9=5d@4mr36DF z**55`ZO}&9ZIccoDmtE??o^O=9vPb}QWa-sg32rj;{k``0h&l6(Hjt0r>t070yJto zf)?hj{oS&g`4$aVfT+FC0HAgqE#PZ$CjwKm*v^_nKC&TY(O|Ms3HWEIE&9?Grc@JI z+!tJECpAZKQVs8`0cA>1xqOFV;@{?lKCaGUjcavmo6kH>wRO%txGiFbsVUB(ko1b{V7o(noe9yk zRwsjrqQLcNgLDY*feQSun)!p?q+X9`zp_$0Gh?^Tqqi6%ZVI@CA53`~^Sy)h^!Eq$ zJ-koX`TPpEwsA4ExxGL;skD9n^f%TR(czqEzv z^fm2<-htod-Ni9!ayL9;N^Ou*Z1fx2O}?u+AnJbvO?QS87qLfVQVKx#aBf8(dVRz0 zh^QP`Pu5j|Q*xGF0$fLa>TmztDtzi~N*N=JPZ9}$gJqQ(1G9$`z8h>56o)GEIEg6D zJC3>DbyqAI>;G!;LveIex_Yua4mWs*M`@#VIJrEyp&`=lB z77>TBVIp-)>OfeVvRsVB6*Z00E{@mVk9I+2|Rf3>*j&(3HGCQsHNy?;s&?tG7 zEoVLYT0s9YHB)!wV0}Cz5n3ZwM;Uy|2==g~w3ZE;k3T$w=LB3o;P;|P#&_gEnJeKn zM{PBxED0OpTt^k_#m{p%anC4$owNKCJSlCN=uu=zA=GoLO3%~NHCX|rGi-7Bl+e?V zZWtMt%EH0pw{^lF6M%^>EG!RVU3T@Cmyxa-^i?X7JwpF>$7ty^2cX9uDCf@*7Sp4l zeK%xkN;l_FLX9>j8+t|`!u~19g zb;mj=bih1{TFOI+>#W&*yQ9Uu4^8RHcuU`j>omwrMp3~wxjQ63$<%Nj?zOd;L)d0K zi`ZS(M~3FY3X!wkUTvR1{ppjUnyYN(FCs4N$2fS`c&a;V#lUbo&PVTj@tX47TMqtS zdQ%*(CpVf@3~Oa-CI^7XllmN)zf9t7etYy`taZR#w7*H};#5sdog&}g>t`3|A_#cKr7--H zH`F0k!I*n$kn5iMMry$i`qyJ9mew)Gd{}WF{S=3}q4BvHvUf??)a>vj%mt5%vlnZ3 zGr=k`mfI(rg;@;(UL9}^zHEp?``{2Y_18aKKzdUn@sR9GOr^phpG%k@BO5b^c^hbp zOzwuPLLYCfdp(={%|B})@WTM&ka7mqHBr6mA}Vrspt~D`q@0k66wdDmW#0E)atA%E z&C^wPx#%L(LKq!IyssxOGVbfy2MfTr*R)~~fhXRhtH*79Cg7c3*J7_2%`dxwHa(u)rG1SZ8?IORoUZG64NgPY8W^=#(@0wQ>~nQ-t~ zqH-^;8nfLb-#WjBYf^b-IHW8LXW``%4^X}zy<#o*>2>%gWehuyOH{5><2HdieC~vkZ`H;B8?S>?#+{;rgCiCCl>AY=|Ck+~j{|gw zI-SVL+#!V;X#xEq0ioF``SE#ktf2K(L8!#`RwQ)^IJMWG=}Yc)PWGYVI`ZOvK)=Yr zmM^&uL5ZH71qLx+AQcvL6&KWDcBt8UAYEDKDf<8~;YiXTiEkMC7CE^W?WMgmw-hwDAAHfx zWnk#9rc?*`+~Vm0Ezv)-$uazP8&vKf)2ksISD&Z66KnH)6Y{F%L$g#4?MpxK)5}MG zMalWr6wQn$m~Bv&S&b~*E*l(eyP=jRBTIT-h=C6YVGzH-Jq8xno##7w8@7Yyul$EC z9nq0eE-!3vJ^PlxH27JLa>gjp!p*Ca#^pkw+IP0+$45%>welOjE?~;o4~76d7NoMC`ByI-TTf2Rx%b1AnQi&WU^rtzC!%swnq}PV!N%jO8OYsm$29KP zyE7OUGY?9^AXU)*JGG%ki5!-|trtR`n=Ss~(RPO6E{W(P4yf`lzB}TEuGOH}4z{zaIv0hxTO@Zh3hZcT7JaJFV#7AoE*GW%K{wrh`9KwSvP8KO?#^=_$SbELLtF^3;*As*iZ}4&Q^VB%fPs(~Xo|;6 z?GzSYairZ80Z$Z-$VuUq^OaUoz^A*Cb_aW#U*udMtgkRP?Y52?`9c?yaZyyj?V{)9 zdg*@5)*%1yR-+Qg(PBV@e_wy8FR7k|ZyvhYasQ{+i03l1l=r#jxw6d(3SkBW znJ<5EXfyF)jL!yf=hqd$v&HsK~H^*?X% z{Y?@-vPb>YWPEK~U!1e5lP3|mvPwYx%yj2i*`DJ?8~~T=%>(35A6cGYD2H$|zIg2Y z?NQ;Ee#gp`her`ZW%69J$I7jzFJQ3@4Ajy8jk2N@MqSt*ts_QI9n&DpaI?56H`2xV zBG@g!mt?Pp6*&@%OL7DvIC|;NxJ>eY(B13eA^Lz1+dJH3rVhskKy-ieM3RjbPTO^j z`rsV7mWZCC{&9-#(a{eC=y6l@!~d}JMnC(1m6!TYum4jV;QuO3F9aM=BA%T-C1Z!A zWBH%1G&_ByMCD58UHsimzoqo!e|UpDBtMzFcuHnL2SOO7oKVqxd9N4|PA;K{oEjnD zilcvyCyBQCJ`b7D7RK=VFMzZZxxp{6Z=oe8Cx<>$hNS73vSj`r`7~7M_H#FL+Bs; z{eTWqQly?feG1u^hOqH`zQ=<++FPbWn1yR}Qe8W*%&PNjXX>k^Sywn|I0o#s7{Za5 zkU%ec_HQj{2cKmk?8QKMcp5a?sNIYY+@gLZJvGkFsQ@BjWbc7LCNiEC`j6mci)fmH`Rp7*yq-4jg#XQ{_15!%tXi_{u+QorLA4e8uKGW7MAB!&{ z7->?Bw)Q}&WgFe^T~MQ6a5Kkg2IhNr@5)i%W*(FK+{AmvLKY0jNiD{ZrHqG%8O0=qOBN zPb+hu!hK4`&3?7Cx`ZrRcZoSH0unf!7F~VnTo8m~KyjTXou0fnnx+(d7{&Bb>PSc* zrwbn{do-~P9MxlDOdl#K_t$20ad7%5M{Cheq@Fh{{w;z1!l{M(sR8&v;!e#!y{6q~ zYL}v}m>oz{lumtUlNc+RpG=?wOYMJ6hDp7=|4B81Hp>4=LdX9fn(jK4+RqOu&6eW2 zV)(Ki9O&-jRxK+-ex+B!Z|>%HlB%_pI9}!TcU@bD`Gb4ZsYIPj&Gb2~ZLU6ApHQfD z?6+eZrPV+D#^{nYD`!=u`Uz1EYh&)bWwtb2NL-H z(DGyG8&c4^O-T&Hv2>C?Q}4b4t!1~%be>6Xc=}*?U#F?jO_Ru4ato)HtgH!!8CSh0 zN2qis;0FTVZkDD9?Ei5DfN^l@mZB@#^^{n9PT424HjW&H>8@+#K(6w@NyD^=6#~AR z7|kW?^xW9q9&B+bM#=>SGlkpq4yav{5KZ4VJN_&yZT%MhQO$Hel;9i?bcBmmqqTJE zJ$)ftxLE}YATLsNTJbg2$dI%!ny`FTRO=J{;r|f-PiH?3Ke(E-2>6H&WRI3Cd z^1Qh>;{&44ECyv-b|Vb+4fwbgEdz=vR{cFYys|{!5mrRK&9S=UW(~$pb44rV!G3Oe z%FjVLt7IeCn|dErE5{qap`P%9;pKv}d`+O)321w|gx9CA=yZK^KH1Hkqcs)|(A+g0 zizd@zZs2ZZTvGiSiAhVvZ|YCJQ9@{N^xBL;VSJ}Uc*upY$RZwZ)X{KzC#Qq@a(#m1 z>o~{q&uaZW^E)Rr(KsGe@oQo6VP8O_VTEt2-+1bygFa{^q)@C9>9Bb@OTphiFsalWt3O zGX;E9mAQ(0X1PMyLOaL&FGuCw2_UfSKsRWrQwOm>w3Mp8?b6%v7N~`_wd>0Mp)(g? z(A|=8D^rA{anGgqBhthWJmfKD6YTlS{K59`i(IMB+-rmW4g)i|Hp4e?kG5Lcfg8hyHFg80={Y{HEqFkr5Mkx1G?`tC?3-!l2)Bo;ZRX=EEKJ$tu|H#3AEklc z`5&{Ygbz%G0!#W_F9u;7%h+PH*}XcX;L7+YQS+9s3k4$TYrN;~+%%bN+M8&#`#P@g zXRooGDQr6_!N+5U*o6A(aqMKMV~WKTb3_aR^h0Ly7 zHGFR>Zf336ZLhuUMFgYZ`k6A^5ZT}N2A^iqAHiUCBQrA->bw-V;)-alb0Liw%r4I+ z_MLcN@<+9t4?&28UFg&y0!v};2K~^971yWB!T=M$H({YYT^}V>53+@X>Q$NgoL$W932zCBopT$SH@voK zMG%~?+~`)9lDah@!s%^%Sj!6hvXd@mTIQq~aA{U%eKIBl;O}pOLP6WJMCvzJi4!i7 zXXbl(p(i~eozeV~GGmqQs|7I= zMIdLJ+0Ttf8oY|byJV|JHYGaFSNIg%vDrP?lSTnPXEVTlOL)h`IGnVq7*%1TETKfN zxVJm`f@*S4ZB+R6u1yq=J$dhD-SUHU5O`8C<7c>PF0SNB9A=aLq**s zG-hGOtPeLF4g9C z%1KuWL-79ikp+*=9C1us^vp z&@XT9{4a!i3h4PC(us$Om(H;yKZFrJOQ_c9E*_yC758b0Nfr~Vtkj2mQ=>%IieC&$ zJXl#1#?&o+&KJTKpaTxjJ!2@0@%}(2!NS+{J}QGjTd4nQq5-(`^rL6|^FIU>4(b{B zU!*?|+zARX)h=dh-TS0IhyP<~ADX?rwAtWtM@zJ7a-xdI$77G+d$mY|R~Al)4Jg!P zhLd8?`zk(Mjw||!kux=FRuY%=)i@dn<4d&wzJGtcVI`cC<>d>>vCp7#Hv>;wnx**- zxI+WlKL_8zc)uJxxqVq~^keAB`AI|eG%wqruAx`KdV~MT%u5iDdaYgRx`*O}s zHcm0 zNUKO*lsL+|p%I(E91*YFz_n?j4G5%x{0|k8boow+kvGNW2H(xCqdfbiV5RdceCC&B z(Oh5b5bBynE>3n)oy%p_(Z!^a4yfLj;6G`03da>Ol6bXBcU!qUU0q{7)atzo!PTtr zpsMsUDqGYfO1dU)E3e$b?%TK)}*NES#@zCW_{N>%pEnb?JjHxB)3wu5%+H{(1vNg}Nw$4!$M z)ie2O>I$E0s)QX+Lh`aY(%F~~tCNc?g827v)dy$#MEWs@;qSxtBKy}Q3l7i zrAbzRmHa2%C$;9kJmN{EJp%$(bY}~cz}qeUa$?LhSYV=DF!5KPZ1AWyk8OCwhWg+M zA;h}uz=XPB#@yvLdHvPsD|h#KTZS*y37*vWgo^lckFt3#Du0Uho;vO7<$bnko7iCX z-73xQwj;%bZSN0f$)50KqV2ZK+K-T+`0pM!STXy&vfZY11{-kcpBd^A{wyiYXtZAV zxf8&cKiTm0r4*$+lUi+wd+F)gkKP;1?>E@GY|3(|O4JL;*k`!iHT}2RMi(rN)YCT{OC5z%r~LP`s6T&}TnIIneQ5q0sn_6l7(Ea_gL|wOJkN+^$h$71 zky_U8n5@Uxu4nra{ln(@IFj2pEl~f4*_*G$<-$Hv&jJu~b+d_RjXT&Ii=(q5*I5zs zTMy)yVOl7$E3vI7DZybEci;t0gTG?!F!L)7nv(0!kw?`NYZk70ncj|-gB)@^UsiI! z)tITP@^xV}FX#jO7L7^$t#O-2^02@Q5UPe~lr1#T(Tze!KAmMYsKZM?dj2P~?9Z?x zwPdUR3oWjty=Z;5Al=%ZNzG??MKvd@59+7ggHL5NVp~vZZ3B{p)xLi>ZR3{Fh)@Qj z#1u488afG?F|`petjWACIZ^bHJ(xu*q>DLqQjz7U>GD8U4a&~KXqGF5ultc`{|j>D zBy65Xk+9>p+!Hn2j27v-(^M$|0&OM1TsP@{+fO5UkGs3_q2S3?)H^R?kFicu{n~fl zQ*yIXeV~<(w#*@Xt-#f%`7Z*ODsmhHa{rdA7KJuA=(yc&T};u>P#mt$C|iGXnJEHb zJ~g*xn2~}$Qf;hC@_Kq{#o#4WH^t<$Sb9x){`!UCDn^IuoU1poZbgYHAU=lv%vnqI zR&4u=*Oa5NLW$k@T!L27ZJ9B)8)jzEfuuj1h6jC|Ue(2cp0D)Wpzk>%o6GNc7wnDi zC%SI2zDGrZAG*40B7Lu}L^Xo-!48(Qjk3MTF`N;oNME}J>fJc?uc$>vHiQjB+A$Zb z&pPj_;706+gTeDX&dy;M2EW*m)IHM)4O=Ckw*xa1(QU1#THI}@!F>P31;k>@+D4kr zYyyCr-O`ggQ~VCkE64kjk9SY>?lsw}hemNzFAfHF!8OjD9_90@b~Ytot?*5?d)LhQ zTSHcZ8*~=OIRNG+$e9YaG(9O&w3mxuGxB5(4$r5r~5qti&n%|$)8bBgP z3b<@lT;X^vv^p5>D>2yz-i-44l=&Vj%EpX2AFC&kn|ZsbKm13GB0vxF_H5YLtr%s` zaL!nZkh9SZIdZ&dzaZ8Vu1UNvu6tF(0uIZPjMC-;ugCVxLX?!eLHME{putF6%G8kH z;z%`YEd&67?3fQuruHb@q>a{z^#-lNcuLA{>1_IUHsE50u}G$h(dZ0xJs!|hBIEbQ z{phXy$L-c$bR$S(M}(WLWgwwMp&Xk~^xR!VyRqud1pBA6M#%3JR70UvOZNq))WG#Y z<-nsTXY;bElK$GgD&3p#`#cVgE*khpwMT^Rj2wO^h5EY|GA<_Gdz;7oDvd6SeE>0q zUF-rU9OCY8f4k0lv9^J~-iN5lUwN}yG;pKiY3$Ncqwd$wl6lXPX&Cz(e<$Q zo}YovD0Q-wYZj*+Hxvu=gDtJM(g-XzI+TSFvPp9i7H!?A4uEurgZPSaWnH8rlMre2YV!U`7IDB;-g z%gxQf;~9C=7f$S7X+bOosX5u|$@O2V%l2PWR({ZQBd`c&bK)`)*)90uM(32X&->Qk z3IJ$@u-Cqx6IS2kK8y2$UN3ahKXkQ#Dtf-4QUQdJ8wPL!Vw-b0{K=P8t=TvRM7qsq z{kfAl44TK&OZ{_eFL=K}y%@gINP>67=cUixtie`yt}8sOlK94+q6Eys1cGt>2`G`o z&s8_YpIhw@sQt1#QLIRZ7vA-VLq(^IL2h?M(3;06d>RpfXhC8s22F+FwVR@MHX6kmIG{ zjvb~b(cw3arlvQ}v0vl`p*F&oQkI>DIcLkP9YI!_>bD_w?OoWoE_;o$Ue%J5ldjHo z^K5r-a~Udm*e%*SrVp_qOiXL4?p$4AXAvv8CVO94=Cs^w_4ov6Vhq}K=bD*`lAoRB zky`Lvbsf24MIe65$nYCGI41ZdVz2*jHtUP76_X>m*iU5urAymV6>zhSts?4gb9M5<88<)OQi@Z{1gy}xY1@qiV|(9RaT&yP_q9& z!Y*ix&cM*#?YjT}2@&qYsL*_=)9?tvPxrNV?@BCiy5+}cb(PJYFca@$#{OY@D=l+JU72#iOz+vW?iPwH`bsCoL53 z<+;v!IT8JiJM3=gzi~hQG~(t~EzbnerH|}c+Ss{4y*i{J_wiBn#Xn`fyRC9ENqBWs zi}``Tpc}Ub>%86pSH4B_xSB{}qS|*W&1D&GU_0OT#&;1Gvt0 zuM>I>UpKCk+te<7i2q8pdc+gcedL?Q48e0}Y?C*7WFL|D&6Oz)wvN!tbccrXWf=7l zkRgV6h9$CSd!ofUF)(FQxs-oR{+_tZ>`ghV zS!vhlv~!o`m%8IOcHEW!kHaw0E7rJ_kyL!HIi9wZ`|aRPY5%J5-sx3*ZZoSJ&EoBofb91##`E^yItUVDc z9Er^NJd!?M_BV+OFPu#suR=zKSo!3@uf2iUohn)0{8WCc?D_!cQ$V#Ytt_ZV^BGCrU*EXi z$(<-_V#3e(q99kR_->zerDU-wVR3V2Z9=W~S9xZEhuL`Pqx-O@w)>s@_wu98hOoyM z^8->4oU@G%?ITs(T=7}ApEE6A3)Vr|&6MUA6Q!y;rs2U+3%lYHw2#v zrIx$}GpDia=Ez=Fgic^LZ4Qu4Xq4J-4GlG;WrsX>@ZM}8rD?l-j9V!2PO5o2JZ4>>WfN~MqG+;{kD|Fgu^mEadDP4IQGf0 zx%glm=sEsbY8i~tA{~%^uE$d>^$l2U$P!)DxmM(isatNB>~8qAK?pXRq^$?vY7Px^^A7;7(Ztk^S2ou68aK){HMt%f#-$Q`>>@d{(EMeEDqJH z)JlThuF?0pT>U4cPqi?w;0U4)`53QzdfsR*Xg9F(_tU5oQY_ zt(*l;JUX)zIfns30*)KC<<9}=x3*iIB?`oq2JYpS z#ZHrl%I9umf%jr!UKVbPB1VP}I@kFFGHgtZ2BqV4%vZ3sV9LfP(qEZs2W@#^B}ea* z&k0I}<_R;9OVs^S>U^mW$hrdp!@rzOMCPIMv>zwr-|Xi-25$yz8$q{Wq_e2T8GV;H$;EW~aU z3HSK(`{(NDsz~BEjJ#u2)}w2LNv87cYm+he<`PZr@fd(ewdJkPWjwglJ483x&h^Xn!Uzx ze(wVMivoQ9fbl0g)=l0HQ%aOFA;)U(kPBkP_r%363T0}UI@WBz{Cmf7-XB*t!a;Fg z`tiL~;NK}k4R3*Ku}M>8R5^QZbtOYtlqZ<<<|rpGFM-eEqxK&w)NAkeyBpyVdi)3C z=h{tcT#HT9<&e&rHe|VTO$Q5~T2B{p7#hU%n%C7TTGXtJPKf04^73v>SyTiDI9PT~ zZC)3?XVNJRKi-WrQqn|1_g{3HgU5fx0>8lc#ZAjRGy^h$?XHlzCU@L+*fVu)&~L=d zgd#+s{}P`E%79Ir`>KXS5y=jP9dxi!WN%vWi+N^B+Uze|5tQ@`eA@D5|6*!CBVXNm zjUzrgQT1t736Fep=I!B1@G4fv!L;F>BOrONcSooBV8I|n-K-{>aE66MSe-er4~bBZ zCW@ew5a#|{){g5s?)YkeNft(s@`2agOX4M9~Mo+Wp}$V zn_Icay6tHqKj78umSHMP6A*Yk)ut=`sH^CQUO3BDt&(~dhr{F;1iq)^XvQf0`0byT zXvX~W0*UaG_f%BP-+lnH`s zWrGucJUWs4fBKS{+j(au`hM*FZq5A(T9p^!>Oed^2yh-GcI-|YEXh68n&SZMBqP)W0L6&4+A zdnG@`1K7P3qV*+;(MZxL=_1PgiL-Zvl8Y9(ub|Q8b@!LM6Eq4dfXd3Ai~7Dd(t>8q z&yx0Hk+CyATAC~D!4&q%uKp$5t_&irhERdKAu38o0peDZPbuj$psdW8wZi2c8D0g{4`yQ5f?yxu^siz;k)1YnA~E(8@5@Eqg( zPKU^9bkyXi`2nLC1dHOdiYEN_^_l-_DI|RRPe-!J%xh(an6x)OUO>sN=!^Fh-0w)& zJ6Pb?j~IoOkA}R}^|w5G@<_|4t@Ik62TEc3nVg%PEj)IDSrA@o`=O1)HGQ=YrK^Uk znWD;+q`sV^n)N=A<-nuCfgcS{rs{6x7 z)XO#CXJy{FdY4DJDPSe4kUZr#CEy|ULm#AJxOvEuG={%sX1fJ#-+Il)^Qo{vJ+(Al zr3jaL<}hC~?%s;B-#%*yZtR1t_JEG4RLS>-g2RT3wH2uNYqAgEj^7NE}<}E4ga%KwD5fsP!|EugSA-QN2g3_AW45btJFJp=CsM z^!`n*t8l0`<$VJ3kVdJiuA5SUfX{W;gE7OS4@fo^*l|HToo18{UODX-Fy$|KZ7gkm363eQFx2KW zo#{2)>##1=<7Iv#=+azV1LQ1^5Pr$`A^1nzHA<~L*99oU&4Sls=Ze!W#l*dJ2b3O6HsffZ7`GQAkm&EQ2>M-3a z^Hfhi&GaRAe1G1&b~iZ0lfyNJLHLfol(X61t3!0xMz_^yQnhtIRB0kkP*t5t8=1eu z6c(e<;CJGUlAmhK%vJI^I-3gHavb5`h`A}Qzxai~=~Y>z(I5QPtYMR2F~Sl{MVE0V zQx(ARiRy(h1|AxA_hua|Oi{pXFmr}v;}Ewj&#@;O=B%HxhdD=5_sr&} z3|QXC%GHu_cow(awUB`-L{Jq2)IZ~0@`1+v|8$kJ(uzX)d#X&A&O>)>IgFS@z9Xk* zJPSk7?;KhdGGga7XBMA>%QyL%-Z44v{gjAvC;-$bHfgLf0g!oi30Y*w_7nEv!PRHeWbFOui=-oEx07t2hwWM?p6)xKIKIx9y+TB@>RdGC#> z`Xt%N3Z+93#;NJ4D%TDU{eGp_sXiedQ3WBigjNqTm{e$ApH~lT|zZqF2 zk%RHCJsJ7f1WSz%{C$Hle(O)x?5tdISFdcn#gu>nY@~qjeQbAl^R=YwN#9#BFK&BVcPFRt6&HY z%k9$mFIX=L7}b3dK7{Q!m-(r8rjc|EZzQ49 zqRvfLBwHZ7qV^psVB?ZBp6Y)|Thy5($R9;;?H-AODKlNN9E#I`Qf zGRWvI>Qs6|Oys{QOL#1m?>%6*kXlab<+g==K5MR3cvY#+!Lp#hzq_j^ytlBVArt$F z>GVi&j4#AV;o;G>QN2%$@gX}^0dbCe#>*J+Baw!h;Dg2A>fHmIVuztmgqyr;@5Axe z)PD(h!SN<~izl8hnc^xGU?v|4B`~k@!(t@n+Nf$CMRM_6yLp#I+4%Z z!rdLfPpW3}v{jb39UoXOinAl4r+?_lrV7q8h566k-TFqR?;|9e(1@GI$eYaiA*#lm zM*`lrzooMOZQkTMo+r3X;lx;6NPTil{=9cF1-mgfQs#*?tMu6E$kk7+6{)udX<@fu zlXpuKqggVqWlXl00tev zMcViQfS2Aw+_|oQ|J}*2@(iIfz%zBRDesx&hpRo%@fh&L?2k~*yqBV+g{qFD&@+z7 z!ig8Vwfnit?W`R_gNl0Y#1nMMI_AGg0nEHxt-OAG%#d7d|KAM!%U1KLR=n`6<9NyZ zM?EDx%C$Q2Dm+{^8-8Igc(}$|bO+1CTVv&tIr{vawEZn&1kChh9n^rwUq8mqJ(o#I z<@I`|><>27qH^>pCn)OFJGIM8a~mHcKRsD=+uYy#60MlpBpnvhDPJEwfA`z`-P#p2 zP+W_9DlZY6;Z`u!DE$17T-++65p#u+s!uE#>6q;mzL%7%uc~%wgBpk^SyU-rt(*kk z4|rxpEQB{$+YRiUEvI~AE3>*{$Hq%!9Li`MvX2l?vDov=#2=q!uADm;XA2Za7%mM2 z%kg|Q01jf(>he1kuBb@VFT7Fj=6r}FgxA4gTLl!ir^_qX+dSS?MNH<{+&&jm#Ah^= zG|P8k$FQJRHU78FpZYImMlxSzefVe`Zx2zOp)JDjNMR2%@Qf+ zo=h*2gvZD8=tf$=GL3@zBJaMny_W<_yZF0KVr-&Z&6? zRMz~9mu+KJz&1_VLO=1db=sfNXAJu3w;9=OtuCavKGI8liggW7w+6p|IoSUP^1zebHx!_Yoe6m_{ zv^#6`*U!br_Y%0yn=6otMWVGCh49F$DN=cEa4jvO?BE zD3RG(A5}eZ)(?Jd1x%>CyET;x5X$*grZio!g@a{pT*9pO_chs}mp87UU@Fh42nW_g z>JBMaWmxArPeAKy(=^%-m}fLXk8s;eZVx*`-AbE-DV`tB%AB5VJV5C9!_pEXll6ZT zq*sdvfadc$qlLYEcZl*s@9+5(gS*yl%Zc_*N$L!?e{oA1JIyY+ID0QmwY2&G4|ETZTL-mv8V4XsBN> zk>+7H){wXd_O{%2kBK+(A=PHC3{GK_AjCD>e8xvB*v;!8auL|4W#f*Z9$v87olG>9 zkLKG=E*y>NTA*S#&*rud`#r)@+MW7b9OA!rfaXh47;&>v&94HiFkk7v`^&2e? zj#OGke89iCxDXF}aSH)lUeN!m3b;su2Rfml6M}`6UpAm;D7zFhBhr>oSWnG#yAeWB z5|N!@O>T=JB_*gkjc_7L7p6POi>c0?Fx)M!dD3A&<#-w>QgdjBKn@=b7rMHYkz14>T4E-JCxgOX;^ICBL$58CW*d&CsRax??bPYQ#Ne(>xmhTlPC z3RW+58&lgePx5P{CKU^G3=mKD>ew`L)N5Vn=)()jRk`-xk3eZtb z(#mnC#dGPDE?!8d=xUgGPU9^F)6=&hiTFZ z;daB@6GdMgrR>{(UF#i+~xjkNXWb3Mc-Ir!|iu+t6 zRfX0l*7aBx6~5-IZZw)GG^kv@ax5V7TK8)WfrdQ!+GT8ztGg`f5mKt@j)V&SSwuiMr+{M565QY`o2@C~*LGe4<79oCKvjZgl zK7J1p$XB1&hv|u`F!Fexg&%BXYUH%z_{b$t^h&q*#jrTMG@^HLX-haP5nQ?_!@oJG z4U&VnkY*0pnQ8XT%|kXOr5qJAKYg^7H2G)7XrOt$z-UBTku^pN!|%2cm&`B6befr| z2{%vNy-D$}UV_c);UyF@{q+O-v4ZObpPk>Bq$}j_henBrT-1Kz zgUUwERVrH%dg6!x8tbKrs?hwHdV!ml%hT{I8E^|Ts>W);EUMHl+E+_$9yo*WZ2&tU1E*z z)-ZgSG?p9BE2W)h=WAg?i83W#*d|Y~@BRbjAE~^Yo*2y9U^AZZ>S;cUU#D-b^xXWE z`A;oFTcLDnjQl@w0hS`S`fEuet%4UaGBPF=2{YS(NS#4uH@A3_`zPJuQtaGvw8S?f zrOa{x#ett96Eh=zQ)?MPnh>R}cHz7O80^V>7l4FL5XRIuN{p*RbRFSEwkkjkNgTd3*mL2|e&Y@dh;_C!U0Gy~Todnfdy8 z$LU4b=^IJF`ZAy?#%#+ZhWtSHEcotmPQvwfJJ|rx4Oj^8J@M zsS>pM8W}y%Gx~60d&xeg4=P)Co79_T=earke13GrP=~yqC7p!hq2`1-T+^S(*e<}v zr@p){;noxEY4Jko_&8ri`yPoWx0kZGR3@#AT}ojP@0BOq=_{OM@eiPdPjbtxPM+SXfgY|nCna&DbEJG6 zrKt}K`u|~(xCM}_MlB?()-kmD_ z=W1KhV^Yj7@+N>#ffOYxXj* zh$>jfT?ZmIsZx#v<~;iqCJ}zw%93GfMWSSQ8o9?-gczR zQzDD*DYvYXb)0|acM4V+1YEkqW%9t|4QBFY^+8#KjhHKo!JND=ZtP=T;HW&*aSO-q zIQ1ZA+$yk)ceYwus!A>S9j{7lX=71&qtjc9Uz3vD)lO4$wU1Z__)X->8zrz!%E?qn z#WVDOL#(FOF1a9)a-P>Tow%yFpB((QtG}e#%l6jbANR{RKzC73fCUhp`;O_{5t_d9 z<%7~DEsn&E>!wc;)MzxtOvd#-XzXDL(iv>bca^HZnAW`~f3SjXs7aH{Ehy zm>gP2=Krp=c2Ucn*4!ybH+{zeTrHz(2pf@{sbXn=^Gaa)li0ZMh}oI?*AYMtorbM} z%@MjHe=5+HhAV#qGN}uq4qQUWw>FHWjsl>3LKf?CI{qN_vRQuI?j3noY>G;ZCB4JSrK2VlXL}6xonDk^6 zn^5RHDy=-Y-W3WQQCN-G%KMxj7MfkZ`jo+D67-W$DF zw!BzeRTn)Tl=_`j>3R&uPC?Yz+0Bz!%LX1{Vur^^wzynxE~o0NM{*VGX{6|S>c}$V z0e#JITvs<8!S|gboj?A!z5^%^DSw&T_;Ss(gRziNwq3Fmu72B-95*B2^hhz$=d`Tx z2X(#cz)=)z+@*1dL{1?Wm%Lma=m- zjj@Q#bvjX>)w|JXIxF|I{6~%ze6_HBXFOOo0a9XXXNSPhm_NKF8M`da;C_wMID_D$ zm=e{W22bjJL~Kt)0*WX-7MmGFBbird-I!B2oz>+{ zmJf(|@|v4oJXk7MY(DJ-d$u7N@boUWg@0>nQttD846=W?q>S6hj@a(NnzB6w7?|x} zl|@*ean~tVc2w>%Ei+EXQ=z*dP1Sy;N1q>TfalK_e>boSRJ=>K0BU(QAVZ%G9TYqr zPfm%`uMMx-tnnE_DK#S}nlHxf^`|{7#;r9b8?T%`jI)`|$7Nu6ja9n>e0n6A1bjJp zwI+H-Kh@iP8cAgNg$I(1>|Q>5IV!Z%*YOYbVSKLrpaGGrkYRx2hON(O0dYLJ*UOj$ z3)JyaxV_(E8wozHp!eY}7w{|R#Z(iW_}>doo6mdG$PQy^huqIpoz-^*ZJxiFWG?lj z_Ncq=M!zlJFpc72O~2}f19>>gTIrex%UJf~NIyc>1r!Tner;|3P#yqFsw$|wxhwtQ( zKd2JEHd!rC8D*XU%zu0doI){Yi1n;t!=6n*aBP_5X0x{?1zOAnC1ZTfj_T|0jtz7h z_x(egcBaGs5aLaxgUS453H0xM2Kp5z?50Tq!x>`|*rQ+Iz-KT;w}J^ngWYlRNRF5Q z3fFIZQE99jDBaVpJcB0LQXef#?iN=<&CY;!q(9F`i@L$!c%bS*PN%r3z!9-tj~%)g zLSV0bv|2hj*%MP^_`y=zPaz8gug$FXRD61CgW4gMk;zC`j~3cLSlphmXSLC7ak)72 z*d~&zFT6g$!4ANEhz;=0#|TLxJ6RT@$iiknT`eOTvQBL)4HoTMHvk5xSO9GtT8c(R zW_12PRR}4aH;yux$OTAaPMz8?!z`HogwxxR&M_*UC`oE89ikxpB_io#y~nYa}O${^3t&stH}ZeZ^l z9kxdP%dRV?`UNXJpe_Vx94P1pWLs+MjXJx9Mp`&ik@@)ZlM;jY1ON&(D)I0ndyEO^ z0Y9z2uk1I>dv9%ISINJUy4WOM83Poj`}O_G^hP@=q=1H=d;ixP_Q;aU$qi<*)030Y zfPhGd=i1udThXOe$#p@8E>sDcY94C4MK7J17~4a2^$ZCe45Pn@8sv1b(Ru3*W*kuRTG`py2Ng2CGq;gs zJ^>B#I#Pw&A`l(aenM9x0+N)e9MsJh10m*(M_N>WK$p^bp0roL7@D`w1bQaXQxfd$ zpHyJd%?qyN&69?>RI~aX1_7h`FH={VDezSdf+yu6e#hlk`qPPxJ3~Hij0bl`z3eD% zi@DRmo|EB2WF&Hr`PoV-IQ$2ze^^dPrtClGqs`yI6V_{H^_7%R*B~JC;Ul2%yaAcU zY3RGH{$^0kUvNBb?4k7V+f9e_END6dY%w9=?g4e1&6I%8&^paHQg%HNRKRfGiB9NX zkp%=l+IUjXyg|I8S_)fwG&ur3wi?B!B2e%<$V5ca2X4eF zUiF}N=P9rrPB-%4rs`RBexTwFoSeC2wY_$hf)M492J zD6Sl+F#KI3D(&HxCGk-aWgYStfz*d~y{O-{G>A=5SBmy%HmRVOu@;*%*$2r-L}y%l zqa-&U_OP7OJ>3RpO=tK&kG($qeZ8%e{E9(v2QE-FHvKWL9x^f*Go{fJ|FD!*ap0}8 z(@FrY{ML654BUppc>$*Agr7yIn|U^GkJ%(cezGPo-B)#QS!tmo0%c|DMAwAb<4+koT?dlqecgzAq>&T6AAxLPWe8vE|${3D+n%?|kP)*!k#eyIRIK2Rjf zW$EHY88XNU1YN?GazngILZ&KP%Ey==h@&7SQbUQujgHRNY$mKtPF8CpTLF>rQ3!1n ztS*yMHestNzTSH8>5=JgO10i4VX*FQ*uLhG!dpl>a<4szk~y`(!&{J7K8!7cyYR)-SQjgEe~yj`lCXZ1z}S{Gq@hmO$L zA5L_rq(J@AFJ9^W7YnHO==iQ^odt;a)(1IMU~ki!xL1tc8z;@}uXiO<%`26ZG}`ga zN%hpn3tHqL+nT$c1RvAo_yA*O4w*WGpGXc+=bC!*HI)D^oD|XGOu*Ej+B6p(cpO5; z#XD6Eve4&Uv9Hlsczb%uR1{3;mk}dJMmt|s{z`9llb)&wg%h z;6Uj4FGbwCfx~04@ja+XcSFg@6D(a;Uf#1Q(}ROms!6g?X?CVJHmQ*2x0^EpBtrYO zO4CLXu#(Kq)CE#6t(wl_(>(J<)P0AMSX;=sE73m}eEV$s#^F0q{rKfkd2qFdlGg`% z8bJ>i>HV$vozJpGl4%-15lOLKZ}KML1iApbi;Y43PUX1B;Z#wd_jGlo?M(6-6YTAt zO?Vfa|BQ?`$>Z(Z0X1HDTIckgA8^p|hv2LPn>RLHnVgh&MQN}Jq3VHp8fBi??OXq` zIpKw)sTZ!^pQpN$HKGlAMSPV_FTNbcEvI7ENeaf%@YmL)(hWd5Q@43_YNv)gFJ4^7)o)iju1ObrOx$e=^Duk2!+ay3(ut=S6&Ud0?yzUFaK( z3QRT2+^d%am?4pe65ChO4@sZx%9oGt;J!a@7ocMk zL=}dfD+}Oktdlb-0|Iv0SNN-tu>3qY7ohs|Yy3IFBW)=mK#mqgP{}V;q0Tq|YU^`xxJD_vmxgB@BOZjS_4v zm%T-w-`sxURp-&PPJ*MqVG>CLG~&qoq)L&d!|%4260Tk{Hx(Nu;38gATq^C?bgI1u zZAofaKgJ^A*)Wrtt~NSPuA~YS~Agg@A$`bzR1dbU@>c_@NT-ILJgq z@KoL2@smfRcS>=pg8?0!*CTDkRxfm3Xlh|vXW^3_U#`aWb|{k{;REs!2-NijtXTW- zuyD5=Xgn7%%}F|Aa5Kjus+dP75sVomvL!&M>x67f$io(F8narBlP606Dv1Zh=Az>U zRMF#uKUZV-tMAH@S6_!6c1g}JS7RT~OXrhd(t?ghoAAA=+BPwSU`lrfX0iAv8~BG; zz%=BzVZ+$M+Ah-lQ9hLbN7-_i&YrK!;LQGzx1o)f`K~olY51dsa0h8+qwS6Kyvwsl z{1S5hQW`cU=rDYgVBa=qP*YXa%lOl{>t=~s2o~0;^4HG6Zi`a901q7#S!50Yum_Y( z`_td=Pu4-}CO^?=*GMRiXT8iXvuTgjq42tl!xc_+J^ps0t`pC^>+5fxqGY#)5`aRt zPBN5$?f~Vl3xalK6KoZQxNTjTgyiSWiu-y^u%a4+$OjPHs`hC-sK^XOUN4m>4{KHC zyPaKRrn>7AVGjAVIFef+;?(HZn-8I5zY?<-xnG*<27YhFKw-Pj3-#y>T_sS%mZby~ zHsuS&L<=Axi+mLH14228RQUp=xz;-AXjm207Ju-U3{}#`vG9D<}@T<6bCI{S(b|bcvA2i6RfL_U&nK%L}Y;&~iBysu6Q* zD^u<&2$~Acb8qHWKtym=49Lrfr}FWJe_3OP^&O!jBN41UhO zq3M$H$}c{)o6TyTap;LpJzU81^mO8H15tD0JCKhzoG|k8B8_n4VZr51v4P;hb`OXj z<=$F&pks~;Uh{A1* ztx(mLxE`qCvsq7h0mo59cd7Eb9IR~d+>SzzA>qr7G1E=Kd`DB=$nAVa=}uOt4CLP) z@5E5<$gu(`-wnro%c3g-3@Ls+YX!8}ho?h=XWKm?d7q`KAfBg`W)La=TK~V*dnbAp zi?uqV9XQ(tNvAWW1L%@-)75(uJ^K37vT1@N2YaRVsG*@j9iSKVuLok1E0U$nb?jWl zQq!Bu4i4PEaEKG~6@{zMp&RpBlf$s%z)IOa%%aP>;s1z5MSeWJ%0)_42BOfv=|fM! z+gpsduk~JJnEuG-0KP8p+SViTi2dd_=F`=YqLKzDtECy+evNkm z#T$MC;Cmk{H3q-wQ26s7o_9EJpLK4p3%kr*E9Ly8iAKHkrGkym&g%`a@r^UT{Qd=6 z?Bnis#DuYtKoHoVVQXy{nmhRzyY2#I9pYo)8Jq*u1bQzMj~LMPkM3IQd;YkLJH`wj z3*hC2%RU<7t#bSDn;eO1JY3?bcHv@4_w8R!zzQJ<3=nICn)*{vJ8gZHv?p;B6Ae}R ze=hfPLJr86pIj5XM|FDdL0Cb(Fbq$yyO6`pIyw>eBNncxTjQ68V@@Ow_*TXi5-#DX@c8yrIqD+nu@MI#CPuU`i>F=GL9Cb z7so#k>uLkF_q%Y=mA@iNn_|nfY1Ovg+Ge+SqP+7gC~BeU;P9BL>wBu$QF6YW?@}$I zKZS_Z8VA@2G+v#*9yM#J&E?Ky2ulXfei|VYP$X{55>ehUk4+4YB=x?D&AK58QX1J* z4zNzJS0r3m5lE>9K$i)C7GLnZ)l&`eHNaQcNl^9ccfQqqYkr1xjDQCiDpq-3;hE60 z_3;TRe&;#TXz@1PM7RI#Yc6hLqv$fz7Uv9L@khrq`H>8KjLF}~lMWVlH2OCNe4Tsy z)PtVRgiMFX?a|-kvL$c^1bUkvF1S`fJk}X~*%gm5kX7-^*kS|=?hH0l+iPM( z;+P;O+{kpyT@r5>)`TAI*0H*)k>$={(96y+kG*?}EfRx4^WZ0QP6hb&t2jtJO^ZGF zyHWu&O|h2#kYv3e-$HJ)4m@$FVK7P1Qdcg@5z~8mOO*~&hS^R0E2}4b+;Lkdwi{w< zu9zyQqsB5`UmlZ}M>`u>9E=#fak&)HTEG#lr+zq^E99T~n@Z0zNQ4~Zr_xtkP_IKq zS~q4hv7wO8NMBO+1!Kthd3HS{D*lQ)4di(F87$V>shG&EoKGsa^nGb@b}L&UuLG zm~^{bDvDSJ9UVC($@@Ptco zR$8W@p3PT&7r6Z+Q99T(=1G>Y8aw`am)ujHu{hbu9%!xrht_2>m%$UJzV9pM1C+wD zvwWkr9L%yi3u8do>u@zOF~P;fB`xk7l%Ei8O@)8JZvf^Yjhw}$;r18Rw{!-&)QYygqs>N6)V z*jNgoJli{zR*-u3>>0N$oW=_}3_-e%j*a;mqgz4Y=HzJ#bWQw?ViFRU#ni6wIV)>x z5NKp%M9ACD&dv`x=MHiViJTSH*4Cb^1lBu%{QrDhteWZVayT0)@DT_E2*ilR#t^sBE9U# zdJ&EY-?XWzsW>JLnS}C!f&wTMT$}wu2q6`os*pbc$>s!s=FFjhplH?;x9OV=rZVU` z148f!_l}SG`{;Hgahfv|g-_eWhi$6VHPzK!)NQ06>%V&QrVJ3vF&qqx1GV9XJ@bh| zuv49`c8(gOw^!+RJ7Z1E$Bz$_c(aSXeEDLY?%9p+_05BH6nv^Rz?m&6Dk>^D^p%9u z2R2>jZ0_Mvs||sed@TPj77(uw3IItVflh%R?*-gN8h;;FKzht7wXA4pX)l@KQ0oN6 zC^ZJ;MeS{Nln`TFNA6%r*C~!3U2vm z(>50u7eYD^tx8%UA$?$wPoMrbh~asKCi8+XGVKJ;-YlY~rZ)MmW3@LX*Qs_R`#PQP zWj#*$Q*v^aH{ZTGAS^B4y=$#7?Jq1U8ZfIg=M4d_^eK52>U(vjSMOrE`s+Js1T6=L ziq|ek3RqP;s{gI($IQ~^=6U0CsR=}ZYQ7h;rm89xU0jsmKQ}E_w%&lS0hTEXtr64Cq%xl#L=EzG1?B3zw zA!D2juysQJUNH4Mwi$JaXlxYRPrtnM%5d$DBxm}lqnN2=F=zHWam3Zq)AJ%7-(BIvGjOt6$A zi-#8Yd;Zp@e;0W69w6cZ?t8@$NF4mrr%y&kM*qAUYQ9BX&CbpO z3Qj^o0w}`{M*n**9A?Ie(*<4Yf3~Zr+m_D=u&H#)H6cIM@#g|#KPLjrhY;%AM9~{o zti@?Qs76l!yzA)b$jZtJb$1$$f#v^WHVmKt=^u)QRuceWw(eiBM=i^VDl&a7PIth) zQ`v89#KgtBy1Mw%22K0p0iA<^Hd$deSKkqWODXJkz1S8&8)LaOnky%K5AFHUxyGM# z$XT!IV}qL&YlBqB+LrH8<6Gn88NyYQ#Dp4Cckp~!rPV}XYwNpvKkR|ye|>S>6G`6S zyfFyay&#ctk3(Z%-M{^3-R<}g7xpduTQZ!Tr9KxB^HJl*9;31kP3KenSA};ka}Z6w zSBt}+w%39J?uF~X6`lq5;~%t-ng7}LSt~c$_tHs4cwE-XR!=H{69t^rdq6f877_E) zT02`?9Pa?23g{gz?$I&s)BE@T+Ns*Loii_F&a=#IrJ%b7E!jpOaz%>KJ_-~rIZbbr z=3=3}!;J^sP>z_Gh5bjletcEjeDDEQZo)@5m@vsC(XE*5H`IPHW=v3Bv78XsW?oUeR^3<25vXE#Pnw{}B3Jz+cXW3Tss zCac$5SMlHX@~FJvqgNYUOeL<@`3k8*I!bCr1#aYD5N~MZ69u_ko#gD}&{EcR$~*ES z(*S?Y7B_-I0I`=bl)|R~DHg8TPZ3TlAwCQaNysTAp$GhxSvR1|udR{MQUI>db7EpT zv_%~2VOI_)ji1AanGHi{Jh}7a6!3E-tCO*BJ&)$h{nN9*{8SRScw| zPfSW{d8pw5>5gg0y#FC(OY}BM-zn$T7YYs>yq-}+9X9*@CNV1It=WqTFZ$v5zU)fax}?6qk^A z^ym@hNepwVy@*dD8k$F)5`5Ug8t?w&zxyou9pW=7jN11X(p|BNA`?K~Ef;|ipYMs{ zRa83a#=3>Cw9>65smTFp{Q`Ap0WD&QO0r1gnta2eZO519ymL~JF{m~L!R8*J4Ew2N zP4TP*kF-KWdThC27 z`Ds*5MrD6ZqmO~ehSSH>)t+6;;4hs?!CLP5G+(0eL!wX@SJ}7M$cZrm9KdWjY*ba% zqJ-vWXPMDvpJe!+C3P%mR+`EC_yC=W4e9^tc|#0Ff$O%G`mraOsu_L=N={mD2@XxK z1-1BLu_5mdH*ENXzHoIs_TBI1z@fcp9IOlZDoG)9xVIo&!&Q2DK$nsF z_S!6d@;NT7lCtF{h-u)B(;$FoVD-Z{DmAOEZk!)32;K_M3n{YzXKSxEHI)qTCedaS zWfNHW$Z+ChxQsh703Gmj?N84W!!JO0l(A}Nja|Rh)QH}3kCX{LeU_aq>{Xoc5y#1plq)_WO~1N1o5`Jf8aJ&ngdmK#9kEc#e| z%|o+m$yrn%V5*92VfzFrO0Vay+!b?ASfuY(*m`YaLqkms1C3|BhtfY*f4?44a&X{q z_etV+`8{Z@N+wK$7ssBHvzXN80G-xd^QH7_Mx~9F6ps!#<8lm%ts=_KweqRNGWec2 zArA|?$fdb!TR}xpjJaBdfKMN&QEa=SU_|dxv=YtlMtAL>cS>nRTSteJlQShb`N2oT zX+q2O+H9Tky?gfnxAsn-h2K@rp`ktDIyE)#*S*j3t4EQ%XwS^kV)=J@SO=e51|-aB zu&&|RzC!5MWaG1%Sjc65#{z2WLuhV&JzT}4$!YlP(s(kEG^@D{?nnyT&Ip@L8^K?0 z7%&Ok26!!#Q$QswDrmL%HJx1c(q+UZBmiqbLqmgscH<6s1;RI5iYg|6s>ghIzx*Cr zFvnj*6#TpaS`XKQQj*}0?S6TH>FVP0^H;&ybYbIc6Ie{nWwZ4c=M)yw77(-@{yEtP zZTkMMnN8i!a7?fT^4OSOWIWU971`a;kzR|R&@r*Fq&d+h<@3iLK72UmJOq-KmIh}3 zD&mj-`tendpXvB#3gLpBSSIgz8NN3{OCR{E*De6{P?vm^7#eIDLrc>w`BO4dt%#sq z{dR1h$KFx#+X!3OC(a{)PkY`iA}UI!n9hu*TMD@CEht2bI^)F+1=^dFzZY%3>7qV+|Z^Sb9xGy#xq@&^;bbpsK>$t+4!ADaF@g@*)!FfwB zB04AiWa~jrW%yRTqPJ*L&=LtCCupvFvvq*)hlMt0X=C#Q2L}kMa!3?1g!FWD1T05$ z{z{e7>R&;(cuY=`>vmJwTvj^|8*pmKt<6Ea#mPK$+?bYzRPsCI07h(cf&lk%(X zc_*9i*b|Yd?fyP19|17liiwDB#8IXA#N1wG?T;Bhz;hH}VPmrgZ0lVI_vuRWhiGiT zno{~-m!I4p-5mwI3Ls>L!C-e@1^s_ElN9V@m1Q0HP_k_)wW1{SkP+1^%i;# z!U_oa$V&lXm;UQiKjC99K+3sBfDMI{E+$A8NSoJIH4M_W?A9Ni2BPmpI&<~4+)R?2 z@7v_OFTfz*!r_Qbt6NOzEhF5&IyfGilgrDLsueIPv3?s8ESh1Ke!zFHzkscJ%|-Y= zPzlqzdxW~$x1VpEHZDIui6R#`E4{rey<4@!#Kgcb1#WOuT_+oM)qoW%R;J_1J{@#DD7g#|GK*P3L-*R z>S>&0{q0fD77~Tx$w;s8pnF`rwqQGR?Ou9QzH{cI{>EizM@*4+<;EZNrl<^*X2{Za zdA^q5?_HaB1sM-MRUB}9_-fGq)cyAAMwnQw0T*V+zfF!qahW-Sh<)n9fR*4EYtkB4Z6Qqgx^I?$dw zsl`zEp$aNr%=edTP~L3xP)t1|C&n)gn@T}X&$hFcNhc2;;rG}*tqJ1r+EZ;wb-SFyKgFi2ZoWQH$0H@r zS07LC#1Tavhb$}lj>s3k|6X=7?ax|>Y55bATgzx|2vMt&vWkj|ni>JHAE24}8(^xbJofWk zJm$bvO0D1iUV+>Cje!5+yZ=`eKGew(TY4vDE3Dc z@WF61F0!t6B^10bZ+Eognh;~ZOK~z7f4T)qM5PHH&xcUd`JeqdimGW??|(E39J{|Y z0&u0DBE6izd2P!rD=W*(TlcQPqn7CG>UvE>^Cs;6qtZKP-PxH}wn- z;EdXq%j?8qD?|kF(Z&9wyVQad^p9~3SPF#PQ8w@2hl244Gk5`i!y@-7ET+je@2ld> zx&J7)W?y(18v#KaLHq<3_~hKNH-;`ETsahx_5AkNE#H|GF4lx)>E88N(Hx8N>p_3i z!n!@ZfcNt1be5O6+{sk0DxEH0uzUu^42_py&~wpl+N~nz(%e-#l@CH~GDGibi)qw@*Y!rm94ZRbG>m_OJY7)VqlHf8)x(wN)-8mj4Wb%+vgwbQf$mr*V|JgNv`s{trrCawXV9-`W zu2B|t_HslPFs0$)VK^M#Z5xV zmH}j_fZg+G^~aBw`-FbyOCbP$?B>Ii(BH5}Ga?VH=*+(tnJ>C|06T4MWtHx+9QvJF zGVC71n>Pn{)-P8=Ts$a$XKs8PP^e3l+LR(G9flr3h@9P{ceBrYQB;F zGY24K5q9kC>_XvgyKu0!HX}sCDIKuh>DAA!kfI{7*aETG<|S~87Do-Dk&*3@q0St@f`4A zgmmT*w0Aox`X^|(8{gro3JMK1GF<+54{xgKqx6BL}K8(zfM#Q^3GfU#%m zWBzvr_2R-axb|tsZvY9FjG_SjCjk2!?fm_4mz$I8GOl@m1pp|4cs7GQl9l&fuKZ~L zP9=nJk%JDur$wws-M)VPdVb-u^BgfKa(dxl9Qm z0s{l3@@dKQ8}%kF1!z*ZZJ$kMsTXxQa9NJh0CCoT;h#TC_9O<6)Q+g&_dY(uCLh!G`X~XfI|=)r1TMO10>Mse8a^@c5ls1%7rGJ)AUSpVs zJRa6I;@F_vSb*}#;U`~*D=W(=)3I)l+pJ4&4gN_c`Ok+w!SQR8PL{?E*5e#r6r7K3 zsm$zU4gQE`dtR(zyfWVIRd(Jg0P~BBxBoD6Z%DDp`1i!yfC+VTJ6k*!F4}ywV4VNq z`2ZWF2!O*Uqyz*sZ{)OWxz=G6@lbBkkl~#Bn4TUgQh_q`$7>uxPj6%_XMa`SPi6E- zrGp!nfGQ3}>{u(Q45j|XveDk0|IQFLFyt+o_GEXibfqWU9AUdx_%Z@EQMRT|ydGoL zs`RfA19N#qEdK6e`>A5e#1_re=U3tXR~V{olE${4+%5da8A-gvGP_s>1lvya@%oqT^*S`0YlUsdNTVBg`P9I%7+H9LmB(7cSFr?!8 zIv&mSi=X(1_LIdoah0qKR{UNU*KOStIJ0aPVf?0DUoaqon)pu*tu3o~i}NIo552|? zFSUzK^!*C?vQ^g68PwYFWRL~&BwUCfhoK+qu#`v`0cFjM_l$p_Uo&H$>(4>{Y^Xzi zSK>@a;Fpu1ze(_TK)fP|`s#?h_2iY3?82ShjejQpWy}`czJ!M20YWigR|KWs0KLFt z?gZT1lw0-nP-fjs3^L^Ml;<&_NR@h{9(bl)icrvQ?x$w%PgR)4PvcSjFXQwx#JPc6 zL{Z6SYP;nIdc2;EDVVPB+J9APJfqF^NO&b`vLTr#Xw5dK?XkHF*WNcSS1HlM12ITr zm{?mj8?R#>hI-vbiq=Tb-=`#t13F(02h)`tz+qEXv(wH$-WQCYn$9SglY*?re$s=l z_^aa$kxwz+AD=QUeEquFc=dB@DHOL}ozd@V`4dE+hG~;p>OLnpkkMlq%VdlQ-IZafu#4YF!0z{~64)B4Ax{~1WT@1` zkt8Ij<)6Fz+p=%Q($=YlP&&Cv!=+bRU*LGQ!OlMLThI_?T!jdOjVmnR>?bjP-jIfy zO@Y+6hZw?IJ<@9i)pNuG`-f$Qr67ccnBttbsfB_y=gBOGT3Jqi2FsQ8#l*??typ$u zi4L$26VIbyJcXjJ$J8!fNyC9dUe3j~zHLd@3gXKjbAEZFz@_elUOI?Kok8=7No}}p z#|vf{4K0K5WQ0MI;q|%(n8-T8U8UejsigJWcC%t0Vt6qOh89;`Frt2TC3BneZqC?! zcUDcuVRnzJuidl=vTH+TV1r8OcUqkS1XY_4odNe0@?4w?5w9yOH9x9k|?sOrUxz})K$ZW9~8!jw5yHI zh!6T^*5#=-i%}qdPkDyVNk6m}J#6dGOEoK+W0Qh>9Y(jzD^kbu&mnkB%{=zZ40BixcWme}wtBsNAnbhL7#weB zSEtfqX`>b`mzctMl!9O$Bnr~Cv^@nC#J)@#FdMWjX2;6qe#7`$s$R?}+|{zKIyErR zTm3<3k!(P=O03MQaP5RQi1Q>-^ecV^(RZKCQ1R@c=~7{JnI|5iy@!ud7{vrYG>qc) z!5q2ngYQAcu=vhc8h6rexLUI7w8tz_Du)OUovzo<_R-}%f8hf~1lEvy-`ig++|Yay zo9eLaga|EpDh4@WVJb$JJqr0I`RQH9i-J8_&9h*-MdMP32*Hsli&IsaK-!WoUq=IJ zTB3_ph<9#&$>c53{vYf4<{XF|;5OC2e*LK; zT_S;6W?yWh-+*5+q%6-`2In;V5seeS7GqVw8Y>VpiM+VdD7w<7ujV24Q_~#Xux&`)0xQVc&d(673@TJU zUJb86v-NEr56vRg}OMez3+)b z1s)Xay$#s)7~y%r#OTXGiCM?b&`O?cQAq*$#m@+#gxdI7l&QR3SEulXRF8?PKA;1= zKuB|r?+CUaC$^c87Dc@A1a+-xypPi^XGxv(xy6UITdlI>vNxaCJ45EUkW)X>u}9vG{5Cio8kTN^*Ir9Pi{*xXm%Hs4&6|=$z1l1 zv)WDxt<38WNrAb7!VjtW=rGKC2HaKLw2NxR| z2?+^UPZaSGw1S5TB#Q30r?aFn9&XNn-Px1UskMhnPAN$?gKG0&b$kyeW z$s0=vj|j2_HAYpQq_x}ZleWz1AL%b0u9RpyIcTe?$n;#D2MtZ;NyV9rSj}s1rkOd| zPxzpAmwmFwkD_jZ*0-y0^ht4vDB;^NX0>Vcsg@0SE8qoxwcRwc0}J*EvljD$80}=d zmAwA@^>angvqX*L_qF?7kB8JnNVq^{Cb9Q|xW(4$nFqs~5*kR_+V{#|sNyp_28D1o^R znwTV(w1jFda<-vfCziO^;$oIJozS{A(@ zBFa_L&Bn8oFuiy8q^v5D)!d%d{Wdn}t4=27vs#9~5&GZ#5ly@CnEvj*t=pw8F9v5i ztknOSk(1DD2{|u2&}6Gb(}Y}T$v01Yu+DbO_tv12;>mBKDOHk~%XgCBGlcYC`1!K^xU+t>@1T@TP7(5iq2@0+w9SlobXRQmY&eIKNU${|tY&6$} zrOkAvp2XbL^nPxn&p0EdLg4oWw|mz6ak64&t_xPkfV(i8Os>RO$YqsCiQbN1{Ny@q0S#AvQ8017-nTYSVt z!Q2a!XRUmGv{c*V$ktMP(SXdM^Ea5}GeUtSgeLsHE3{`&MS~oZ@ zIKWcCaLT7l<!jqW^r+GUS9W-0|r`aL}k); zIIi{;Pu9mKc7Nw`ZcE*)26{v?@8X1vY$=yUL+{9qj;5`vI$kaTCUllF`Q1mEk|Q4) zkaWXkS#?cXeIfkby{uf-T==#{WA?-XQDp~gY(gza&gYd8Z276_PzpNzbwi3%)`Xq? z?s}7r{tLG%W>>lehRH?>7QKl@g44HK2=Ph15?+0G6q0bmncO$h`vE0*?}Du=lWay`;}VG!IAMnx)r6x(%=+@^qK+mf? zstSQfU6wM5*%RCgdqHY;pZx!x?~cGMfBaE_UC7|xb`wkWSr&9u_}vyo(mTAT!yA}( z?yvjQ9VPj^WjN^xR=6WlYNqs1)zbq#5W*ntqJn8Fw|Z{npj?}<$R@rRtST*BBL8Eh zx$*q0(=k!%SyfJYXl6a(sK&`jao_#|11Mu5L@hvtK6bBxd5&5KSyZ(kglsyVfOnKp z`nm?BG|*qeohf`h8L#j2r*Bi(8HYEW_60RKHTzYZ$)}#}Eo6yT)jH9ol|)s=74J`+ zqUE*@{F^6*W_S=978b1^@cHwDVU4iG&=i+Y@@zT_s!{F;j@!blg!iPeR$A_y3{qX% zNuxkbUUVgkJXm@=RTnI-+kh42H+?)UutKx)a-9`%`pUT`CL_9G~ zWoK6g6Mp+}XQ|m#3_Wel+0@}Tx&E5kCm`v*kEa{QSfyN#gCLL*E{V@AN{=4=ycXjL zcze#ImfrfVXwECdnP#rCW_nK52_sG}bB8@|xCGj@?IMYLO#u3Bo~hI;NVikv^$)+T zLcO>aKC6^AOn(?)2&U%@cOyz%x?t zsEcVUS51{bu^IA`(KI%1#@OMVX7ub$xw~X`LGBQz)~pZ zbO^EfVD$KCuOupiN9$)1;?Mh&J4M|P*J z$p4^3AWm7n2ArA!U%_#3H@CW3lXmo}RryGq@IjOpW7`|a`4zXv)XUH6VPEdIZ+L8< z55iN|nKD#sxQRFFQ{ZoW!>-}3$mYr#&#>Wpg{TB>0J&j$;UzS?0JGTAzQV=G5T??d zHHHN@LK0r+v1pYwvk}8A2W;KwNIzE{2QETcEgD1ls8q{%1#nqyiUtx$cYOzL%h zC(^*&+qI#7IXy7~X2wkGJhWRqJ;g=DXjxzR;m$9=XofuY&}u%dhSt65*le#Jh_y3u z=4N^i&%ca`Rb{>NBUwI6q=Sh3rG_`+mLPs?alQ9}%F;|FZpsgNgf2%3l6#`kz~wbE z>f&?pDAZNg;jgs8+sfAzhlm;Qd5l}KzFgKRZP1>`!Y9j_Rfs`a-eJ|DX{Rx)ock0z z7cV`89eh^%*@0z4sQO$x-+bFsgWFUighSJ9A0pY3%Seq)>HGS}&ikZ%OXlZ)*xk7= zed<)|Tb?2M<7K8rLQkJQRaO0jnfh@jP44oVpYf*N*kGr=pAtWic2>gi)-5Tzf=_kJ z*Q0q+sqc!E@YK0C;>x(w{XP|ml8p%2aF|$r8?s6JP;oj%_^I2STkUFBvint}RP76H zkGw43n^L#K=caL7-F+uBdzbq(;q^5sx~uYwl?#EKPttbSRjZ6|o=A2oitmVd=PN#s zT)~6J6Ianb(6!SYW)G>$z+VbApHoGVx*!p&r3jAEH!(&BAJ0^YM3N^gjMeZxx_tp<>_FCqaqPFVG(Jj%;&(<#Nq;_VHOvlnlt*gJYeeCci zbF5IYY_I+Kg{vJyG0(Gk6hCe5KFPjXShFED{v{+~V=i}WyEOj{>0y3Bdvf)# z{g8BUY!4sLcon}{`n{s_x`H|K>`EG)Or5L`j8C7qZ8`EH@hQbUWAR%0 z^vt%Q`LwT2<_1^ywTnVcGE))c#8f&`c=O9{ITfRdm6PFhR0|A)J(+h#ot@LU=U1Oc zu;_Hq1yff{n6PnudLNs!_25#5%x!J<3B47W`5G$IpvpI$CV9%k!4K|Y%PM=8eI8U^ z-wpcosd9=~RakA?`&&i`WI}WyH0|QH{zh#5CLC8{&z&y|sy1jERQ(hq!0fGWkvDjO zQ-e2&oQLk&Epd{9F&`=GY}IRSiSl#Ng5PW%7rMI%o--M(u{AvL&XM4xs8QFvlOb!Gy@iZ=uMo#GW13j$WySgpA7m0amieKEquoeW zE3Ge~E973}0*TNKh5o0){J+Lzf}ay@ihDW7ew8mh^wZ}sZAPc{r?t|E6MVQ*&j(G~ zYc*Z=!MjPX>0r!{(`xbdbV=97RL+Ub965P;G5n_g1c7QM1Bs!be zBc2eClLh5)Tz9sg+jYH|a&!3Y>^ezr-BYLMWQm)2JvgT|lzu2%4>|Afbw*5X@q1CC zh|)Kw>UkIP&aw8EB&DyFP)W#C^moh`K=%%OiFkqQZ@eLTemT^%eqhgEOmflGV6DQ@ zw19`nmGb2U^|v9ep~a=|7TO5Nc}Pk%B_0(ubW1C`7`Wuk5lUU7kZmy9cfo5h%Da)x zwJ$s%}Q6U|NOCa>RpfIC!;)Q{IKDSW@IGpn{d`SAym$9?RA8FD(-o z&hpz==*t;V6-5AseZc^2T4(j^$tG2wHkC$6oRkJ0~Z3?!)tv!{-KeJyWWT zyapAoN^b4R$~-ZUd05lU;OFU=l)>E`TXq4aWW?T5-4e-EpvBKD+}jCKHGW(fpG^Hc zPR?-l&67Fo#rpd9li_q(+ZnXi=Rm3`e&9gP8($q^|L1ajs#Zv3L zkU5)}ePGh@!jxlQYUirZn_(?3Wd@gbDyelHCfs*j-XW4(BUKPFDs(tZW$|0~nH${= z4M?+Gf`hNhcXwLMif&|wRX#ct6(IEV+;?EwAU&G%8=@^t!kH7tI05&;zcSh^I){Lb zTAmo29ryC>F2Zc{oh9|)%V8NsBSYHuv$wnq1+z_NeCnkYgb;9j< zigWoW-YMR`&GzMe^(~fw`xcX!>1UX=g$e)1j~}1IGGpPIrotln0RIJx_-|bDE*51= z#jD8gH&R8*S6=Q}tmscnuMaQ~%wDubGEdO@y%#=><^I6^XoAbEa_q-hxMK_B{Qbml zLn*jDkDmBn!Mcfs#VMn>_F9fcW#s#p3r>cWbyUvS7i|dh^7167r%1(H^?drXe#{5U z$Vf}$dUYv|O5HX|6fQL7k*0ldk*=41F`vS)am!TLmm|*inMNKJpM3wmJRRYOlAiC^ z3Vn8qmcFCx@etcxulSrk=*%KaXfsDNMI==MMmqTuSR0%8ZD>agXxWuAA1Z4s^9OHq z^^MT2b6AfAQucE>I6R3LEj>wMlfl8&7*NT~r_|XHS`L5Aqg_6>S)$=l#HI4XB&E)!w5BY{YIfBUPA-Qh!F4wJs(6ZyOq@xm zi374h;H9f^raw=U9Z<~Vr(sv(j^`hRSz50P$IZ>2H0-!_mzx`Q)`wqnQMd8PhLKPqZA1s7-8 zkSW2muzdT$2ep3;!!oY)>ae364m~Oe;y9>DL(Ysik7v@0H{@nu^YG&v)mSf=SMYaJ zU(p;n{5GI(_tClSQ8DsSVE$S`D-*+9<$+n7>!D}Wrt3BTiiyr5SH#*71^+yohj5We z;_j%uUT2~6++x86*Y&n*(YGrYmV+C5N?-gFRg4roYXj^2)&?Af>74KaqL%HOC%{wYlfL`^0T3B0vLN|6e!jzLQ$I<78;Ti z9aL$j5mm1G@1AT#>0d5}1uZ9roNa|o&4hB4yGY0wPGX_1|JCx%dh7_kI-Nx6)Svm^ z85ykiWfD1waA~|kbTT{i7+agUb7loU)a9d#BnLt%Ox3%kiskE*lI9-URx<0&h73q& z6XdsUpIIANKsqQTi%~p7DeAG#-mnbsPb~VeHJP_5KUqi^-Pl%?zF~Mlo{`yL(Ls1? zu^*v*C2X-M+EIS9-3iY|j6{Tbf?wE1I`URGd+i%}L-y%u9g2SIo1>9c+MO~FSKEy* zXuD)RojXv7cC__Q_R5w}3NRC4{fExfflCS2Y;ZkE?=~sx(dwBtb(@YLL`SNAdB=_z zb#T>k>ZsPEvq28U{zyt7g{6+tntw}^L`oz8Ay#dZ>Qm0A(w;gY|)vi+v%d~T7+Kz?h~7pY`=_1x;4YWF>d zjTqTy#EV!LKL5<#UZ1@%MqqePjzU?30V+`bs&>nmU)x$-b^f6q6_m8B+9hv*yWp00 z)ASwk>1Di3jhOuzR!P?YZLXYYkHwAN4%kLO%+x4NehRTzPsueS+>k$u-Qpm;Hgv~0 z)z_?GR(D#3V*V_fVm7@@GSQ1{KCx=Kt-VF(#~qZ*`zVE_<;wZ?iwjkKQ)wSm@nM|C3c9@Ow zBa0`eSK17)kQ@t8=yf_(^J)JeNq@v@8c)`1c4V-ezdL;|Tb8(oH^xXYH+KE{IlRo# zoP*$J{uP0`tp}RoJwbM1wYL1R3|SR4xJU<2GI_rIgYAwbrVJO$>@VK&eQTe2T4N=g z4x(E9)%oh~q|0aROw>`kIaLR(f2C9DC^|?(i+Heb?8o`deG5aowvBwc(fjM%0_V;L zNv#p{5w~icD=xBj^T7%u7Ky~yX(abhQGN>E35vN!txcZUyak5djm1l$y`9pBiF_T; zBRAi{<&;sZn~-Y=GZ>P8K;$S z1Z_>WYH-pTeR^3beRILn6zPkN`w@2qi{Fotst@+fy{!qNa9m2PL!ps7FI6O>(4V{H z{IkTC9jTi7JV~Ri&4b)Dg)7knGn5o2Tx0K*4G7I$XBE_DEG~;Zvzq@HZT`mir8jj9 zysLa4jo}^jWHt}u^?Mg6TwaRKvekqJE9R}fyYBh&l5nAZxFbVK^FbHieUIC^lk5Jr z>+;cXT#|~JBTd-d-K#&Hb;{imN!YJv=~O~RYT3$u zTh@@NqV(&SlS`P1vtgpVkT~|7C?jV0_M*XH=Uxk}L9aRa zXE{;z(&cN{U`-NrcyrxmquTb(2iLBPScBf`h3N$ z#3~1`F7^u>YL^s)3ytE^&olZek(pav%o4I~&6kP{pJn%av)vUMs9pO_uVJdf4oaS(Jbi<3-ckrN6GQ`1IV-B(bisyq zLrKsHcjYH(CYRbYV3n`py=hpidvs{7bTE$mLikrDMiU2C%STq|BM%mNgS(A3xv%B5qxj>b z)4L}7_0Th)>Di4TfdazRM;QbW$)l7{fKD&WT%6S&M24eO)Ti? z&$t8@EL^6Urk%dVGChVpaCf2nz~JY?KvoqY2RH`n9Nean;=+UyS% z-<0Hc(c*GyFZ%Q6Mp`kJevnU;Y_~fs6!TM$zx<@EPA3Qzx{wj7EaVvU9Z6U(Nud%| zwbssV9F~w+KT_|3gQa=-c%>8(ZoZG(`kqjh!&Q?)?!0@u-?zw|)-%85qwA92ti|_+ zy}Wp~j}hfCS*K7qt>`q8JUn0+CmsCaEk4&1trl*%9fLKZM0No>$uyPp;Eqi#t@pAX z>Qkg@*Q9wPZ_u@D3#-DUXm{+AzTAME_cDh;?2q}xlNEA`tuyJ{r*6%!Rod%zk~;Zq zMcWzXX{Z@+A{Ka^#CtXtzAyGu>~RcsKWB2?j34pR%K=GcZQ=rjj>;pZGdkCnk~Q-p zTTnGJhH}Fw8R8^MwScx+eQUMT3N^Wt{V~gzD{_@h)q5Ah^7_U2y!BcgS1TKtwo0=K zCgr^n`FGN(t)eYkE3_I%R>L;#s2Nt=>dw)eri2jn)&LX~D;$ey9-!%g1?PAmu+}G9*7?_+R zClRMs5=cX?wj(Lm6VB5sIpwYq_=+wa)Fz}C%|+nxnx{#KiE|S1hRfqqRcH*?F^&9! zM5Ne~1JOT%F4+At|4cJaWp~d@A^GGLlA3z-fo9*L(}@`-h%022+3!^+n)*REcLu_$!q*jmM258L^f&ZL_?{IdE5-^xrFwwIgI z)xj?VYOA zmgd=_5|Z*ut+`V-8d;NR@~l;fu^(GG28nG18}q(0J99lp=~;A$r&u?C$SVeO=bpaGZA1sGQk-xmjswvJ=lFs*&Yz-zqn6Ygr(`=H{S55h-5z=3;7jQPp0!=YBuz zF(?#O^{?H+8tZ{N_DAya^4i+*#SSkBd5cEI(i~N`%(nk2L(QKwmIy1G$x_d?bm?`> z%WTt`*D5^qOjG(MP!77Eoq%S9Dp`(AlZnD*f4FUw-XbRDj&Vbv zG$goL$A+kwOybjh%Z;DyfSR$qM$hOFBf$ZgeI52>zUOUs(1?(>nugOc(PKnZqwhDvd)@#n<h2=5Y%Z6@ZrfK714_|m2GuVhlR%&a$ zVXnAUSHSGHdbyq8-7+^VU&O6>L1C zoCwY4!POoPZB8dVarF|aI=;5rR}9N&o{pRC35WA`RZiauE;t<|BYC&q+eHgrA($bS zxXDx|PBU8W)*OaCeDQposME_tI&CfF!4!6k`vHEE}E8V@XL!WMJ@R=@eaZS#T8k|(;*~6ua)@zR-Qg z6lorYMZD$_1@((AEFR79t0`I7qIReV>t*-yTXxy_^kBM)&BSRw zm6LnxN|NNdYRsObx`VN#x>|#}zw_$1V1(_*(owaKTKm>?YIcqJ+MU7KCIU9PpK<2A z%1giiuZ0enUD*C&NXk)s>A}Y>Gv(bbgo3%wuH5&73D1Icz3kna%a{HkiSEm(wjs+< zs!~Ju(G9LUeGO>UR65*WQV68IyctpnTJJnFy-@qNQ1KIkg+ToqO9hFyFJCpNlz7*FYb0V!y#l|LK5S zy-+J;;i4Ip1GCA{9%*NA6`$sp8RwNn7YdQtY`KyrsyUIqsG<&mwc3J9tn@bfgA}P8 zeA|QcK9+-dlKmW2=#C)8#2*RfeZ@C->#zx&4ho-D^%aUH_85ONNqiE#@Y#TnH|WLQ zOLNZ2Ck6!kRof2htmZmOYef~sW`jNo;VGTbPBstPeBiqHV!E@Au8M2%h|xS#$N5Z? z^g0(xe4`5WNfPIWeTl}Jt5aMXb}u4vra#JMAGB>vu)>dWXH1+{2?y0zx!Ho-CrYS| zvKbaTZc+T&sKbsq-|os&E-4haQJdZK51J0SvQmQwv=lG4Czw*PqXXAU-Ggm$Tho8G zb4QwHZ^@q{TeM`Y(Jl4EGUJW(Ao&Bm1>q{O0vK?us%fqW$UVc@(KRM}wdA)UtirZ@E zBT+9}(_<>u-^3+zoeycmGd)r)2ej3aVLzX{=o8bI zBjzSt&u*a>g*?dFepg3h^mtatT1_XEL-#8sY41NpNYT$fk`wQ!x!mB1dD7H0>bz0r z=d}=dWHMT^Gc!RBhT3r2CS;Jh&7}VBb|B|zuh^EBmJ5W88OjB;_wGqS;c7;KZZ|%m zP)&Tdl=f+?SOsc{PXTs@!lBik(+?Nl{<#HUX|Awvsa`&$UDC8noJ+8)J3z7RnffBYt|VAO(S zRcD)yXnB~}F0OEEdPdc&w4?mfqKt!lT>FPWTW+dFNA1%vH?TMH#gr;eIcFPhnf~39 zVZj#N&8R0C{9H6#4lUIYv2m!0)WsWf=LhjoLY@BY8y234K4F*zf7MVq-W-~!5Bwu$vL zv8e{1*GuSONaSz4f4=ASmErzIE#XB!t(G z3%bS%pBn_p$oMAS=1_{iTqCiV*6JqG#-C$47%iix;*`unFCUXT)-J4Y#~9*P_hast z``|c$rvL^dlPm#+Py{tHit{^m{3=XdnK^w^p8l2jD^d`+{GdKg5sENGA`st$f`h%y zkJe%AKlp1o223>FKmIa2S+)CLtI9Q396%`nVvP{VpV%4GTK`-B)0y-Ps`1XBKaZ(9 z6r?zAnxRYZ5ysvN#vDI9=J$WsA;CZYJ{=a3^Z%>q{?D(gmn@(V4}*!5Nya&I1`Hb3 z*wG%gnC}6Q%Nt6}0KWGRHT5-&nG@7FzB$b2|L^-Ccm1UWF6-~IE*{`U&2`VcJdah} zIiUR5v;yYU%k%RK-=FaMt0?E~hYvkQ1-EX&R>i3NFb*l{eyA|T{yh(axh&i@u*}ZE zIdtB~zM&C6#mo#h3CpdaFcVmIVp!Nhn%_n4Y5sMyL{CjP5YnmFiy527-g;K;DV)H? zh2lR>a^8<@HUr;0FPwH~+=CbO?}77<0;(uS50*yL($X-w&oRobnk)c;Z~|t*iCBTk z{jEVv=^#+D0Fi~MG5zq@!be2E$w--9cGpK>=>gSa+NuQWM>4=RAAC!wc8O{A-qRvi3HU~e1`s=J_G4L#~6V2A^ds^e~GQReyP z4KKj6fTwiv?Aes;?8E(KH@C&UdnaP`p}sAcTJiu~Nd`=keUY32GXeIVjoXb=lbf2toqwGH}O)eTksmgoLcE^Qx+fg)}O#5iw3~DOP8>XeBgS#tBB6i)GOCA_djzVDJaCp~9O~@-M z-sS56qFn^9LoWJ!Fx4{{zp<=;-GFI`O7B5uvLqoPp&RhRqW;h30!Ld}!d*wBi_1jo1XfTr=8k!q!LNkT%c4oh8q{Z7U@#o;Q& zC2vX=78Y7svE)ueHj)!VvK^ClP`f{OuKeYkw79rW*EKS-rE3NO(G?d-IiP$Ky3Y@a zzwUJKY!NiSo!UR$9}|1uzkeU&x+7|I>hx*8)!J(v=~!9fP-3p5lOTzNI%D1?sDqGn ztktnNts}yGj zHRL!R4|WO*3w!iEVIbz|BOTEgAm7?|%9AysE-n>yFYX~)?S^f_k)NUZ7dVM8o$2@_ z;L; zi&zBBpZF53P6bOkKCxVpZP*7`TTg0sc9s3=XFGpcN;g!=tg}359B-adU9erp-Hw%| zJ-riOQUcxVQRxK?bPrf-rp@6@Xr(2DdF5Pb98&fQ#Bj1@aEmZpnOBZY^ccyvA)vT+ zcAf{rpDv%dCfg5Qk)A*t(+$m8A_~ohX-$hAH;v{7_jl2|CchNiHWi>_tf9D^e_6m~ zgnvG`N=nMZMvDTt;`q!B3lMl`yhGpuAtsA$13}ygZ7&s14=$v#7G=Xm(@y28**2$uf?VF*Xwu z6Y|;WOMG=6lII_MZZ~kIdbuLXDgvZn3ZN4VciITh(9l5be{l1(l9CdPQ9!1kiK8DB zH$()yRexNH!P^|W1FRt(Pft$(BL$Pt$!9{Mk#F$XWG#f68f`JT*BC zoPdr5(dZ|-y3-I60o1ZoA`uf2YV3c10~29Khr2Ntp1~69%bAU^d5sD5iXZ&-?5@BSIfF+YLINf5a_QYCAJ#TBh-bqB7-;Y@ zoWCptH$*_XdU}Ng1^bKo0^(v~S1mf{UIKIhPHg)V{`gK~-N3@zp@wMKAykSA2)Oyi%?*Sba=`S6}>#L$Fi(;>+Sl zg^Cdr(+fZd5UT*=*Q%r(vVpj`2XK}RisN^g)jDpWqo>Cm)dpa*QJY@1lf4tm%K|_n zN|k~*2)@1Z7v-b3&JSqX z17$hZ{3UnPhU4nQW6X$i^94%E??ccW(56rxcauGgQ7uhA{qzn!+%ID|_}SkmGk4cx zsaeoJGBT2dl@;(b7*7d{?9Odm%ag#g<6lFQe5vnCd}Sp+Ev?b`H_wy%N2n-O&f_K= zZft6zrJ<3c4TYOKIB<|Ni4}9h4aRQ3s=={(T=@hay*4MB7HY(MdTF1PJEHD%T z4r>X(xdogokb$c&`n-J$X0Nc{1bz$eBNq9Wb@8lw(O=R|s0ji}5#NT(2zclLP`ugIQs4`l}KvDys9f)dx=#J?ToX0k!Mh%EhUp>ww z)_c#Fhy^0W4r2&Kjg8acJU_AjSQYp21wb!BrxHtKp`@%Cjj4(leF#|@iC+>J z`22hfjBVpY)u;bDhgT|y7#SJ#sb2tDkHH10_$pDu%dR>$7P!Kjo1240=HVm|fUUK7 zxvMY$zuS?4kr87kT#gHSm>2U43-$E$I5;^Qz@H>1Ul<~Y7W8P;F)}o~0$=StwiC>Y zH+b!HK7YP_|GxcnQz)3olSQ7YjQ;Q61G8NlWBz^+{?(e|uC50la{z>%+{`1IO=(*%#96Ba^YWH!ILXqaWz7R0&CYHE zz3-%*?EkUSe19!Wny^I;DQRd>5EGXewBLTk!*KKFO%4uuBS_3J$g1eSPOJ+wozU~M zf4azfyt98Cv>@_V_kW5RvrcG8h>otV=#k~{^gM@)3k2}r4)U+$EiSdP(sg%hNm?7g z&d@VYVn2ELGGCPu;EF&8y9}+w_g9~P{V!1zfq-NX<4tg`y15-fMkEe)>LF$v{neXq z{REgZ7y;mZ{cAvoG(XyaNl*=b3g-O5b0ECS3q22hWzkxd_u&1{aI3qBYF5%&6Knh!3J#&r?0!WN&M12Qw=Jz4=O#>7Rupp;TodP~M2u-hE z`C;C<;4)F%f5VaaI^=Y)6@!Km5Rw!#i^RM zfSHbHm6wvjj3e$NCLRPSh8*T2L!Fv?=((QpZsw(mynGgOJb;!ki1Z5=UX30g^!^PfNh|J!h`{P|fKsx*DsgYD@@<_B=0F-Nsxi*W#J zg3EFO0PJfhRm8~*2nYbb1k#GSIwO4PUyaxu0qJ%*>@oIzX$bU`|mAwg^Lm zm)0IBcK~TbGXP@24r3{)K=|~}g_GdI;#9$#EE8X-7Xm(hyl*vy*|os^!214;h=_=e z1<0N*;&@b#&7NcX?O(B&e!to=z4MnYNj!Uo-%FC91)O+H+|sLyMh!zU4f!G@wS+{V zwOmc;?da&}>5+%C{`e0+t^=U?6*PL;Hoyxa+pWkUX zI+m(3q9bg+tuH}{1(Me-_~{SHKj!YvoyR*-Vwwi;{Qbk#eU9=5L7v|(&SD*d;Dy%4 zpv)q0zOLY69@`kgG#eWmKm+;^FkB_F;(dh^+}PF8aT`7y`W-t|Gg$V*#HbDn^2RlC zX(tdeAkDiDpZ`n--wIJPZD^CPudlm1)~j2>jNuTG!wxC;L?9RV`LBmFdp9;V1~P6t zf9IQ78!&a47T=eZmbUt}qV>mr-U6xWw_{TO_rjSh+V!|@Zf?k%7dsY!nFxXi#whH> z$Hi$KSxw-sVXUUmv9B7V!l>%it@83Rlvu@ecc^A&W*i4ZQ&aa|Vv&Cpn|?c{c>*Ru zz}=dSR>I~WO6lDggzZV8NQq<+1JhGdEbiZch!r}L*6*;_tkUvuzv>*8+TP*e!JQM; zOX2G?6}z00;2N@~k!}ZhjXKbs)C-J4^~dONwSe#p`(Xe$G9|?T_Jn~a#dSKK(;<4x zs9J2X_sdd1py~{{{Y}tVHHm)Ay?lKXI`ABs+v?0Iiknzi2E9E!^SwP_W`24=umic7 zvQZ0yQ!>5#JT4wM6VS=key1@FE6)XYlSPzUt*+(!at&P>^*o4!cJBe;uKw{a+6IGR9#mC=e18J0SVcXK@vP zqlXdJ<8yMJ!;2d~R(i+Ucx)Fx%2tOpPeZT|@8)F9M~UQXN0#r+akPy}fmjaqKoIu> z?G2v(9Ikn^U=4`MkR1|@96K4TP#h=m;I@?FKL$HHrRp7|`(5WY1+});D`K#QhxBXABFJPuPP( z2|!2KjpoQ1deI%ny@Q#FiFZKD7`O0+oEn6>lRqD3_x_ir$13(=JZ$AAa>R*2d@8+mB=%{lPN&aQ>z{j5c z;Q{{eFCSQ)a~vaDAVlisV4fI+6B?N$KqE?FRgi$XbCmIW-TUitbpDt$G^}{T$^s1P zIn2j@w|G$vOtxr1dK=yh-S_0llcy{gjxvKk4zA{Czl@{%tY*+>iMT|@{SP}lH8}wR zl{^30bLZZ@eG8^ydU_f&G6araW|6#*Pz^>*WXHv+fpK^Ghdw4meE%L^<9}Rf%xTTy zAQJ?q4yi8ezyz=9T&w`^E3b(O*e?Uii{ki!RDS-56QT;Hfq)}Q1x7WF@PLItQ2spm<&O)_&A^}j({_QkVef;z@U7cn3%?$bcQ(L zx8n6bCrtvFD0qJc1_qeld@3q>rMeaJ#{>n8W&NKM8e;a|DJ?C9#1O{$g)yr3|IwS= zwT?Hq8JmHrUJ5YWH;;Quu!w)Kd^$eHkW@1t=TR2dK0ZwX1r{l-4G9ls&zIs<$nP-` zucqQ(k0S-5riROAXcsdPF2;#gw-SI^tN={_VgY88-?1+)=c`EEnonglgH<+RE)Kt@ z0v(fKHTiv)`5~Kl?7=G2HhzbndE%`ueOX0iNw~e)(LwhsU9t;iUL&piLKO&bBt)Z2 znI&v=JN1lhzc#8(yVwguw4e`)9wepaCMI1*1rSx^S2KVlMSA(NCIkG+T&ZVy!1 zgj|mb5t}@@Pj%rFi`o3ZxgDB&fs-{g@vo;kx=qZCDu;$tr8Hzzb=w%}`!qCLHZS>n zKiMF8k2jjA>%b)#)$xvriCtq?3$>Yb+Do#}e)-cX8J$AohZs{lHjy)2+Fu3iD<&&< zUb{@U?*Cd1^puQY*4n(k|LgnP+LaZ1xd01AA0y)nEpjFLl4g9y^f$?1Cz~x8y*exr zuRS~zGL@&7k=513cM^o@I5-N6i;E$5=i*Yr45Q!Gbr#e#*p$mLeA)$R@il>Jh;=2c z@QzaY7lB%BlYt66?`!-0&$5kuAi2V~8N(u<;S*?w3Oj=%s*#rR_IpV2Dq0ee_eHbe zGcxxA&g?Hj%B!e-t5y>aF;tmg)_ebu$_avJ9}C-!hH}@N!`F?r9Zq54s=%Ijm@by^ z$F8e9N$$ii!z7-XA8*oP;zO*ToQZl(2xA31QC3{`rfQ3*dF9!Z4s*qIg=}hXI&f4q zAk{AQYPd15&$LCQcN?9ypsT~kt)Nh2OcdU4)lKu4rTkL!!PlO9o!)ti_t&3o_|QA| z?M0^6IZ66nQzOAIFohQ5}~r1Xco7>JvFU0oY!&Z>R=G@N;J8ed#N*m0WV z_Em)iA@nrM>M1Nud(6HAEqZxkiOPWZgssIOLX48I12sCn{XR_lV9Y1?k?E=qxOt(= z@I_MXBT|_3^YnU>`ufQ!9NMU!iL`-})=Cmeje+0Rn#}mQ$^U<2s_Dni$1+Ko629%8 zUhfAeIP_e@2Zw4R$9KBDQky4k$3N;U$$hOxEs3HtFD#(w2zf(eocxc(JK1+TX_NJ@ z&4H<`KDa*Hs~abC^=jwlJ(+JcHdn5^`lc);qoS%FruJe;!=Q4g^@}kJt~F|Fr~Dw? zTar9A{24K)me%Ky%f!TYf-g1NCyXs-t*db?yQ-kSoPMHDq^GLZmG}7@1%DF*x<8M_ zdE@rna6IKT%G;kG>b7pO*U9Pjhchk-7}HnAEimDPRIe1*#-@*Qgx}~3a44RnV|cbBxh_Z zPq$5cTl9;#^3EUmtc1ky`7>g&D0ZFOpBD6%8onCxwhC^JX*G*p#!-H86R;LA>$ zokd}Xeq%_w3_DV2&eSjasx=Mf)=r(y9_4WgM9&UC_6@kiU)P<+(CpxXE}|f?WyEj3 zO-QVlvG8%TT1r~>d6fa`z28nc<(Xo^$FwOkd&d`b8J=4W)tFZsNF~FmGd%|cQ&AP;x2^q zI_L}!wHbOzeAAXzcblkk;c$9sp6F6Cvx;wA-ifQg>5o{QIbRc^qHJ=hyK1kQT43!N ztG(15XO`21kkHF)3bn=E&8p5%ts`GQRu?-7IFF$`BLywBpZG@aYt6C=wfB>y7rEFu zyI*|XjxFrj&wU?~NO0p8_c2P0vNFv6#Y9HZx(^z3rAR@@0v-|+CyeE?adyRoz(mid zyviuJSqc`Chjrt3Hog=keP*$?bm^Q#^%b=Z) z&P{G}xr%?ZQ+Ul|!ioF#t#6yGciXiQL_Ga z3d_Tw(5kE*bhvr_W>p>|Lfk5obt0PzZ)#Eo)%>(1yD*8qS5-zyn_o2`Z^79?KtJ;1 z{`i2d540?oM%KZ+&ezN+nTZ`g624o*yMe|mu1eL;04^SX(VL|P`8miEFolL+Z5j?z zN_el5lz z9L^F_pyxsid-zurmK~9dKV;YXMYR*lxQvEghx0ISqtpHQlE(&EEaxNDh_7XTiswar zIW1GQS4@d)&#Y8G4iNrT%C3Ld(x$UZxh(jW2vYDdB`5ub4H7?smoS7d#iDe-JTeGIc#tKpe zXR22ygB(3p2nrMczrmGu;)vcP0u}Xl_F8>;jQ;!n)-)=-i@cGYAo0;lS#_IxBvBc$ zv#%XH$ITC2hR4Op9&bEF3X~os3$?!U8>}3jCRvN!%BvApkW_Rivbz*N?|qR6kHV74 z1zjEY_?)`9?v+U)-EU+DL|6EwqgrQKE*p^ zEGqYo&)ykdH7!Gz32XJatyK%$`NuWHs^B3xe%icEsdscE8*y{5jTr0wra~RpRmtd8 za-H$H&Z`yY2R|&q2{I({I$}l~mdV zz1O6S_a^H-PKTa^<#X9-X-1=!F4Ee7f5PZT)lrzZychCr2={-Bj$VxadhJ`J_)*1< z^QVL!jlC7@ek<62He#31RVf8M`yE@CHnD41Cbf8RYN0=Thvgr^z*Q65-BXte9jzZ! ztXyLIMEAY5qBOrSw%<5aHK}}!;x>!g)2GSf#75^U%ViffE7n!NzN37$^H`ziqNG2r zOHp}Y8Ksqu{-H&Jw!}Go7fT0I>-Y0UC46_8bHcAJzk}fDszU$Bf=e3zP~qlu`BvlU zla$LJf8d=cwAR@p^?r>Srp-K?Zo`KV+aQcHGEy2WHy^&SXyUbI#)HfwSl;zCC8 z?Tl|%wj##Z^j$a0hVr~LkWqQu65Z>LSJK0?vr@Lv8iZT-6}}Gx``%^YJO57qfB?Si zr7pIbw?9+Ov(y0AX~4LvB{XB#fPhF(D7sZ|cdeP}_v{x;?K8Ra{B!&E?SKABfce7Q zT+FgBh-c$|drxE91h;VUsqAWkv z6`wOrUb{jf7t3r%N&@~8O~CDvHop21!8%|qVcn+hg3DISMmR<0%DcSMc@%M_Z|=-l zUKv;G4Qj}U4*rEo40a*593R5N=4!o3%5L`6+jz!N-!4-8L)P%t(t$D+tU|-R0AXs5 zawdtN9_+GwvN(RjaUV94k^MI|iTkY)F~{DgZ5FqkR$I%wduG0K_CmXm>H}i_@t7X| ze8z`EzseNs&s=ewEFW4d_}r;PcSCx7XI|Pk`N~cl6VxjIvNuXiTohUK^CSD+c`1yMV9oP1i zOj64xIhF=ns0p9_hPZa4rS+0-G7-C3U1e3|mCm=sYJFTcpQpqs8cdBd}W?M~GXQiZMY)Xo>Uq-WXZsvdI z=%br#+ZkH#ui|%G6ROc5Q;Mr{TsSEte5&b@B)W|U_gdf@iYcm4&`)yHuCAP@MlX{WxXn0h?bD(@kYlG@>lu36UY4YF)3knN^av|wPe#7Y`tCYKYqAv8A~NWxT9va)Xg3e*Dmw0f)n-& z>p?p_e6v7s-(@Qjk4G*WvE@=`Q)4$BUY22GoJEs@Cg+WxZ*CdgFJ0k~FNvBqH{JFw zs$Ie-ipnx-b zDU~@7BRb3V$-j{E=P#{#)W70p$AJY*O4{0!M>-MgV@*skS`fWih3fUi8=TbglWM&Aopx3dn}@STRGo?w>KAsu>w2AgH!R$s zTY9Rd*-ldQ#@<&eUHge<|3g#MXof>SvzDWM`1(0tbq?)^$;!gkrn*emUU!9;A1KZA z&}K(}H(btA3ydpSR}dpkb2D4Iw(OWpd_zu!_0VNAMuZjeFdvX~L9~Mf!mymTY`Ibe z3)@Q~IauB2w;WdbTUsTW`qcWkbwp?@ouE}Cr_czkKizu!d9WpNo>S&_t)MPzj0>k- z+i)lM2j-us#0CD4C0D+pgogytW{O)gz9QM(WL(sF{$1}?j2?Y^jTioO@%oAXkFKu{ zYx4iUp78;L5TzwV1SF-ARuPnzmIkG}yA%N_5$RT%(J@AYAl)sDjc!JV zIowZEplRf8d2bcX=n*2eXx}n&`u6?v(LviX$JWoe0WVs!tfCt}m=!S#D=VvS1haKK z%UQXDPlOUAR#JTR{sbO? z)%el8+F&GBi9N1@uXc{ix(wx+T|v_;vA))PtIwe>r6X9X6XEffOXj`DInk{$d`H9;djE}w&XkeYQSEyfd;B5esSO`uJT_ zJOynLNyPN39H4qyCjLOvKuC%e8vG7_evo?o6^Cxkf|HOjcL+!l#*_a;gy6aKIaUWX z6gd3a57Y&ZVSI+pX$iW?QIOx-2V z457s_545yX_5O7bM;m*@9BFqv>Z9F25>wWpZBu3<|SSG)qT!yOqN_5L3U5pqQi z#H<1G7;Kp!#v9Bh3F6<7rNKb?Xex5sw$bwhcly1D<7h%7=7`Y8+r=*%BJ|RDqaA~; z&mml-DzbOd)zWshS%+(rdB-Q1(=^czyN+#Ph8rK9%yhF0GDylJiw2$g;>(SX$Zm>@ z+0|*{xm50M*zGaL#l`mUwyBe+ z0!TkO0WEP+si<^&&Xi6UpFAS+Z0$E5`&czzNCl+Y5h0bH9DT=>Nq; zSw1&#w{p(b>DY>Kc$0(}I(C7vWN+G^+S}O)7%H&w@tNb`8U0p_fHly1kfAz?mdC{ zKLv7${VH`eH4r|Ig?bnM6+`&V7Nw`piDUM}v)X_>o?dI~N%uck0I2miUv?OA7pk+8 zX98s9?_~jd$9R_ikiQ2-wVRYcZUMoG0c^`E><=$P_u^k(rsd}!IYUNzy3fU4+D(HJ zfO)^i@-gJ;|8ZP_#aj)hMSw}mQ#CO%`gOWT_7+d<|LO<*z6Izj0+rEouI?Q z$C1~iZ-96)7lK3ge^n`g{|R0=ss829QbmQa-CzFd;PLKj>FBUwpYVMwd5gXL46_Rh ztpUdi`fmWU2lQH($+4QWeGmzS<>a*f=YCT_pl5;&iQ=jnw;?=MTDId?D+U*)aTW6MKhUwHQh@wj64T~jxY zXGUMk*OIo%j>Zq2HR#iG3gp9rZ}SB?Djs^Xg6{(H^$+>8>hy_LT#SzoH0da{U&9A$ z5tfVsL(u2Xec<7jM1aLGVoclB~8+^?#%nzcG$me9dSu0f6>oaB%SV%l|!-pGm-O=&6d+{=P4OZpkNB4&k^P`({@K$rt9IY-Zb3Bkc^3yhZ zZnh8|-?3xQHIGf{ybwRU5QTOMV=(pNPu^U9a64Kh8lof-nljLJjs4@Y#0#YYP8fA= zkdWoM$>^4zXT&M|@Z55%e(S(_?7ZEnu>KIeF|9T2;_GV}uBK%My8?SyIbV7M=tZ@7 z*keCldLJ0Y_%hny@|4DH7)x&MSE|$q$>9j;qV#F#IlR(1+PKU zpV6f-bGSwd_T(;S&%BtlMp%ptToePRNMIqr(Dt2WP* z860%7&~|nV;=a$r%8+uI4{sGVYy)S*NX#% z02Y&AE(Y0R`H|cY^CPtkF*&;TS@@mz*sSB*eR82?1?NR6AI$ed@qq!Ne6kjGq4JHK z`pRCd(GjPgR_0EkXU2xsX|S! zmN8dcimeT`TU{2b(HLWf7?%!pR=`Js?S)YN)${P4#LSZ6P+(+)d_d$I$sT+rC)|2B|f`!Xa!P)Xb+qS{#eYRo& z+J0K{xwxC0oD3A8m%^~}k+?_tHh=+K1p@{W=J0z0`{hWP z^Jd!)bqDCG8^>7d#nwxcd7YN3p7)&XyCWKZ9v5{2U&wO<^jgd;oOYyFUS^`KmU(kZ4Kj0 z)MkLfGScqbfj=@pyPq(fV}0jie5O3J@{0qRB;#&(Q%uf7-avNQW%lc&9P8rnxam3` zHp7XAgS{_{u3}0_P{ke+if2x-X%CTJCD}Wh@4a6?AIc7kcAMpJ9pMEgb?nI$zMhW8 zM)5}icQq?Gj#GH+m4A}Nu9Y1kQX(ma1va$5jeo2qG>uPqbl%uq>Ah9tDvv($EWMGn znKG_nCI3kkRnzX7&8=ts_|*6U#WLK|atrQB6j?6fs-3vmXYgve&hQU-6fiLJhm#`Fgcdt;e4eVjFfe{sdfm*)s`9@~`X?&)+4JLB?5wNE^f z2Uqq%RHtcrR@NdAIR7rE0i1l9TmU5c_9qJt4-9ac_ECWyLQ3!t{-%;MV$&xrjZa2U zd*D=FFOiGSRoXf0Ko>Dp;U%eN+WUpdHu3wQi73*iJ!9?1Q8`&cb-1oy}e` zE?Y$f{3=nfv8=F=!M;%@_CcKXyL#+8>zbOXf(A%kib!C#$Dy=oQ!?T!>k&mY0M1_nOZ|TTBUSHjuXc`)9!^1=NNfhxnr}M04qWu)dSr_ ze{AAtM|D~hn%S)1g8m@aH5uwrg?ZStO`=oZPx)UV-B`4jX02SWpd8e+gpE$7P}68S zm8`#;xS8-QYhV2&JeyNg_eGCjRpIX1YfLFj7{+O znfm1BJD9TBh~QdGLo097HP1)Qu#*J?*U-X=g>8K_*U*Jz3BT{?+GHhyZ1%05`4Evw zfA1r`rO@&!#G%@_+avx6ZWi@4aZ8iA-Ujp(_mgzv8P;C+*4;toxTi@KLrGE&#pTwzaql$5LwVj__RQXy2ZPtgI2R+vAIL0O3(-*qhEJ-%%)>$vc&vxkZY6?;g z6o&GkYx<$ZNSftuYrdr8s&RZszHiObMu`QB$B5wy%JVw*z(iM^0)`iBw5Fqy`{cg?8K?)622wyMwGAau8osLk5lc0u-$ zp!-BE<9h77s9UbeeG2`PYqC5esmy(+(35c+RW%#&INb_H2Bnl=7a^KCH(BHxG^k=f zAAb6-om?!c7JoWgAEkE9_=i=v<+h`Qh)@=n^MSs%IgSP*!8Jri)~26EX))#G$+zHM zRY6!P{q9p0w)wpT66DWFb_K9LTDV z`g)siv5aK=XB~+yOiZy@dFtHYwVZ=Zm@7+QEFfisSq7+kN3{vnHTGF;U4_{Z1gfCx zcHpOu_lGw#5|i+)N7PPej^+ZhGb;pK=RVbY<>K6NWTiB7G zvRmZgZ>YHC)dL6N9xung6^Cr;ofh@3^PB3@0o&2&vvgR^F3EL9cL6*}DyTht((RGUt8CkN_npi zYE5StQp(4ZQ^gggPiV`I{aC@fLb{IZm1JD6?-p{KBupuF%dfjHGW?bLN2w0uux1pz zrnrxHZGIEF$r>)@Up#`;B6}7d2iGmLRCTH}l+|O~JHhzprmVP01Z<;Wt;cSCO|=P)wMo z72ZgQOy!7c)TsIRJ$3W+)pN?-A>~~aWqyjc%{Y7#X>3|sTP3-Lc`TN`Kh0Gvc;0&T zZ&nZ|B2IBF2@-nnQL$5I18oOee)UWKi9XuZlTa$%hLrqtA6>*;WHGPg2Qh&$;Xz42 z?9ClsNNkx`sM7B0TH&NrVp{n#H9%#p*bmy|K-t8>2Xvci(5DM3?@Bvpf&3fW-C+>@ z_V)pQCU3nW&G|6@j%SU~8H0vK(!}f<#w;Hs70994WskJ|Yf)*k|6l4Iu}qi+po@RV~9-TG`5RYcF%WU58W+tzL!bHZP5I+ z)r3pD&vVIGsZ`p%9Iw+tEn0>;I~v*J)}b1V!gRLq%B-gP4AVFztXor`yDm8-R|@^KZhI+`7&8&H8qM~cB&TY1| z+Kq^y{ZT}IfAlCXJ~z?s52L`>g~ct0A&KwwPLagWkI(hL{h-t3?|a9QV^W^S!=T5` zOQ~B?FzI!qqC;Rg4>27c9$fWIm2)>GUREluSb6E4EU-9ZV^vA$MNuncnPNl4F>ZE0 z!)Pv*$Y%;W>`n9X^VdTkHGHN7_ySf(PF|d?kOFdyM*!Q&}ntSK{DzNi|s>e~0h5dj*M=Hgmh2X%7qb$N%n^HbCO<$Cr zrgzo35>6Al&61Ib|IO!9n?=aKuZm5^Fy2(j^WM|$ST6dNrls)_HY~OmgVYFohMZJe-HBX{eO2&nfscrHe z6!i4GF4e^NW%`@WaM!gv=C82bowCJ4<(;w$-O+$9O1&*Bg&tRQ3-3LJDs8RD7jB>e z*i}R08g`Ttx7ZKT^T*!gmBGe^HVPH;5XuLFZ-#d@;-LxgSCZlQ9X7aBHx`HH3` z`Mj=|7ih}>#r&UU?WSCY&<<&V6Papj#7)-*&pvyNCr`7l%eoJY4bj%N;Ghb3m(odS z#_w6yAQfaITN}6qb@Jr9!UcV5P+6~5_x4d)^g8N$4NJ8w^C8z-mNYN^y6oeVW8kWE zifn(yV2p20C&)r)9rdVphBIz)vq;3-HAal>=@H7mGB$u2_aaL=5ATqqTa=nZ@yP#h z#{0bhbf)s`*gu69MtA{p*ILi79MNz;z4fh|`vJQ&H`~|?i|2M0GU}21BCe+wA%+Q0 znufji3Ru01&*`W)539~7F0VRJ+6CYGhS%UN=tP!~z*+Qm?QWGI=2}C3@Q8aS=DZTN^0Vr01XF&Edm zDCDLFzG{v}TWhYD|FV_WcFk29jkh_rBv-2$`Q~=UHV>RCv_n{qY$q>#nJ)AL0<0Ta zWEanVexu43{g~OJfP2|4E*NvcO6o%UeBa0T6Uc0$n^g8CL z1g1HUtca+{?$7{Gm%JN2)*F^!1g`}{F0>mB}Po+u|JMgA9h@Rl<}*r= z4XHO>Dvy`NUbXe~gsot#G?T?T`OX1OBLI_dEI7=2=(zYFQCOw{yq|ay2d?#V@ z*N<$>;BFJ4=?3|CwD4ptN>bUD{{+sE} zz6s)0YyOeshvvSSB$suT63f+4o4Fc&-7gPJn1u+F*v^;TR?PS{j-frlc@tpW6<}ij z%pBg|k4%kAxyun44Ti?f`8l_U>uQMjG)F%xJ!y_9*-|Ehm=|GWQhE9Gr+VF;Sc+8r zo$n6%ktl1V9^c3N^~1YQk2P&+7bbh%ec6&d`zvsVjuT<4bLo)yAqnz>=Vpfz;0>VR?JD@_@oe!!b)83vkzI#|G-9IwL z_w`2jMGBuU^xPMN;BTQXoSt1Zv$WZFc_QYzX|c*nq1rPH`Y=sEJ}U50%4$pGvwZJQ zdofSTS@fgbQzH#4U1-W}OVb4t)laE=SjQ=o+x5IowOT}w9e$sEe9+wXpyPIncR9gN zUk<00-Gif)cTI+(o?0<=59p_(Cnsj?oMjJoJGPU`=_^v)w+!Ux=@%)yqyy%%CG$?S z_n!UiZ6$qy(Y#tiy>!%V$)oYWVj;46!ndc_1^$zoBY{y5jVx*P%4WH5yg1NGskN?H z=R3A-=umYn?Klpf2Q#;6VOJ9k-8S1Atbz8$w}f>DAEk6PH*6K~@VGBcU+_BS*B6e` z6L%$@M_00oYaI@L4cVu0bT)60BQ{}b@21z_K$JLWYWO2pwJV>HG9LWU_pqB59Q9|@ z6LysSv_I?ele+NY&No#C=(!?V#A*MD<|%z@>Wdk35xDQ(rO`v<;6Vp;b~46%m<2{IiXc)DBh_c1ID9^NC@f4x?(_*dvbi z8LDk}4wc2t7=|3Ejf)};HO(qJdo4QX6PZbjI-I~d3pK2_-`ZC<55u>?48EUl@donT z^}{cXw$L(B&M!0O1uT8+qK=iaX3bf2A$6LUT@gp<=>&b$2pvwRF_cAfH?0-Xr{#_O zd@_Sf;$3Rs^{>t9RR%%W0mGOsQFi%Qgnvf$aZg8*IZmyyXJ?;$Os~J#{=B_fG?_74 zOv&cxsM$B*>||F_z!fEEQ+NOX7+353QrOvv(e&cTsLRT8d-d-Z?bQJPDAAbgZCE^y zXv7?k^qba8oUFT@=!fPHyR7s?qQ_n&i$1H#mAN@BsOC9#(P4NE(5^DJwzksJ!QG2P zt4@H=xZLyv=zFe40j7*I@vr@Z11$9+qJrP(#fofW$i}@I7FlO*=9Ah;lsR7hX-kDJ zG-S|p?VuA{6u7iBOd8thK7CU9jK}hHX)5M>vyzI4tE+QHyYP-rEFHh*ov9%k5KN?o z?#tOHbxat#W!W)eJ}+r%odlcnQkmbD7n<>W(ZKaoNT;f8Lfsgq$}31icjumpJA711 z>gK6t652_{F=O5S)MH(ZiKW&8f#^M=-R6|>vC$37kzUl3Szs^=RM=aDH}7dl4rd%=K;9uy)AdX;yFom8dSDYj%4Drd8F{ zfM5v3u>`j=`kfvFw>?6t#|E6_dY)*$OAz{@)#0278qvfrU&^rdW?yMzlffee$4EuX z^6KX&w<0nShRk_oBB)BmAM@PfNtuK@s!p|3IsIxWVoadUF&EOe7J3xe^i|(EDC4+f ze{|P29{d21$DyZrgL-NL||4?qf_*!>spHL zHz1AuVNRj_!^7CDwrBj_WV)d4hk52I~U9Cg}dE@i@7MA1Q-Glb z;PoEq)#YMOW)O0^7CVb zUz7hx1IRC9C7$f;2W}$Zckr(SO#g!gOaeCnV8)}$q6RwtAlo0Yzs$=ECf9$z?4}Ih zHxN40c&7a7)xg9AGf-K8&#R`PfqV7p!otEs;v1J{h_J3u@Hh@#hgu}|$i(u8&;^)} zs*}_HmM_M$TE_=8w1NImUOv9-5b_%u#d=7r1LkXBI<}uB1#xp45-t}IAJ>Ni>ZhBF%2US+RDmvF0~+o0C*{}Lqs?*&^rJn{69Ak z*aj9}eygANMybIN!g}w4?!wemnoE!f*%KMiou93o2auJ&{Fr74!AJl7+5dZMZeYsi z24q0`KF7v+-kVd^z;(fLyy!Z(T_CClob!KQx!dK^($XPQ6<}}By$JgG4fg-x;W>XS z^(MiAJzav_*wE1K;0s#a8YgK;QwFfC`@dg26TP>}r|@tRxgl&l@=sTw!3qN-W05x8 z2J|)S>H>QkNvt~`#P9nm8Vw9{>}!z#M^D3k`<%qUCh97-)RVgV&&Y29y52Xy3k}>I zIOv`pZHe2>_^DDHhHY-l+eY18O}2Kq}##_u};fAKRU zzICg_Znjx1UmbV|Oju#Pw?OmhH3)gfXubw8PGJ4*Lk8M1A$Tdkt`794@Ui_iNcx}G zI6B+wZ#)>&22&Pnsi(kt5U5d@ATu_AO8|_54czl&f3@)(ZTpwsH_!*kgyjG8ZOZ&T z^B!akR#_pS8`iR>6fu{{7me(BhA=F;{oTVneB779w18FFT>%@=(x}|)00af#;jFH& z-vMXcedBMyb~;PBP^VfMLXHD~p5OiH88?0dI|giDnDN^*lZJ-)vluC2i2b-GJLmAs zlj_GEyKa+=^GLx#$3EhHrLyZ3deKaKXz9eN1fUCaY~u+sj|8Ysv@_3lQ~A7r+f{PF?91$-S(Ka z?O(XK!YO^g%&n@jQXN8G?BVIz_-pPa98oyF4vdxl@|z_F?9OYfqGHg$(Ez+5Pn%UE^CmkM;y-d`K?2nQ zSPFD7pye>@5SR`C-zo3;HvCHlBBig{z{H7#Lcw06^f2DzI=PFXm<=3ZQjSHJ=C~xV zd&2g}!|?JY>kzdYamJq2oeMQbvnAy7v8vwjc4}(0k1T4=OW16r)`X=aWfTz&UCf)Kr>{Jg+dI2q3TR4CTj(;c?;n!9ChH7TkhFy3Hm-a_|c zXrsDFdyB6Twx``bV$>(oE&DO}CLuvou#JXBDALfTQeBgsME(;2*OZa{b!kl;iEMsu zt%=d0F&%a9?1F}R35qBI0d_TJp*2No9zMbW{5*to;6!Q7NXNF>s`(+{Q(e7SsLoN! zzMdNLIhASxt{tl7Q6`Jx#R7r2u0O3850+Z86EB!n`q22-0~v7Yxs;=FZ@37em-uC? zo$e-1ya;J9LYgr+i`~S)~JcMo8 z;qk1ATc{j4*kfl@*Uyg=4P7;i_ZREcCO;vH`wfr3b{RF)E*o+Mvek?6@i*TpQuoNT zZ2Mar;awMu6HOEvh3EJd7P#HABI8b6@lxjx_Yd7z9{f}0Saoc zhyX7OCi%n%kgzKXIgViJE58LRJ=pawH>pH^{=7XOAr6)jxdd)Wu#`pq(fw~%bVRph zj>t=YPghaXV;Jq;DqLl`AGJI5B|}m}XD?xnalG(Z0h>%(nV>xD^Q3O8@YJM(ao&`r zF~+=;gkuU#{H1wqnFy+e@|SEXlwsxk40)1qP-xeJa4cKj=SpUVvoW6LxCcSg_S^g# z@-SaQrR$F3J?_rZzmO@zvY4#vgPL`x;{5pOb!We@#HOv6UArZ5&5s!7JuNmSl;}`w zQex$U*A)-h5tfO3BNANH*UyVV1`ed}d&^uHum8k=kHe4lg8Puwh+W{sXK7^wiK z-fxwg>%4Vw-tFt7tC@utWh3gp)OI{aul!&Htdggop!WWn@a8ZOZlHK=Cnk7K1$@=K zU2i}8PxYZ@qgVMRisRRq#3li3OMqK#2sdeWkA6!mta0xdZ#)M0zn0C78xyEHni^XT zp-$qo;*!!&omp!)UcoxOPSeUhCNEYaHmQ1&#(VO&G!?9vMPH3wy!?ifWj|I%7@Na@ zJAL}ZX0om@wv6Ggm~8zcrqHCw2KLjLTvSCAs&2|P_q)=TVpq9Tx?Ov=i<@a43pXt!M~57A5mf}uhLvBDO735`(Bl%(HT-!X5biLo0w0puu11^_9PDDYL zmqaZ*b!U^6hpeGgRDM}B458!}nIwUQen_pl86j zvd!t|n|$Zsl55D@8TQg{Mw=nxl)RGFxV`7BD_tL>tt}5azi*Oimv6VN?}v8+lxdlmdZ zHh+Er&%#-;y-VU36qlq@!6p)ZR{)$E;4m04UISYN6R=={kk@u~cLPfUiD%D%-z+#` zN@6`etbt)*`+Q_)?ia92r0*00mJUMBK9VOi_)4WRuVQz66ti3VPYIAiU>qQ
    #!yEnO)X%5CBbvf$L~+ zdfi3!jflNh2jZcY&@D~cSzA;m%Pu^}obSY48;@@*-YZnaEjftwh<5CO^|Wzwap;uAJovT86O7b8<(@yLS~+gc$fhNesg_CCo4@v&4WE1i0A)6YC2*a6}+4 zy^N)|{RW#|-t1AE<4C(Y9-PLg+sZ$F%&3yD{E>-XrBxDI7xkBuRtVcsHbq^KBX1Iw zzjo%6Y^~(f^Y4ktgzD_w!Q?uhwX(b96lG(&Tx<|QxSuqpVnny&p+}fe;74uJH8LW3 zd~vHyLmw4svTe`r?KbnI+O>AOfLCG)LcJ>FwEc z<77l;k5y{2psPiftK02=nN0+#!$;)!`?^%|2b6*qn_@Gr^xURPj*TFQjhjj_+1@JO z$Zcd=YnAfe&-gJ&LL{v{bU;a`Q+Y>wQ;z4?p(Q`H8@xu|g89JQ-9N!uf;0;La8%$@ zVEuLxP#Y8-hV1pi^%B{*m|JM&%^quen07P?%Xb`KNg^B*%FNbDN{Vzk4RaH=+P9f6 zqw!rf=uL49uNc#?IbK_ieRp#3!SpJ50&soHTw4P5FmTpcTU!HmO{;c~ulkiPPQolf zEtZ9~SMD3h1Lw;e4lz_xqxBy;+~`j~R~YH$?i!EUkpkC@dFg}~Oixqi$DMI2(bP>2 zeFja6ggp1=`n<4mlf|6Og&c}y3mo{X0Oh!>y_kxme$pd-+^0}Q4Q-rU3s#t>bc4Va zJVC60;^Buu8xH6ZBPB19$6L&&pO#`PjWPL!U9&e8AF3&JAp|e0S$sjAYAC{;Jr|!; zU)69d?Swy(tm5EF$3zMdYK+FyID6a#0*C5h#4&lwEK%&~=hoG*1k^KM4avnk=p}+R z5o@U6y0w`yEHP-wfSBlH>h+fJ3*9jl#5y&SP_U}Y?BkB}7)MH(Pxn_h3ri?#U9t4@ zc5<5(BF%K4?7_S7b{YwNhj$qwCcGS@YDH(kx>f-i!C;o-EW|fzkqv-G#p*E(3Rc$F zvCW^l9B~S5W0^$%U;#OAjVWd{u^ux`X5YVHh4f8kv%GLY+=oAFb*OFV796tVI6xS2 zYm{9=pSHGgbH?Ii?>feaH-!aS#GG&IE97{<%1b1r|1mwQu-@2@vM#5^#@fC=tif~1 zIe?`(cgK3v#+u4*_gX~$;3o0?*SXF0t5=tG6&(Z%!mrTe4pR5fxxQ(twdY zdg+R9=QokHXqQs}^neoP z2t4zfkc?^gw2vn+GR{)229Us|84A!t0Z;^BIK&YroA==C2BfLuZ%&0>grgtApc1!o zTcU0Fwc`SkXOd9me$4K6HFL2HZ8K66VO0p_@_udXQE?sTEXA3yvL|0PFzEP3Cz4~l ziO(%AUt)3s*=mx-GDK_tQNg2Sf6MkGE)AWWyDekW;x3+DfpGir;g%e?b2Mfvd6t<0 z!!9ZEQBLUdT&i7gPWh;Fj4Dz=r>9WlZVu*c_m#B# zXyz89^!cZzop+(v_V{ZklTrv3;SaOy5BZWmwO!ILYge;$z2RN;xqB`~K$t4)5LP2( ziHrI%Vdm`oDvJ>BOo1repSFK8k?& z@LJ@1AH{>IN8I;7iTX1&Wg}%ry-M3E_J*ZrMvLCMor_!fy{?`xlB9zOE%fIA{00=W z?=KXwOu{3dQU-XXM zf3aWRA0x@YexfatX@{uvbv}EwTrX-bSimmV?`KFK*X`)wkhyCrDH$+f#qsbV7K#Ga zju6fp0DVGFl)td3+wEcn{Vm{`kkHaj1B*m`eMPKwN@Un?K$BDXVZg>t$caR+F%*@i z&(0Sec(vuW|CU5P>Gw$1xW&8lH{v!5&8Y)-3tiE8`V!w%2_6&3WR>||(uO*oI?~{q zy7Zkli+JU~eTiEm82%3ZYYvU8AUjMT}mz&CgBNhfg34a4x@Bad@7SL0LcIzXl-Hsue@iL;# zV>ZqS>+-ZTY@O#b814-bpD}bjD~gjMQsW>aJunK zDUoBoMXYc14*z@n(pf`7@0$F*!zUCtb*e_tg_oGiclxw)mQ8}t0>a4!N(5EGbjqW} z`YjGpenhgxiIF< zi!Qb2V4uVmEvkQtmYml4wQWUp+5&|Q{goJV6DCPK->=^t9Z4VwyGoxH%?riL>rrG> zub>%0IdMn@#5a>Ix@fM>(9?(ya#4b^v`AmpAXptzdsMxRpknlvgP?-8_ZCuP-SHri zE=g9x#c}ewMoxLgv@oXSz6r-Cax>NlJ7n@-Px%KdT}BfN8?HTbGx@@?7HoM=jgB2| zD3Gr6QHL%W*6ux)w{CEHR$`N75sd5#8hD9BBO2 z4Me0^^gQp6&52@oPHAkFBV`|k%r7H}S zT1j%o)+nhu`@z@`r|R;=GHO5XNo4_ArJ&SPdsccYd!2?~+JDc-Sj3&ehty(9AAAGdrl#X!=3|{p70csR8D1jbBxrGz@{6=?{%{=_f1a=w z*WHKprwD7je%+BI^DeTuw)PYV*RUr6w5@MGuyOYzYWY`HnE9vSUxY7JDt}@jY3ZKe zTl8U8+k}AZ27s+tE0=+?KBq|yv9KM3f!P+nK0FMIuSGRH3Uw3S^3Bs~oI`*3SQ|3N zy=XOG*V`cSE+1*-fTxoFv?5jf!YAC7#cw5ZBb{$G;evw92s%-&PF&dga%0U2uOzZv zN97_C)(szBE0dpaFNM}Z$ICtzR9kjq^f852-!5jm8AM*a0LSZ5Yt(HoBbmYMy27^j zTGw1&$~Y2CKc>*-yza+SQzoxHC$NHosmQywS81EKuZVnpLm|g+X zXTJ)L;&Mz|EjD`@9<$R%r`0DAbe_JQ)oDSl+rKqy_}JtpTvG*h=~DA|&YEWuGF0*cW>PI;RK z~&sM11Y{YNqH~zIqN_7;zWMdi5ybsO^jUCTVR8 z!|;#%Mdf#{(-fvAA$s(IF*$=?={-gLgk+Sulhgc}=jf$#YudAO@oCH5v4}ed6uvx0 zXYx+4pSB$DzrK=eJH5HCJ4Q$tFV_-dx~%|zrBfrUlmM5cO^#1ii%mXQm(YzRY=Fw^ zT&#UD*b=!@Bz!tkaprtv(>e)Da{fM8Q$dJ{t}Glzt@sAYQIHiKjt>XT=xWfIPG|&d zeH0Lr8)j*_122%DD$6eaKD(>FXk7MCt*T8t)DInbIPw1xGK`{LpvB3AoQ^%58w;s| z@}d2Vy|S_0S?^eCL|+~9xMMOhdunI=+jVpkAKeL<=-k=ui0S4p&>0&wa6KWViVVnW z&NycpK?CBk79p=KQ!!}CqwMd$+}ABGx;2Kfi74r{H$Cbi*qo-Rc8OA_Nvk;7#xxXs z>z!by*0@81rcLMQZj7T5+^E~TRy3?;?8|5MD^%=tM)$+x;=G}pK&NA5-C<2Zt2|Uk z!mZJsJP@Eww*fJu54Qm}yE}mA0GWkH|; z=&b|C^+^OGs}g*>^n~KS5opDz^0}2FK;p9tV3;EaBEmGfb*;rQr+v*weXcH{hU4g5 z_hPg0N=i!!4GeCIiKU7^VwxcKB$j;jd^?9itHAU^=T#L9#T=*tA1-q~#qfEWPPmj8 zwj%Dh$|k{nx$Vqh8XWU`=hK3|w#=_J97!fjv;OL@JI^72x$DO$Q*Um$r#FtFjKwX* zV7*3;!Awi$k0wr!7-}(^W~H7-j2i9JBA<(0Ha9IwF3z4VPPDz~^l}=Z?L=G{e?u=F z54=nPRqqsx57Wc^*e-XkLEC!ry#I}?cIc^(uuB3|Cicv9r?OmqS`hzJqx-XSa~yL6 z_SP4Qs*iCeG!)m5&#V7xMHF5TASV&MbstX)ddKYB%d@&SknbBC9n&i-yN`@8_VqjO z`^x#+r%`&p;1^_H@>g4s!qW%E*0ZN0+xngaNlvrqNt8>P4?zp4i*Dv^mf6p@0q@3! zh6dlWd0KG1yxeqccyjUtFkLUhK_75nfat7_5+v}ieBf@(+5Feb#(nJh;s2v2095F; z{hX1WPop_&^clK`ETPA*qUM>OQ{c#xl4bC$y-|7FeD3S&DriL#nIb(8wUbixUq~x0i8;YysFD+B|NEc40_luo@AIv{x%G7 zHu57KAKHD!hE$bo`KE0@Z;4dER_*Gub`1TAvx8F#hbiAS2bA|3acS@p*5pfgNpZ4y zfPf`ykUX)z+>hCAJd+N)osb)aKPJ|nu7YEr~P`^ zH36HcUtkd{(yLEQPCi@9FTMex|I%r6f^~`A`UVK{%aGK+R)dZ~E=qQ0Ce~~pvJ#|k z4;lshOusKa_HH=YcjvnAPMO3$^df;=+9**jgKiSx5I;$yI=kt@8>1!=oYOR_XRmN_ zOXA1yq!mY`*bmO_s-6SJRSCQRf)|bf`}Rd1uljxVC0rjT63Eg<3{KFA(KoGtU z6XZrLeANxB?n1iWYfmyLEoSn*@KBWQf$0bvW_bO7qn&=8Fum9H&jEEX0}$Tr(gAO;K~ zkPZhyUmt>kZbIm_fq0^27SLbuAbt$HqWo>MOCK(x7gr}4=e32@twejeT8nNGL;ga$ zHvEGH_`L&3eT7;unfy_o`QfRlx;L#iRuusFwYIu?JQw;%geKy<0Z2w9g-F!_%MoBx zQ2Ol~ohT5NK>UdQ?wpQnc?+G4G3Qyx$z~gaM#DFqtCLOEC(tpYBm@y)CJg}NM^L;0 z8+$;(xPm7JPH-?S0J5@UGcv>uR9Qwr%)wQ^!FwR{q{iC9V%JTifanUO>1E523HUG& z`rvl+qGA+__cch<+mP^ZBq+GWFLwUUc3N0k;tt~9b^>#Qm{vzN64FZk8u8RE@zB~bcg3GTIYab5E2mx|Dj<|K;gLDpzRj>LPj!jHd+fIw+ zfjF!@;En1x*di`10dfJcH}a2ru^-n<#;E$kNlSpWpwVbiw`~JW&4iI#K&YXwKbW?dRtg8%rA34KyfCFaHU6F<={BKe)?pHDJcepJWaAQe~JFgj^Y5 zmEaXVQ-UZDJRp4r`hWyEQjjCXM@dPEJs(T|IUj?P!5$c1fl?d*-T%w7<3(<6Zi2wE zydA{!h0A1x`Q@gS1%2n=L^%CFaXEgawtvo>9^kUI1!r?lkYzYEKF$OYDfK=*)U9=4 z2dPR}JFTYN7?6F1eN)cef0YK0fX4?Yc#XIxKqx0DbIOZ~I4;dxSJlwaud;i3lKfkb zBgF$uEaJL>!3o$>0q#o9_3y3I|JiPlK;)g2nCPrOkFivye@NuAv3J37n z{@YQ&4oltpM_)1j30P+Vz8?HkR{BPqzK=Ut6+wK|C4&J1S}`j)LhbKy%=KMd!oT$m zz}pOd{3{Ki0g)tZV2A>|8HBrnu`vny?Ew{W)$b{1QZKM5iJ>e81M0)|J6HX<3_<=J z9UUF$paf_Uw!FjpM`qJjyqxBHQ3yu?@c>Z32)l09IV^@iwx6V=q!=3;E7SAvr~u!8 z$lHzz^Wm$%x3n++KK5}Ci@Eafn&1Vv3tin5h#l?Us_0rI@Dq|(R3wA6#e%7V zjdAje{Ktg5$&HMPDmgChXh`;vwDBZZR>5L=4RR$?OgJ$qiR4Br&IB68TCc1TN z8`SYYyj6v?ximw-pVUixq*?7?j5Q14Tmspn;OMl!zyDiV1mV2==YwsNXVy5MWLH%5 z=1B*S6xt`rnRE7i;Rut@la^voA~R6+m<=G?-Q5MA&g=go%#buN%zukGfn3A<_b?}K zg#s)z?BPF;dOIJ8JL^EYFEE#Isr3e4@mR@`-!0%Z;wY2?#uo^&B!;wMwMc&}EdKty z5Q!iEM`Cf+Pl*)>j>N?=$IEGrVB`Uq5JbN+`0b#BXnjBk$}^z1vMjD=jN-aU#2LHl zR0AeF$YG4ty9_aSFcV~pU9Os5+0%%xPca|@Rp};O*tB^w z?vd>)3NsZCT}C8@(-28!^OZ>cF@w8cV{MIA%wby*6i!de%9i33fwHn674tbI7K#b6 zt)YBAuTht+%`?0}$k;+x_nA1WdN8N8d&U=absd2tgK2kN4sz|jS{WDqUx+7fw)hfAC=Yh!G+#Sk$RWU>%-prE_5BAu!2s#g70xo|4xt6U zN?RT*(NO%SzYy!o*nO;q%_F1l#`WI42```DJ!{WWy@7l58UN1tsHSMS0{s$+i_s4~ z74Fky_CM;DDxd6l5H#)MTF`afvv7ai=PP!e-w3dU+)Ha+)j5T zKY4P{s_v98pNVmM=Qp1!K7Oib_fZQST|tM>OSk5|zDMzgFbhpJHN*@vU_@vCtk!Bc zY(LfoVj-k$R1(x#Nn|aDW~XURGHYxp#i}`Np$mG)X>i&keCEOQ+Y{WUBWDUpO=64xxJ|)HPVq|=Bu9sQb1YFdYYJP3%s&T#3@kEf}|41X) z)PB}GKjl+$Rdsiry8gAXhPd2U-6p0XF`)sWkE^vrb0Wq*w>s@y>mfa_Wn5EaW`5?= z6G3iccF_x}ap&rv(Xlq9(kT4SzPXap<7Ts?z2>HdW2uF7P$xcXo)r(d=K0aoi!c7|7`Z4vuSnX5^Ac3Xn<&x}!}jT>vP+Hy}T z*i~~&u1ez#(}WZD*~xFx5+%-MU2CYA(@vcIGuy7M+`}APb<#ZQU11+B$?c(PJi6az zG0*3#{-{ByW9pUimp(qUOFWk?NWF`Lv3cUf22$ZFvOh$b0sas^o`@%sDe**}^!YW# z_eVM+Ryx#kW&VCtAYZNv1mKm1AH(Kt@zly=^D9n7JF!f26thm{ zeoK{S-=BJ8J0%;sT0JV?3blJSHX+7u0*z{7^dsfOnLqW=v2?fHP6?{%Sc@#QblqXj zn+L7Rw+C6ZzAV%(ocO;=2=~tHeJH7#bSbj> zEO)~mtELFWYhL!Ionh-P)_0_uZt+Mp=<;d0JbcuCG~&zV$~W_My7h`jb#?p+ErwV+ zG^yy1zIc_8e&-{{RrZX#vKp5TsI%v$vxu}dpUphNvY%i5aGhbX;a4RYtC-!a#T%kG z_UtbnNzI%KQK@b?+I3S-JlsevR+a`$5P{M(4(@W>wY8t*-(FBjfd0h0>qdI!rHSdj z`du+8{_J0$CZEXMcP=zqaz^sOyQ2MDNg;FrQiu@~F*eAC7^`|?;x*)08wZ`X($em8 zN$Lf53uofZ0OE<1)KX7&2x$W12jUM!YfOqJ@+5aMx=-SEW$w;4at?ekiZeN=sA|HQ zDEIrtP|j!*eJmRNE%sax7EDi`u`PcU)}dDF2d0Ea?W=wFTfSD&hIQw4YqD2`Jm^TS zmsn35__ZT*?~=<&-Swkyo<{8*rHj~7)12|exJNPgN)&6oqN?3gjl=Mk|I2DAEFrTfyGAsD`eCxxC_sA+NSmiz1P83jQa(!L(Q?5=s7dOg zc7DTCo;2I+u`5GGV{fYhXSbZ>bA#C6>zPAoQ8OK3O6_WUJ?lrBK0l-#{!-k;e=%)L z_QY~vA&p8b_eW@&L<*Rc81I21z{9=wB4d_r>l<~+^X1PC;^qFMXuuw^TY*#K-HQV! zV|{uBB6>Um4OLCW)FK=VE(YQi_3YA~59O;}bbnan*yjDm&@i1eXF);R85r1Hcw@kE&CC>LW`;oLk)?Tuy&pXy23om1-X%x)ZWp;%(2DTzVHike zvCdU+5-u_G_@oz;=D#Y?|9R`zxe{m_WXT%G9eGpI(B63Uj;PXggZp(l0=}m42mdt7 zo>8cVLQ=8t&SUZ>kO2AoKjeeKZ?|*~ro_}&9&<0C7i6)!*2u}fswvZ9*3~3(GI3Rm z;i|gqJ$tT7zO!Ln<=0P<8*p$UcCibn<;V+!;~l`3@2H^!*h6#cY3UVL`lK{xD?E%%{nDBn zTcCc)x^&+v#>kB5Mlm<@)|32hFW;*>Ml~Bz#g`b|56Zr`Qt#{6AAxh9dSapkYMX?2 zaVeApP?;K@-p4@g_I{N}A3syo+h*Z4cL?@Dy`TJ+h^*=*0c_RJN4XNp^;iY|)Sv2c znKbfMBiS4Au<7bhkauU*6O3&|-|Qdf5gsr3cafS)>aUtyAJ0^<1eZWwNsUXT=vP{* zJcBlWljFOcN;C%ZIcpw{8mUE=H7PuMgpU8I!o152UB@>0CC4kVWv(g?)2@;z7iqaZ z5E&64-_~S4Vkw-NwV7?Tjm{|Pp>@LzMLK2$wvR_I*=O0{gaq*;xa6eSUW#fDT_?t@#C4YHE3+7xMml*m_1?H3JiDY92|PXer>z+ zsk(og*|M}mdhP+c-kiT3`*X(!EftG?En~X+d|Ae++H9LIK9>!fO@%@dXlnBu4Er@5 z{aC~DXM?LvCM)NYPTyTMF1L+&^Uf3A$!znEx5dqC2bDq|yxJUO?H8KUTQ)!w8R@L4 zZt_s1C-jJv_n^eO3r0PV% zx#}oUBkP9gz(ix^|bM<0=h&OJ4+@HQjN=?0-0fUwK ztbD^|-Y*xIWwlo*T}w^1-QF5-?p@Tzz*(mcygu8#SJH2IQK8?RpY0{1<{GB?{P-yj zfm3&WX9{wBc+@TIAJ(YayXjeDY@77AKeT4KZ0BK+( zw{P9TyPz?FbBo&w9!T^M6jOCBKQ5i0he<*glUS-ofmBDfD}Gg6OZ-Air1WYSXLlnVOy)dkmW5KHf|< zG(PB?Vz+&`@U+?{#Xt2KA2OHUwiOStq$WDf|K#um@Uk8)_A|J`6BGk&mnSUS4;|#U z$<+_gH=l~rH4S;*`}Rx4VGad*TPglX8Y>B}mi^t^BnJjuPI$lFuOui{zMUkyhnIKw z9~x!dcg56nYu=Szmtqm``Y{wgZ4g zZ#_rTIR?Hdm9@3svSVF16-q)4g$HUhefZh!8tcl%iJCVai`PVW7U+gC{E)1>oCp5!auKoe?4`zQ4!i5*nZV z`Sc;fEtBQYqtgZ7>QcLN1?Cb>e_aVtdfEEr6OYAp>+$+rHxXU8PL&(4MaqS%Z`Hrc zE~*e}5pVrc|EBVgQkq7JREq88fmB`1q{x!7n;DM&O}82wP3&_UetKOv^~&+J^=XZ` z`QJr9H_4~9xBqgpNf}bn(>jHDes$^QIz~o)NZZ8bCfa~eFYz>Dd zNQ?F(b*@mL_f-_dpO%aCFGiz)qLFOt(2y>P`HV*?UBqu z2CDj1(+_u8-@n|(tF*=R!RO|XSJhn=qB9D0?lzXaWjB|-3h(0-%20IcVSLKEha{2% zrCvcaX%4mKqXQ4EUr3Y)#t(h;|F{`a1BRu~fjb80i=n$v0BMQVaU)7CBtfgRpj0Dn z+dpw7DoK9VK41h^#CpxZ%`UmU(DE$%KLR0YM{Iqk^5vSCmfnd86mOn}ZOS9A_()MfmtLo+$rG2V2A6#qkuePbl z62aKCW4Y#<=Ee$rR#E$xA13>HI%WCXLJDJkHdGuk$hi}hMG$+7&?v`rVG^gGXh=^$Hh0g1^2fgjYOFfF!3aTFY}C`iF0~hc?Fz zB(pR^_BoXTj2rBJo|q_$wQ314oM>0;7U1GNx=YIwTH&LkGhVy0HDU}_dkRoR6w=2^ zdsc1R>m$(`WBf+>B1h!TlTQVQ2E^V!ie8IKW=!1E`^;zDZSU*yIya3)=63!1;KbJH z?9*g(YYEz@*{$S*qWWTtO$F76=IYC;YOTB?Mf7Y2+_P=1c7TdfXJ^IIYBKLgwKpnn z-^STzV4PYhU9#+ylJ3UAxcRg{zbg&|f0<_5uO~D3fi7U>Y#;sn#`rBI{iOop&+qyx zjcHo=E%$huocZ*|Wy5`yJQ8CX*>&|zT>Leqq4TZIliu{3g?6U0b8b}(^#4*`xnc@R zPLpUf#L#)f&b^7bT#Zy(AJeV->xfbzz+nt3#>~mpm|dKf_9u7#&-{RO=6rwi94996 z0*w`RoEpPGbxZ}ve0dJiSU#CZOB$%|w|FMX?O1iUwY70RBc`(R1qmGirXj%}d&tGklxNMvnh> z$^FQfEn;P+KZpeTXbx zP4D6JIOvVwFZOzxfNy=VrlwYFcSG-nxxFMXpwCe)*UbN+sqi2z{STkO0JYjuNwDRm z4c1?Dlyv&66kQsb6&yEOa&a&rMnt|whGPZ&3Y2f(ym|9hwI4mqlr{|l(DWX5XB9bw z;@dw@Q@xZlP|9=#i25y@XW5;(f`t8~&%CwQFFgC*zW%1n?O*2`)a8v|(TK8rc0H&b zDkdB1Jh&{lokc=WMZgNBI&oAf6&0(tYTmdk5VV$x%4vSOS(X?fD|-oOI0}2{d)&Nv zGggI_)N+$qHe|=9O%~``MjLwf_5k8rH@2@Ei67pny%vwfBH(MpF?lo@&_wLpK?$uPL_1CjYZAx#RCfj24i+D>qf>*`ScdhG4Jw{fpaBr(a{I`dMzGQbIS zQr@nv-&05gqhGwZNx(?5+y4GE1^2M%IQgj!Z6Ti&Qyk9u+VU};`O=i@ch}$>%ItGp z?4xIv=6&C-^td)&4KrFbP#XAQ5Xy-dX1wpmQd?Chm0&0?_hQgYx{c1)AyiPSYilE4 zy#jq=h?YXagjHdsqN1{gLKRw4n`>)r#=6RpHjIyrt=5{zQ0MBl_hulPurMK&#B&(6 z$Pa#Ze{P3Bz`~D3{dgI!s%h4++ld#4Z3Y_9ntY?bj(-VtGR7BmOL*M9`}ylvS`yC{ z6wE689)yI1yn4kU%!we5@w|3{cWO1can4x?eC^IP9|x>qM}aZIp>U&2c+gy-dy3PxwFcKx_WwO?p;f=-D^Bnuz&@qqukTe z+e?&-R2RywTH*eDS1SNBU^!938vnu)`b^kHh9sNUZ|ywU?kB(Z$Pqg%6KEeK;-`m* zF6;pPab}dc3%8p>j^Xz>Lix-;s8TJpRd-*{G|<(xdhJ@Zw^tsc)gM_J%?ETRw5*|y}$Uncm6kY1FBM2 zr@Dmt8>-#FPf}Ek?d{PXzx!ovZ$v%jni6M@YVm6`P;Nx%QUWah;ipk(+&%TfbhyjF z2lx7JZ?~94GJyH&u-uPvzA&pGt=_S_Vn3eY#!VTi zrC$m4jM(b&-@o}OG6eq~!MBErs;H<)^^NX1OlH6=JxITNYTHQvC%m|&m^;^pfjfKb ziSr3t5v=0jH2OJ72PgBrL`YERj6cLO{T;MavE?Hb^ z2}}zAfiPWH_ZMnrHeeLM$MI;K(<@+dn%KQ-mlj$Q!0&-ZRv;$+sdME*=;wEDEkR3= zStuBzQ=i;lwvE~i@xuaro8DQSs}|**aB=)$R6Rrz=;-O8^${*k)ZZ}>=icJNCp~}u zyjH`(`2DRF^3%8>AcppLcWpm!lyAr;*kkArc@(P`1kpwU1-GzA;gF1x{QFwzemd=nY@i6UO|?+HeMe~n{XTd+(gOXv-GH8pk~<{e^A4XtKT)jg*@ z=%6NIO;$)~6-RY|{HOC2`i*GzK!*!%CDGVmnlt$d0l5}mCg_@EHw)F0c8iRKVo*M% zFCCvQd;^}qinVlv0bj8KVmkdz+0zIMmWFxPBi4cDia)NQ532&S?-VfeW~u`!DAu!6 zR)m%Lck~_QXN56Vk&o|fQj$5YJA$i?^sbnLT?Gx8;7`z*pd#7UAPCde(OE$jT{2o+ zHubv?9$;hJ@FhDIC?$DO+DRAYekVGg^CDOZ{yfLQdwO2m>pHmB2fsR>GV7Q4{T>J# z#AaI8=OSutFtIsz?h9<;a`82O#Y)P`%={Ym-TNUBLee6po^1F^^sBvIn2hdXj)>O` zTrxlm-V7*f_akI$wa#22o|Tw06%!jP8CFglSW@2Ijc}fnAPw^)iB}JlB9=b?)t{m* z8+XCk_x9%Oc6m&kOwYS_4`4xc6XYch9)LXzpGW2EwkHNc6K;_z1Ug)4}jRZ`EaJ7-(!cHzGv@V#F;3*U=9HB zLxsrqoo8&W863|%&iyBlVP>al-6W<6x@_$BRnfS2|2`Wh=WK26oK<^~FTqHtX1!_9 z52pv8=tVSj(y&)9qyGZL3;q~^#u|bk_I7m;!c^i<8;l%K^J=f~L29TjN#_h?n(~rU z5sRkP@7%d_mdvZ8f4}KQvMgdw!UcsdoJo=f3|QUk%O`K#m;uL{nBg?8%aEGuuz!5C*ehj0o-U7oJxxd+WkQ3|6Ek`BnBrUcnbNZX+!$% zAzB1MASFtq!Rt(dg2PVI;)6-QoOA(c)ZgVLeuV=};UUW1eDUHk?k0l!ma9k1f1XlP z3xcgA$9?C?+`N2ph)=8?j|;_p7k>PBc*lcVx0dQ}bQV%oRb^vmk0=C6LZ%LA`P{vK zpK8sT45wMd1VpTLa)Ikb;^B$r^99!{sClLX<^;@BQL7%GJTp*2VGg(C2nuZnKaLb8 z0n(|Znx-ZN!kVXA3XvjALd0-gQ>4) z$2ZF@Q0zz|&txL&Uti9L4wR~z8d3#TrREiGGRWJ=qFdKXj|%D(QseV-uqUTZ-5FaE zA02%T%lv{U){_6|eO|wQjl&=tB+w6U04FLoKa-rd;Rrev@!Q#aVQw}<2aK$OMZ`t_ z^I;A+S%?UPgfv01N?FulB~&*wyIa%;s4ca`cljbJL9i^tvrwJTB$B z`ueaF;>_Me*8rG?2xyl+EN&Z~*cwE@`|B{|&83_Jia7@kM5Uyp!2ja=Sr&bXtuD~7 z3uz|7s`06+@qej-c{;<~0vk7NdIMcgyt+C>Q7~AnY9U9abPwX)hYrK5qv*<`SL8!5 zPibl8GVQ6ci8Bbd$#cdo8X+mnBRF!@i)w3mI#e_?bNNhjLpUs8I&a;&RYh5uiaHPl zvNY4sAx;5a<2t@2B%NCTEyx!9-$4<{fr=5Ke5RHvWvHmAcw@L=t4W%g)&>OW*1vm{ z=z>T*g{Q>DB{Y8|Xy!H%3{O&?_u@GdiH5}=CbaF?d^>^hKqd-#!wSYO9C4)Q8#x4) zXmO0lR%nqgmF>ZW7{z5bk5x|ZHRr{1Qe4Sr8L$@YD1!>9fNX3!ch5HCptBrY!& zh_)np2E&vofZPZX+=|<$Q00?m_Rc$&Gug3yUVF3TQw|t`j)^c;P;ZGAHRNkc7Se#c zM-a?{=+KwXpNSSM7!>eiIrx_km?5t3=`nn-y%`Kb*Ap9e?rf~Atfc4@Z(cmn)JbsQ zh-q;of{$F+M?i(lvJ1iS*4=(5o{WQJc0xMiPFK014Pe-3ZU~uK9kdEP;+|Y0Fp$sEy{SR5F4Ss zhML;b!^4{aqP<#(cHuEvE66Y#7a<}JWnWFimk+o#)I5rHMz?0nyYoQzR)Wbnlv8e;@`bfHJll+$<;|(Lu&KF6(J`y|Mp+R82s`QxGO3uA*&umbMvJ7 zemb4=uxLa=QAl`p$0^Gtq&6vdX)GUNwZfiSJ1o-nMXR#S7PG%bMn}N|M97cp58z^V z;WJo&_xu-%Y{ade0GdaTL?N%W4hKx$Xl>5BlFhnsi5{AH@PmxWVA5q0p^C;MZkAzWV8t`e6Lx5){Cnx_!G)_2J zyv)#Q9H>uA1R4U5f9~8lEFe1()M$VQj+HxeX2CMsKGi*V(b{@E%WgD$Wd8l_Rk>3= zvB(X0iwFTl_#2Ci8wdyS`6PdTde;->KYm}qx7^`Ls5DDUyfK53M9H95QmJo`RA|*u$v*VEzlSnu0 z_=&DRbLzMQ%aPSCnsxf)fD-1B)y{MX=0XDyqx!H}o5BcMJUbg(cp4t`efJI=Cxx`bE8n?pQ;>ox|-Z*fFo#$>WA=%PrYcL_sgV04Ot&3=N$-TB7~ zMK0C1Sgn;Miit(oe<}Qp5iUM_sgr5 zuhuYMCE*no-LMYAokAo@dgR6!)S{60f0TpxX6g^4irtv{xSK?J_${~R6lJN_FCr)O z!UW1WGJ={7#pwe-r58fHfS32L0zgi^cI{@EQv&@E5FpTZ_`_q1r-XOEkdQsxKiG}< zcDR2KP_^&QjE2rzl5ZI59_UjQd5U};)*ldiA@pTPB*sNCU7ngdZfzxT_W73d5 zjMOtoih1~>A&zTL55#>?!3kHxaZx9`xLu2;u@I2g^S-`51Ul&SCnhli6Mo9@(>$Yb z;P$HZ5?(khXNXKWZ}ZS91}+L{wd~)=<4~tVeu*wNL`WU*h&GvxKa+MGwI9PBWAkN0 zKMZi$uSos?x?;1?t;mK`I{WVlZ?+xM|JB*)e&-IFGytJnT3X&cqnrH0X=eCLbkdDU zivTHRK+()RJUnP`F}%P*hgTnXxB%Mm6e5v_Xv4y3vl)2%!{~N%YFb(vLB!dIvh}`w zXk3#+f=6zuS+(dCJssJ6dx*eN)Y0C?#s*?j($Xcn1mVxZ8%;uaE`NY=#bLWE6(k@{ z1bT!0{RD?fbn(2j-BRZ&{qp6Uw6yPi15k;HkswU(?lqc-ng6^~bpbJznb~W^>jcK^ z=I)-^IDomSxGQE%i(7B53$PKP;o8-!ArJeJsBqcaGY^O{J#-G|I(fJzg{4J25 zov0*`+)9YfDb+*g;s>UOqld2%E&$J%fi+V)L!+ z#?euYS4QBINzo)FK0W*a60F08+flklV*h9^;PK<#e#JN!{@bWmc8_CfG$QJ;MN{CS zg8WWA43f6qb-=}pft+9;4xvkl*gN_S3xOd?gV`H#Fh8^rnVCB%$u43!;@1G?iQvvc zbc`q%&u&<>tfcblYOArX^U=JpZZk78I2$h_BRdg>6K57*6NIUugr+Z~ykHg67o$+p zPjo8bY4G(Z<{@A}O@jbXES>G_Cg*0S9UL6+Lbl{G@}I?5p&x`m0@5w9b6yyeyL@HS zOweTj88R|5a&nKx%Fr#*VUdBuLO-!=9g$ofb6z~2=1_UGnV*}L&%eSagGQ-u?%)81 zP<#neFQ6isjmRfYa*BwkN5})K%O`8Wnyp&2#~?f0SC>+p zdl)afdQv5r>~W9&yUw*LnvT_#m4}dnV`*EL$a)vIqwbe%P&%F{Yd$l9-f4Ovke(6g z{M((y1cg2l!C9R0HNI*jS|UEsG;MEpa}$w*9#n>0~TTbt%Q?-c#m$LE*kn^|7Cp^nZ@ z*-+6-G1&+fdlBMo0!E_4Lr-po{LlVQSRmsYAGgj>&w=Ls@bd7Z&^Z4Zax{~2Z~%@j zMTQKWbTXw-VDT#tS5#NG%Q8ZWp8N3O3nEDii;OL|Kt-2~01REg6#~}}y^*LCA_@5y zO0&bgS!Xt)IS(^ml4H{`RfnVeo>~t#*_&f-x`_ z(A9{ZQG5?z3%u#wnVTyKRpq>Yj*aV?sAHW;8awTMyyN}WjSLKdetv@gjcubgul@dg0=DR^7f>Y+^fmw> zLOg+`wd_JQ?rXaRXxY0yk{t`bpddUh9800#;^!~a)Ynf#0)DOL2SS}&0K7BR zx^>x>-4z7y0H_o~82|qNDc$lsAU&LC{v5V5W527hf(j0kzs*z_H2?W6b@AwyeNW-uZK|IA?6eKfS zwb|rO2Rr}pKn{zyem@CUbqX8vZ@zIF3A=;UflA?(Tbf((XNp zYi(_94Wt1~2|0E3K9psU8-Y|i*yJbFurFyNwt=7)6p{^MU6ogogRC4#;Z z5e3?;pgEA`<;ONNF)?A(UPD6z`k~>%5*H1oju-0S1DS*2g^0}Y{=3Be(X$(;mL?Nc z{x&rgASMt#gpQ5>WAi*!7TrWr2jT|&e&AjKN3Or3e~e`8omFo*BL^@@A7Q_C9=YR+UF-r zoh7`8rejobv~@u-LODW4QW8lc0S_Xf$EH06p~Or4I(S!+dRuzTPD7jdP5Rr+9riPj*E(;pBCs-Dal=X}vwkT-;a`RPQ2!6*1Lg22?K%wOTGrUvHM=&h|lr87=J? zzWD~mVUVM-LMB9VWE9oKoT9&ho5XXUNStt85!F5fkwxdRVqtM{@md<1oc~&`HX;pz z;6#hrVcba=1Cf75S6M%Kl7i)U&~a|kQIV4RikOg}!AtQAJ0wj^Oz7~A6Dq-4px5Nt z{tMB3kuE^~fP)`J%n0pV`buV*XX!9{xQCdWv1QAaq(>qy!-Y&&T4I?uK z&n1Y-fbI+)4gD$~iFP1DM|283kzfA&sqOt`AJPSX0pmc+`tRSn2pd9bTe%RuYK>(e zXCNo#2!LVY>6viHK>#@9Fm4o$#nAKv--blbL`zkGm&Qa83Y+P_PdIX8!c10HR!T}r zUj7Na0UeS{6iK4ZVU(a@ajyL+cG(5PZ3)6QDKDK!-pGISdH^v%THC|#^e40Vo_TkK zp9WD=q^6FGjwU#z0WLf9Y%yiznV2)ACwNT$9gToxt7y)q!L4cQ>W0%NhBypl;BW#_ zhe-j1gpwzIb0z){gIaLH$+5fEjepStGl5&8xTSaSqQ3qs`esX^MxYg2(1f_|Gp=GE zv~w~35T}5QB-6_-TFNjt*d~H8M6giKp1n)|2jw1A3AQpaI)U0k%yr1fz>J#j?V46U z$ti`)@V|;8If4YC!!SjNpI<5O41*VLF!Tos+AXld=R242xKZ-8A4UQ-KbDX@$+-rx zST5zQ=yJu$f5L!Zl$!b%w?nS{0b0VrpiF_YgIiD7OP8dx5Mi+til)>++eo?Nl*J~6 z!f2MD7MNdn504Qw=WoOpcvE)j*dQ6_cZlUlS_jEzy=drb5oY)*d& zWSkhnVF(hEc5@DH3OrlkQy9Fpn$K@?o0D9I|3z6;J5Zs2=LiB7f$9xlo(`w5urLqL z>sPP-0u1=<1~2i=SUZ?#a*UdHVG*gO?+&laD$)1;qC3~UYDf!yp#e(}(TXOh0}C@% zcz=!WTelQH`@yll6>pxG;G_VyP&CYfO-7Eo^fmx4*xB2+Z-Zt-G!77CCK5e6*@>

    J7?$Wp?Nypj8Jg>bzsw=9k{x=5%iL!x9PTQ zNkNPaJn#rPD!~DKqq@o^>3C%2m-Z3R72uBuc^(jF{|ZHXlqC!y^bBXMzSm4vA$hd8v*VNQF&-Q9klKF>{%--!C9SmOL-D+kiJ?rU3 zU<}kZvIgQ`<9~@_hqXR0U%8U&!s}yfPersYt(gx)$%6Qo^;Ru;OS;GL8)^+xmdYNR z>X=3$mkY>}p$f{DID_kDn4l`gGW)lVm#-~^t^J!dQ%4}Ek(tW>xoqh=r^Kugs@~_z zNQ`l*sWtk^=bE|cI92!Y&{kAdg3cOlT8W%IJ^g#zFec`tfTPF!w-CZ(y|8fOpgxGS zNQcUAJdL;m*#l0|bT1@`fJIcH$qvZf7WzFvABo?kcQAk)ND+l(k~j-I2gScGKJBT$ z|6!mht3-AKHi-!XDmX9*&C$WWK57~oGOq5Xc;m(mC#M{~R9pkdKN0$N*oqT?DW>E* zEQP={=D<-2qga}}GKXWspuCm_@*`6;bB071j>VlS2p=Yld7adX0(wp<29T)g2Z_boDPk-4PFh0Z{(}c}w)cTZr6Kc;KChwSgtQ;z zfKyOdw!D^f!NedMK!fDpE#;ArmHjh+`TPsH@bGZ*34wo5j;$8^^e|6>#W7L?0zJsh ztgL>6W0i)9upI6DUt0mjWh4r9aW5m=I{m8!%W9ZOKYU5d+3<^I(*RS;edJq{^ZZWPJi765KNEOJTA<9-n zCynhZn7d}MY++21QEsK9%K=aavw>}iL8)A#Sr`+U5>TMITCop1Ku^JH!o*!xuU@VD zfR|(&6P}p8G;fGV9O3T~juJJhn;UnDh}^)*2KC0o)N~$oZKCY*`ZW(`_4&Tnnm}Cw z-Yk@C?&wcWT<0Nt0pM8>P!doOlkfs7aQUX@9lVqt4!BqJ` zP90i{lS}z9I6n>|A&99;pB+)q+82OT+ise*+jkA6+zmS7X?%YL!hhiTV1vCQLmN7) zZvxdJ{#pW1o`$0etGe*cGCJ?`W{Ri&^-DLcKKoiNSJN!DmhO|{cI?@$bJ#DP#rjrD zz!B@d3s6qOvw=M=PxZf%r(b*`oCJ)YSvOyjXG&0 z3Sy1Qj;5M4vT>Wf>t==tDQiP*05i^yg56S|YFwYB<4$r3(Jgy^91!V~eR#Dcq+^YYMBtLaMzb}wK7lY$Lqx4-DBChL1ucGHr|mU+_io|yA%2O)x-6P2<2dz#|) z>!UW!jx2&E8SQjnwB6LWg120MbgaxvuxSFULn6T;8jSE5V60U7$dAUz-j_I8E?i0% zE)aQ_v-3QT(@l+QEqOO{XWF@c+s`68L`nS;K|_0D8k(SC;Bhl2vBq1sa3G|rLVsgh-QC9XE z)3-BtLEN+3rT#pTsOB1{H8P*8z_4F3PGiYb8n>B@Il44GYz`z|nt1IeK|$;f?XK|! z;w{&XaRU6xohabOs1joKTH(eemdK_L*j>bE^M(ev;jZ?7@BO~j>{omH7cbmr2Rpkp zV@CuVEh`&snmHftLiK(w@8+_lq<8txkKaZCc;7knSHg~JEQX_HTP!RrWI{#bE4#t| zTrm}OK1{qM@@n5Xrpc09v|eP$B{$a@I6vlh zkz5imt{_~I?cq*g{p26G#o2%Rm`J4O*DuD`JbAJygBJwXn{$X@OAzEo$P@fIVC{G; z9jVOKA6DZ(CtIi%)Q9$VypbEtPPQ`BD|~^l{OcL`qQb*yVMKsrv z&W;-4HH3qwQ|nDOcP0M1T(FAKb{8Z)ketAw9YwT7a{1O7VD%E^df1z5i`G&|61{_2 zQfJS~MJq0y^*`|<9k!$WeF2~>#B@wzPU@s%ERZgK;)k3LS}6; zG=m9qF!*Kg4zwE!3kyMvHG*>?kojW<418D09_>E4?;H*_{L0v3Z-k3X5-I_90sj7H z?N-|2rj_{fwjm$2KD+~I%^a$RD@c~iClbsLQvF zw6x*!(K~29;ar)B-4oA4>#Z_CfyQ#UDtxqvQ@8 zD0N|fvWj6ab5In{4@j2~H@7V5*qOEK*7?pq!PG+V6|jf(i*{6~0+EnwrpWEk&`?ss zDaLDPz)Fyho8+f~jC2Q$g5$A|;h6(F zHXkVS{=R-IEuD&Lk71nElW=O8&*_~tJmN(OY8dVQb)5sdtM%8aZQbY5T2u?_2ee~~ zZdamUvGhvl<$Sl<5x-oNevM5`Ou#B5cR@YaylvY>px8*-{QZee#&+E3d@4B6h)QS<^_=vCr&L7ZxUseV=X(xSJkIFJfc~%!|a*! z*=ySmxgI;Y>XD(1_|)2$I6%})=@exe_qzH zcPrJeS;yHLr@Uz5)^4iIEneNPVCO1&!|>&c8?zOHHP7r!54_#I`=-PT>Ib{bmXeOH zYu{#ebmck*K6DJA!XV|eH(1nh3ddo5T)#+*v@@@?R5r0ThyVjr%gq8n-^yyVnPae_Kw+D|$3=a`MBIic& z)KZne-D4ptO?E$Pv<4p~{7` zyp)un2Ff#ri*R$Zpj6s*>FfkbyQHHJ4?4954LXiEtGvq*Rm<*fzW76v!;#~ByVJKP z89k#v?QhPVia8W`O?_Qsdi0ub=n{VtGHW3SWHYm~(4J&>xrQ+ej5mhfBUVg$`4S~V ziWxL(L~Pt1@kJL62FRYq7|GY#MI(>) zDblg^`yZ}v|MhEi(V^#isp)xIZmwcnO+|Gus01`xbD?0oNGVL`(q+r`pL;S-h^&q= zN*y7Pxv@<*leTTP4A8uA=e3V-t$^p+q=6rAxY8t+i-$K_8OUqZZ?E3s?46OX$a1T& z_+_z_O z4Id<84<0_ao0Z@0XZj;jIoCS=%)0A_A+hI~Ht=>Q2!C3Be=6|F(fXW51tTGKmK8FN zeGx~JLh|{md^Vs-WP(J$Kt5zCXjcnQOO}~9>jMdJV7uJ6|yd2mf-f5Q_JV^NwD;77}C$-Io&#^FE zPOmuIw|$_5n~RlAP;I}J*T_#DCi@S!pJuP39ekp4fGw@A`uO>KZHHckpIXIM%9OT?<(k2Zq$0LkdkQ0( zgsy(z_j~-BiR+QSGdJKg>OO`{~L(>VvV3!Z%za8uqM^F}!*^JO0E zTW{}DqW^#6yud}d@mCo(`VD{+`l$I!(zi^MtyA@?fcD6 zqjijR$eA>HgmAIz#;ZpA0 z%oK^Yr@e>AWBdxEU3GsXTbBfm_8z{U7SAnOsa#zk9YSl*%ywQSYdFy~?z?LB&+(F_ zyK~g{`fS_=mF!<*pVy9(tlK&sihZ%Vzk!uiM)glqW+?6VZ=w-YZ;Mn;7$)7i@TW;r zHBQUyl9%ts>wa90hR;%8jMS%vs0G^PxVs$deV+ieaN zt2N$a3w5;~2^F)yEbp=QyPQh8e)o2E3uDfj;`$`J?1uU9=}^0?-(Cc^B>hRX#O;fwY~BT^^fvn8?`p7_toru+c7trbVfRn)0sKj zE3-9tgeLSW{{>fOHeLa~vuy@~ep|MCD)06+wd-Fe-8JH>lb{%Ne4D**{pZ~`@yt^4Lr}%6cY)s-ztk8JZF{$!#BrR0J=sNYZ zVZw#RCYulkn``m0+nyaZ?zuD3wJSybO?6d`GM!%^>tW%pgetF~Ln`TPd`3U)kL9Gs zq_b8QK8r0gOQk*D7-GZdAe3CuSo44}RMcvQK7v!s#-T>W`_e%@F2|p*t$qlP>eQ#- zSFU8?HvN)l5pzak(s3f@>VqGTFXqCPdF zCfvAWweiP~YrI(suJ+u&{wB4Z)_J&4vVHe&<3EyY^+O{%rkxp65fA+BYu|<+zVy6z z&o4cvjK&K=qcvf&^nN+BeW&6{WrKj=$`nE_LpgkD*I|J6VCFO<4@cP-Z1hTm^2Z|N z*O1bg(Mz#xCAHb%9R%=S_TnTFq*7WopZi4#>b2W<>@2(EmVTyd*`9;?g>lY_ifqjx z7e!6K7KHVU|CCm1aHZPwO2%fr{;Fo7K|xk`LH{>YMo(5V2pY7Sa@$FCD_{F+$60*YvSGgj*97KS$8FsjJd8RBuq$Mm8mkHc5rm7cQ+<5e`P z`|mVokIr~1?;R*|c_bWqV=n2O0++w^F2{~O*U?3(~zR#sR;_`*?DDGqfvO zjYADG3fru*7iR2qjP5eQYt!Mlw7GMI{i^WQ#E{9uPB%Q zxV`Y1sou3>Gaf7@Ekdq--CXZhH6L(`?itblJr;3QSx(B{&fCPzq+y=`gU0vLo*+_6b7I1JI<(i`03XK*5=O6rKhirIYa6S6=~mA_+V$O z+=1m~FS4?;4N3yh(W$x3s}SaT%rAE)LRH#v7pWHIry?yLA3wj%R}e1IH4y>eF>bBK z?4Cyc_JLOdyN?chm7aRBS$|)KXj!VR>4CV+D<1Ejd<{+@-LV*Z>-qWGWR)7@N5|-b z?^$M@E9t&Hv{@HfFP*9Ov?Wn6Ub8pH+HUw`WbIeKbcZ>i<`k(=w`1=;vp2uz;q!Tae&65icmD8~bIx_H>vg@Z*X#Lw zJnj#todpWhJ|f)fTsos${e2{4dml$n`V-#nbBklE+bN2ft`>}_NTxrC&@96hkBPX} z<0#SY)3LK{hW)2V$KPvOr2;z- zL#_&~xg`rnJWZC0VCQK(Rui++@9|{g5qRD%VBm+MGdaRS>h>MYs>uITP|mA@^A5C| zeZ4nMwns?yD@Wh$qwL#V775y=rqe(B&`{djne?(f)>TdyF?(rUw47G*>IMfkRi37O zPDY@fzj;Q~FH*)py*2|5l;^|yOnn@mn3wz14ojX6dky|}-u+(6Z9DiQRXIo;t5#3a z&oVa5P}I_NDPG%jaak;$+epx2T8NGf{qcB(s%EZAfi` zDeHUepi8B4P|hN^Oq*J$vDKQ8fUTa%t0n#vX{tHts+T^y%7x$9I;2YJ@9sS`C(>=n zTp(|to(`j9njg+1M#l6y#T2MSEDf@^jmZ}1E2J@x#!H8D-Q5unQhdQtS2mn>O?Y_A zP!8F*LqHMxZJxDOQZNftLAEilQ6TkpK?&l-p81c=>*r_SzWT?FaB6E;d~GB zw-p()o@mo~{W-M_CH>BG#DX7ZzaFgrLR@@fvnPSR=|MJ>Tc;WV+~6vp ztD3)h2?z)_x3>#mbIv1BgNbmwm>dQ75yJ7N`pN9}zA0}OMbZ36{S@!~Yb>2~t=Dgq zv*guRar0pTMCJoET4JjSgMizQrN?%@wFP5X;5ulO-x2-vsPpqdGf@;u zWwmuMKcl!$S-${<;%SRYe``yW#Iv{}ug+(5N~++UOn1@ko9dmAtar|GCz^Tstx>$J z;!Z3n{jXTkZNivwzewR5WwZ&tdaf31WSsTvWve$*UjEm|3!96dEaStXNrkBYNjK#P z%~r&Ax}4;PYdoduH>WBVNZWZ?*}Br{1lM+W??JVShPqYN$G1<#EKM~Z2$*W5YfXpZ z%c(0GlCGM3tVsS*AMNo;+Sh=&A;DkVPj>Z@MvuBfMopHtV!UP8yCMdTcn^HJnf$(& zi&W2VdVUvlvxUt)#0NEbK0_BD=HJ?nI-~>y$UJ>o4dP!;Sn1z-=7ozWrra32bw||y z5v3oLP@xHz2Su}nbA{U0^TDs5ev#d&YeH!D&j#MLqe|6cZmr13_vM}p^I5Q|0$!;=pOy% z8jGSHtXw~G|N5DIO2B#OCUt+HD*qjBX`0*usv+o}tW zB35ICY^7s&H!l$?ErnK=-$rqCy2~yF$xls9-mb}~M5?XK(Nh{#y679~CC%fy4vMvx z?^-UF%2f8l3T~NFJj|qicx3|>5kgEs;F|Y0=7U`wVo^BR2KHARaCc3tVrq~Qa^8#9v zoYxi-B8mi#G=GQGFl=ClWLGHBOS19S54(^jvgw-o6ITL9 zY_&X0qrM$WR$LMO6<##6znfAKzv4sN%d@CX{YPeHA&hQnC4|kdY`OM2;vgp~>bgTX zP|O{0Y;yARA-kl?bf^&8oI90=t3b%bL=1|aa2eNrB8J8Rl#3HZ{LtlIlk+8oIeFB5 z+>*$;Y)y#p`2&)59= zvshWaHKvPlzpClGH~9!n-(pvzz8pO_`q^|>=S0W+cz3ad$TfPJCNX=sBq^2eqC&dR zg^^vw-U#LuQ)I_0WTs7fQd}SQ?w|XgW0lv!aTAieG*A|Gy}i{>S4}+LKIqF2m;Q;O zXpIqL#4OOK6MS-bl7AyBx5ySNEvsqxU^1;h4W36eGkwgv1cFsI(|?Hb29{@L?h>&O z#5IGF2SkI&oSbziaXki>Kf>M4$JaMF?7VG5b1t0jxMhlvz_9GrOG+OW%_u0;x4SH0YV-DM#%s1lRTZ=h#ctKf>}n23;Kr`HsF!^0F`hs zt{%9Z%(+RTglnhXq3%Uhff+geKJTNy4t;=znreUo5t81qlRIp z;5t*MFQ@%;wWgJ<^C=}|)>UN-DP%{?d;GQGC5T6aBm zY=wI@-oods9FeqFBL^4AtGE{N`?3YbO_pBZ<;iP)lzF~C$6TnSnX65>{XDB*W!0fKq^8FT(7*t#RJC>V|GY(1G8=tdEg`##+t~^}#--z%p+1B8(JgE`Osu_A1R=I9jz5ht? zS!G ziN9(-K1`sUsnUS;!r=t+hFjE9QzFBZe0{#TmA`wcVZg{#1gARp-`%j$li9!fd4=T2 z<7Q*+QlV>2e+CFrK5TXHdS3BJvv%XDCVl*KuJW*_%a6RR;5FOO`o2$lMUcyCwR^ue zxuhNT&1DBS*ZrF1O%jM=45~E$EST{4m2yp`|QIxM4J&|r@3wK_`7Ab z%U27M)QtvL!#E93o{?yZS}i*qer`C~kQxwPOD)}Ku_tdY)$ioNrVb;->~MP5#O_PO zci=^3c15rV65ygajDe?y0woYI5U~;Y&lH1~F1|f$A30Z^ZEIq$L~-jxahhj1UDbNx zDB}2bXllYWKIlfw&(#4_ve&KZ>&~NJ*ol|}W3={B;VAyu6><*KMv{**EF)SvT$%5j zSH0C-Qk5BLjdep>*~hN~K|Ln=N>7^3BFfDO8S~sgD|$PO6fLSs>iK|^b~j@{ zbNy+EO*?WsvgjJm+w{w6eZ?zlY^H~0TV9?#teUqn z@^MJn;F%DnCaP9TMK`C3TqBxy;Ye2nM`o`zZ})r5ao4?&Y`2{gznGA_#!5-O3Ijor9357x(>_DPU8obv*1 zuU71ifSc4U>-%4n(Zt0sl!F{g!? zYxbhiee9^(XuYr5Vj4mg&Fp`>r5$Hih+<;D7=U!&;PnF}ut##x%M3d-ryP%uQ2FOC=Y z|0wSNDL9j)oSn`1ZOkM@nfTg2#9`mB*wy_Jt0>4UuQhmtVz!N!+i1&9FnZylE9RP$ zR?o!fxz#C093MXzqp_24K`9)6WxG&rJqX9p8!tcUWR5Z;Nr>+;gDm;24x&Fg0;k7g zrz~7F4%6SzJAyd+G$c|JbSboAkW(j#jvlZuM3bkgU!k>N04&3DUkw7p5v0x0!DBuA zp)kVWBO@q?LD7&9&!n&Ev6QtA8-aRTortGoR{|eBpcUKFa~+(o9ZDn-)|bb@5>0N@ zRuFEG!)Zu*TbuN0v{{CVjNYwa_bTRl9eV7BFFOQZwFlI*S3(GJ{i3frax`Feh#t=g zoPG;~jS6k6I}P|Q4Zs-y5|n+okkbNzh6oD^P&(pxRvSF}2VJ0fJB*17N3=!Q$7hTp zEnD3T2O8czNB4!7yXmya&xY2M0&vg?h&dEd*b0 ziDV6M$9R({5pf(hZp1=b?FZHWwAvN_({OuZiyeEI1EghXH zz>`disHsID91G*0c{|R=`~kbkDQGvJV<3>vg@w^&P=iNwcehucT);rMTsxDH#gQxE zi1<>NW__DZzI}hO?W0e!AKnjM5J*QS#KrwuumMb_!=5ODFNVJRr(P@NLqY`W-P^Z+ zL$}?8QzGz(;zk)QB98uCEV1{Xlz(*4KRFNcWNy(4lYM=BpSYoA#C za>qd1u7MBDf*2Qy$$bV6HE0#k{HvR=zy{;xw=mefj*dFVcxib_UT|-|#$gm(&I|4X z9i2p2GtEtq|3|<#ZUuZ}uulucqltHdY}1X8QWewT!obJkAs!pU*MUIA*u;d_X8tSs zo&VY?G(arlUw$1RW%6lP1=oJ`8r^s|ju9fh4K48?Lqta)$XU)M0S^C)QNmzpVw&?3 zicM#I+)>HI&ad#V5oiC}e=aC`{lEX-iak9N)f%cz&EpvrlHaQS%TY8Gx7_lMX0Ozi}0Kp4~iJ2F!kSG6FdnXU)aBX&p8JJ_!dThxH9 zn4O)y@e$$}xL^utikH0;LB0czo}ho_6oX&G!sr7w7Pk-1W{vhIVTQo}yVZ?&)srgs z{x)ccWBwJ=S>%A&1%9m-T6lmP7l4n8d$SwytA>5)L#F{+fI|E4(fh!Th~v>CSOGEs zDHGumM&ErOSRbGO2G#Bt$qOCk0AWg!H?##?+19oHy$wA-5C`iEo?5VW%PK32BxB4M zm6w-8A07KlbMimP1hynyH?-iv+#dTvZMgm)nt8mcoZm`HYO!u9y~iDyv6FL9Lwe1%BaH z#xG@@bb3^DO-JPw?_eg%KTM{_D%6UP=X4v|fwuU=#KxDpK_6t(IXu)fEeu`^e=y8T zq1Bj0YH63)H?*w`@Q8m&H$vK_$Qr$Wwq^ObNIHVs;Z4+z{7^5c$&s^6$TW8#r{kor z?_*naV=-@gla_m_q#nczyTcUJHxXO-${Bzpf;-9htj!_=NRJC3DTBKmv^~Jzvjly} z2G@H;hdGq?e6K(0GI$gk5^|T04p5Gch;!YaJB`dkU8GuG4eRqCj>Ewpei10(z_QlO zaA9l{NpAMKsP~VyQqn$8QRcwkC8Dr)AG2sFgVa{7{V)Ny*Tt_C8kH4T-6#rO*Inm7 zjFoR~Ek7{DV)zlkx5U5Sv%K$_EBKrx5~ZwFZ8p5KPd{WLL~(<6VrL$agy^7xJmaQj zU>Jc1!`)R6gpP3Rzra2RuF6GFUxE7{u|*HUDD=4D`;XzpU6SG&B^9|LaW%0=g;9Tn zUV}D`ldFVZ8IqdsiOU-ko3GL=CLwiqg=g;WW9yYM55KYxkuPTGkX6}OXKqI_Sg3R> zEl$dl)YsPeq^J7j;9Iop)eXK;m$FGZfs?cs{mSQ2zb zP%|>`?tK=!>pp4@?jKP2K`rM7w@oVIAOmci&~_V#-miD{tH0~(;lTGrbf8&1Ks#w_ z5>M>BMox})_?w*OCmT8QJa)S~v z;?GP^+R8_;Kh`rj*t#e#A}L7Zr}G|HwL7g|@9q<86dh2Il~>e|SJf@ZD2Ts>CpI|{Jj zMC83hl6@n&UAyi|uW)C-c9|Kaj2#>M(5!{2@JmPVfPE6yp=KRby-Vdr2xGAJH}Q-gm%m)&dcE;OabJ?}Fu(aQ z%3eh%l73Wrhtj&DjF>V7Rp^#?5V&nj7?_b1dsyZj#H!?2&1Nu}FeZ8LWoZplCJ#zi zCAq~+XA;l0U(-6#I5aEUYE`v@&h0n5VX!)eV;(#$0W1TU?)I0LJK7x{5~5)#PoH`K z&jkz|*zT@30lu@;hXfWYx6463_%~sKRX>*t9(+1zG9Pu=A8E0gtCMjq5iy`Y-A^Q^ z)8uVq;&@3Q;5y?+T2h!7aBat=`Kq~bJ>Oo71c5y3b-hp{&WiPpuIq=`P7JMMCAxk` z+O{pjbK37rCWUpThom86Hy4c;4*qlHmL2-GMzt$dZb)(LYV!f8iAc;1yfglz`MIy{ddCt$J`W$2rI4 zI7KM(Uj!!vtB&wYB zk!vHtW^LFFH~E`u3X&04njSc)*TWLV(a+5Ctz~Sn)pLzC33VrZRTF!|RJ);Z)`vCX z&&DG3@a<}56Rw|$3vLqk#Kito-n2^L&JYn-yLiM62N!y*I38F=HZ=q49VR;rL0mtP zsc@$^kctis{-P1PNJIxJO)(ufN5HxTAi_VcuunS|K+h8_`k?sA_Kux*WlOkeN0p9l zO4A!j*pXk`AdpmUWEDde=0x_Dm+Q)uC{SDtx=Q{}X0Q>H#0Jqz+Sfyy3)&}?+oMYg zo@!a6uZl|TvU}?8*87V|UL#2$R+dm*AEh4JU|LBsj&T#d=XRJ@%h}+y0yU=GgkQHc zxVA*Gj6(_N?^((NdM2 zl&)Z?mwa)#Qs_W%MQyIRh4PkfNmAe8ouQXwG8DY8M>GFK$uvdR8KaiJZwNX5gvT?W zsRvCs9sY0v3<~j>`TqUCpitZcMKYv%txQb?fjSzoWee~Mfa!=Negl7J+|c>W>{xa1 zo9J{>jNO2azhVoT%lexbMfxq|S?Nc0dYz@$s|sual-Solt7u1%II+se7U|dN$Yf_< zEu=Os3X}Ry?&7+vOmW3(E~s6A(@@-`8DnHCJdeEXbT+ABZ1gn;QoBI6Zz!*3@ET3w zH2J3vMw!$71)la+=T6zp72^N^MKd!MA%`9a+?sN3`&H@Gdl;Ve$dgrPVTetNmt6z^ z=Y^?k`$Lm;vDXc(Z)$Lt?~e$Sbzh3#E#GJ(DT-N}-lCvTQ0Cn~a@1Mb5vYph)mw_S$coPgr16ti~n%G%xSr8GNhf4@p@BjiJR-pM1J1P4NBjo7)K4) zC)_)fgI|bnRir-6=|`1#b?ftL6z#VA#A<7^GF3UB*5Ku|`zzRYD2G`nnSVbB3>ed# z5@$DPsLDV3=$psR@wP={?}z17UO-i(IVtt+o6|d=zby$chK)UTFz3sCwZuo%9vu^P zmE5L;LWFYJyLrTcg7_+FQpM!Zt(%WH6{A&k$mOLR{AdSVPqAv-)bBsp++R^%+6mxl zq1lb>&xoGH5|lyj3)o2h`6sw3L#hhHMdRF@$A7q-2Tap86pHKaU17vL+Ncev_}7mp zf%==KhJ~2dI$tBK(S?|v$kN1;2@w~~_mUZUS@%?1=9v2b247=Ko;B+=2g2)IYdMUR zDCuSbzdiz^TUh>)56m5@4%Z1Pj&p_CK`u6AqYc6WLmA#Z0fml5YJ}QotTf2Gk z`&mbWm&?3v8w}>jwBweI_li3#hif;K_?xsHyit*Q)?jLjx8hdQGo96JQ$epZMRu%b zG><*?9z(7>^xl`XpYgxl`yG7rgrf(dSx%enV_ml5w}!e>Mk~fqK4W{Ei{?IygACI0 z3Tys#3`*65b~ew`2)BOlFBg_dRJmX(%<)I1ex<5s6Bw0q*fwWJjI4lL2W_oc-KWEH zR|e|_xA`dVN1+_p9UmPzW5mgHmx=|x<%|TS)X0}JS9M1-hvLeB;vyx#Qc6ptmD29V zhPN$}1{_oF+>fN^D5pa1c_UY@2P@%-CKWei8gKfUj@oCxh2t~hCkDoaJNvy}Qt#Q8^XKWXbt4Rm_P zac7Tpu!mwwh>dbb*}HCmPiMF~jhzmsXgza|lG2!ak+SJxL}<_A7p&+0?VqeQuQ;vm z+&1<;KcO$1?Z`^)GNu2#xMUPEChOH$VPyaM6Eptx-G1*I6%Xz6(`P*GGD;rXA0C8A zkdLyvH4gqR4Q&_WTnfkadl<*Z^_m%$8ft+BZLSOCJ!}t;s|~Y?^7+=L3O;UQ3bHB4 z%HCuH#}dUAK|j(sQ|@T9Gn396u%7|#8dETqjGCH-4Yq(_v>EwwUJ{d5DVp;0ju~@J z^Uz5fA5m2pHQoMsdto8OEt>VX&A_b0Ni+j|nfWHGb%z zsX{fO?7pglZu1|I3pFQ^w>t_)z*x7nz zD{68+h}C_VOtVB}O;Vx9nKq;Ona{3M?nuboWwAYFBE8-x7OWlz)iPE^DH?XGWvIuK(v)Q9eoD;eW{a@uRpZow_PFLfnO zxscfXuR`$+#inDEb@6k8j+#vw-!&PA=s_gD)(zu%Jgl{%(*7M6b4se0WODkzJS zI#k)Jnuf2){K4Xj|Bp+!{!&3z$!+iU>=WX%m;~&b50s4Uil_K@F1Vd;E`A`bvJAT~ zm);X&PTDBf9V7ekt$_8n;6=o?E<3*KQxfu)PfeOS!Iy4ik4By>B0Vq#*T8FlAr=UR zQ{Yu#otM5LPJI_hG$sSNbubgNvPvW7eE@`kF8Y8!2%xmD_x?`6#>nj4Z98w{@tLTi zz`4E-qI3F2Vk-~DJf3}cTfSAhQcWK4g}LC?3KQ+k=&}b|j9DA5e=}4o;*CbS!c?ZY zHB1Zm89De_SszMfbf)*Zx+cWW4%vt5DSuhLJ+QJu9b*1C^1)Aa77pbh;dV9S!HPud zK{wLuB}y$Ts}I%z;--q8#f3=Lfs37%!72)OrzrQZy?#h(f9KQ{=^K)%${z^1`C#>^ z(8@J4tllfs|N}jzM3u)O2y%)6?C_&fu?p;%8Sm8h>k3i!+ z>3836(Zq%QX>Sh;LsOFlc_8Y`X{}-x-InDv9h{z;f+(dsh%I60%nz2?o2#m#O>=@= z=fzQ*O82i1KHoyp6!e(~iHcG;{rdBx_QSX_^_UV99|;$3<07T++K<FZ0SAUU?|Dq)jt{cpE81K?6TI1d4$oB1Tmn>KukB{}&DEM7g3fw^+fR%ZxeIw~ z?iP#F@jfz}8_HT>F=`F5n~1R$q6$e>S-ltXq2}eQ^w@O5;$c==W6OAYg9N`XgqMpa zZ+X(S#Znh-uGIxDO!by^X^Rmf9LvF-d?p*p1y~-Y;2}`qM&az79?c1WC!N9V5K7Q^ z2p4G^n=SCHONfh`RUO-bl!I#>oGMTRVi!R2_iGANq2&WNxxE$hQu0MkvLDV)2p4%5 zP;L~VzArxamh~1{(0Dcxn3S4}Y{DSyYhveW*8ZvV>gyq$Q(01GA)f;7pWy1xMeQ+9 zrV_G!KYvr9`B^D~av(&8n6~@hRQ1!_4e1iOn89O^u|yVU86x<#H#C3Us&QFKBx9b^ zbuq4k{qD6NUhFdr8riPFJXUgf))FT(V`YQ|14H;Dk%ocEwT1flcvY!}1S#{=UM_PH z?F6NZycinZKvi;}AR(DrTVq4ytAnW=i0RYyK>h`zC=`#Lph55duX?D|c}=E5<5lgC z$GhX}9xOafLr2Q!X28j3dg`HE(==pvp9BQNg}pmn}c0Z)g% z8?hw@H$h<7dGo@C}-n6WUWG(==C{tYK-(E3KzNbuOA|KY)V+Z$$u zpDzD%<%wyi(~lCpp$64@f{PVYM)sLn`X){pJSifjlpK))TH4MF@f(g#`n^AkbC)W- zc80fAkb(W3H4~(%fz#ZQ7UXms_IBys`ySo1 zYCJ>E?J%L~SaOThL_>S|3bX({t1SRy04}TtkBc=G$S{guK#(Cob^t;Muoi7qRYt@X z3+ziax=2o7Q=oFIjh57^nCaV5sJLi*0>$SJqsHiJwiGj?- zhi&t!s;s@`cg~HjuiW7T!8Qa7;Cx^&i9ZoJ4-JZQ6a+kYn$qW`EoKKbp)P@;!}5qaimKw5z7~iSqoEcL{FyQdAu;S(|5O zciHqv!iv7lTOu~b7WMw_np=>0CenPr6cwglak!zw%okg+*tK?J%qls+J(YruPhf!0 z(Xi23w^PS$KRI>Qc{C|L%V>9yi;VxIvMlh!?bT}im?#iuzUMY;8~&VSZ8KPwNrvri z92gtR#)fW{AT%;SnhOvLD7YWDpB@Dan2;e3u2g8Pt*yDCA3<(Ma?od>J~LmskG8SV zPA!l(8EI)LtY!}Eto}A-M!N`Cznt1Ck!A3PBBf(&TbPthOcpVN-itHthd?encFRu@!`eu_7w3={zK`{7;2JFSCQgNfie;+*v2s~Z>4-v zKresN`rXvLTw(mGctLKo=GWSpQSCs7h7|L)y&jiW1+0%>tV9?l?4-*rkN>w8pkatL zQk$afU^OXBBEsOA!k;IUrn=GOm^jUFva#|}yOO1>RfMg6wjA4C%~uZho!JQEX}$S@lXarh{1Zyr z@t}fK##ewHlXZY8WBAVc(?%i|m$ib+m7UrYyz7k|+9${DGGOfKWEgqYV@aY{k?-8w zz&1>Ik9YH8>V+P1zYokc$^4rWDcQVskR-iCA=(=5GcI8Q0XiS4Wy2K9WkaIu#-qyl7#BHu)$&`ieGKN zujNl6Yxa?}({c^lsH>HD?Z6sQO2Ljs#8cuZ2J^B-q_MM8M#J!SN6)Hr6-IF``r!e} zzwxQ5I|z-~;^JZ;a6sn^ZK(q~6$mzX^Q5l>NCrK?%VBOo8;nbpW~wze9{&HR4i?&Y zmYuwDmR>ttt@^)$fM5P)#8uU5c5y%d7<}s~olpHx%wS4S^hMidYtUAP?Onz%Mxj~P zz3Sf&!mjHVqOOKrrD)}_(|rYYMW)-OiF%Lcuc%lOCgt|}+_001xU$|ay2f`s$tG6q zwQEj^bQ&a3`T1YM+geIRmdcJ~-3=(utzpWyXVKgg5_g)!+^)O9>TDyGbY&5*-ECub z#Z=m6OjcW6L6@^s%bxh!4t|^yVC#~C^9s{j<9B<6+3{mG-PjJKQ=e`mPd=mTqD7}uG@4KBaZZ3Tz)_%VEQuZ zDC9Qx=U7he^_}_Woh+0_n}uX$tIH^nIxV}Os?^uR{uy>4pWxO+aRke(^!Yxv49POh zXp>j?d3TKIRE3y2#az2jB%P5V)20aK0@iJcbZtS|ml_=$mHWp%8(y#ZX8*yM$AplM zjfcFvQI9TkctFE;My{RDX$R7OPv6bQd+(iK9w6A*vH^w+wGZI4CnqI={-l(w2!U2z z%^POw*3H-wB5AUiRX2CM_p=rdHVf<|nAnbm2kEzUOqarPz=4>(HdCC8(K*U(EF|lM z&(K3vxVEzrk-21+4-!l4(< zz|0ICRGTMEfE7h|gcn)D&3SX`DIaXdPz3#B>)|Ti`6`>SUA4$jBFFfehC#}a0kNs1DnXN%;a8Db@6&ZmB}MD zV`gLO7}1_IYklF3t_$;sjjt`RNdrjVyLZKij!rbGOH3Z}>(E4zaZjRyePU58$0AT+ zB*{BK;w`v1z$E?xZkdngYpwZj%gM;t%ChZlpBtBBHp(E!#p!M ze(DV;kZVOySnY?&<%W^pxZfLfoaneUjEbpZ_YI>F6lJ(~FF5Q>>)!DMvJYv+Af6I1 zZ|JN!-m6TAEj4g`gH#z*tFZ7Xp9xim4ZY?$u-~1A?&eO;j9YAQneNE!?xp>8`Hj&Tkhn>An(N(BTsOu;!YRFnb>4XA_m zut5U|;v@*0%nb~#%s+-6NpQn=><%U9b;EKT;y=;nsq|bIg5dMle)XaY9&O#iRVkNZ z4yM1PO)_Z?DO0oEe9Y-md2~0+PrJt#71n+Qt_2WroE#`HNpTw7x47|Gk#}AUy zG2((6YhriD+)lxW2S{F-Xzsgkem9D&;pz`FKqMLr!dw64#muuoZyum4$H&JIE{pBn z2Ezml`aHn9sjn9Sor7b+{Jen!z!K0Z2BH6U|2wp!nA^ueYmBh@jBx?y2<3Jhb|f`t z!wmw43Kg*K1C|;PTkyd!;cEVy)p~&KUGw^3__PS1yI^!ds~d!2yx}-vfmrTZFVYnK zuRoDD`_FX!A0z{Gm$dtVOi|vmw2dbCAUX;_0>X!W4+td?9Rv@Sun=)e|K(K(R>=a? zWq4Bx7co*DWcZZ|2y<^K-A5U?xd)q@Gt$$)s6uyC^IsO}0l&i*I&)Jz$qGain1Wn} z;f81~%CXm7{h2`e8%`#Tx%b%F&Em*|PALfS@C?n&fJp%KGMlqVPxRTJKfHwFS$>^e z&llYc>u{M~dn%JW0o~?fy6&4|gZ0-NSqt1g`kMIAs?N=&3^tCt!<#V1ngN z^9%AFf!|;TVDCWCZ1@0-KrkECe473ceSVDhphSay#mRtqGcfFIquG8_Rea-8QeEA} z^p-x7^y*vjpYa&+8rL0dn?sgoS5#G#$i&io!h`vr2#q+>JE(@^-RyytL`!>3(w=t?k(sP(FWyw%t)JsWzX8;)-@m-99JofvD z$shLM;85j1E?1aJrmWh7f^=-dmL7K`!JXHtw+SZ>(sI#)<+uNi%rJoLz;60rZ}0vp z<`!S$2i*OtkI#aN==g+}^AwE+7+$A7;F6&E*mp)GwQ# zFv*!noOuv%w1O-ljxl?}l&hRQGI#aPk9;vxEJ|91oyCGOY9ac%6emo=*%4!uUi1hu z@3xad-F^Y-dg=PqogV$&+u=po&er?M_@d4JYoR=>HiiMGu9hMu>eLrYhk^bBye(w_ zCjnb_srfz$u!Ppx(0z1x9poM#?lw1rcmc5G2phJ>T$h1B0bTX!#R?XeA!) zH!#)oZ4dbsy4+xKNT1v_XNhr6zV5+o*EcEbMIW|it!S>c^Yn;kF6PQ_vzHYupbq`e z?(pWRPpW6d+g2Y5eyfi+KhW{>@v-o0r2KPP!cbaoc6ISfrmjpxt9g0lO)Zz#*;b)d ze;YE5c5gg-VA#0cxIQ7CmO$GM4o+pBWlk{-<|Ka8w(Cm5hIXOMzvFRZgK4r2<>GJ6 z4&R_vn9Y!wRtSk?azJkFU5gf3eJ8A?tY`O8cXuNt=F#y4lD^&;y8|6|1K~92h&NC< zKy)}eFw29ZGQR_s@ zYPoqqKSqABJN9Mp?liE{B9ERs*EUl)qDcCph(b**l(=Q<*k1Cc2y<62Ha#isx)hNU za57P%IwwIXiH>O0*qPiIDeNj(LrS~qckcVqTJJRJu0Ha-?L6De#`c1L2-&zm@tcq6 zW25B?2|t7Fls?M~PJLw)uKp;YFkb8G_l5W@{Stj_+6&_B3~y%3*S)dV2REtPU&yGR zOvdKk>x~R``+9xz`Zt^{EWqFk9Q}P}Xvm1JiRo{i6@5TKaYIQ}6=;}^F#m!qNFLfb zedWCwd^$S;yhX2^n|%RFb^3Ms;)%V;_hGwb^{Nr)%G>Yd!@!}vz`aYt9EN1eDSJcQ z9Hw9LhBzW$I6y!DyP(iR-RvftyCjc-Xta5k5{BKmZ@YB9e-tvD-q!Z$rTUi9%Hl{? zDQ>A)dqp`ziIUCJG`~;1_8S^Rfx5ek1zFdv*}n-2tJLv7blf2P<){}Adgh&J=TtTNhf8p%2?4!V3mwtt~9Dr_dk_%^SLv@@0M zrQp(>g+|Jf*UYUMg+=or%-3)Irf-$(`)qbgbtCsujAUKrBNwW=oB0pAyNAi+e!nUt zh&zO~H}M@2*e0KwR`vf2vU0BkWq1zcbTR-f)ixip%ousm@)1Fvi`U@;s_6` zZdG-?lAo77I2}$VpFF@XQHtKK#@X$6+iR%{(0X+E9{GFRpU&X+n@6T$in2>KS%RKF z)$aLkR#+w7sr|IanTVzN+l-N;?ESTL;;?k=_#QZ$-9?SmDbFuBdV##kK@}oV(F6X> zF93J_ybmlp;0Xw5%Z1W~LSw|77-uBU^YMGf<3;vv7W+y|0bq9wX2HZ+?B5SyW6<32 z;Ay#ebv@u-%ni5B{;Ib{j{dKj52o1iSe+0PKbsH9n>H)t=Wpjn^JdrH6_fVq)wHk9 zF(1fw%~{S}8Z@<&Y4BKuZc7D7yfF3nkLB)Dg?o6kf+%Fv+I1mq=4&cy9lV);n)x2|rFB0AG$3m7vNx zM}Qfo0`Ld@a_66~{Y(PSKC%`rjIFEqA!>{to~72aYORYL=r$+AR{8)GDFE1_y#LA_ z9=%Jq%3iPJDtWPQKF3X&NYpjf8B?=(3Y9YXsS+{g+Lo%yJcdG-ZL+IONo6X5trK5A z?3AuOwCB#|w~SLvSHf5EeHL_bFh)iXbB%;AWtmcW^M7jr98D&iVWFWA@~8%@;wS$* z!aoWzkg@$7xtjHG1ED5wY(t51qb3WgbKO^RIeU`N)WaGG%`N=EG3Tjk%p8?fLZ`@05F?x2z~k>jBR= z;QHt0=((h$xwST1otpFRS*Zu@)J|gO(YUO3UPEnNf%(Se=SbI^Yf2>Fu1`4rmWv3{ zH=O&ME>BLfcKJm=+%bK2CUG69BOW*zbzR&gH^T`JTTH2i!MA5HkzYjWDeE0TXz$G@ zp0_zjl>xd$H#ax9)kPA~m8B==y9ii(Wap?M(a6n!fOdGKVY$aVrHM zWUCl#t9W-9uHq-j56G^!4&7MJX>{vL5;%ERD$UqKOyLvz!n*g(PhB=Qwq|G6(zqUa z85uji)%%&6(bfwuJZ*MxSG9zH=ofDi7mj403h#7{GcAmRX|a^A-MRBtrWYaACDXfP zZ7oH&V%60j;NHq#r)XULDOxwEG%&kkdwrf5a?++}U{L{!*64pxfd`bH)i>D^0S|>C zhY{y3+o0DHyUOGZcbY+Cpl&$M(#M)jJL|S|a%~qckP!02ZR7gj2De0vrT>0zPvnhtWPA*oBPO8a1%0DA=m0962!Q; zbuUKn0$Quf9e`%VpkD`}2?@8Wok;}S6RE1^XU2H6j}?>)@@^aN2WHMNjo?N%DU}IY zy-aVv$|k$emSxteT3|AAE9J_U;?s8XHJ=@ztnF-USn(zH__@7Zm+THCB$ciaZSJeD zEV!w);;2FWU}-m7!p(90j%SVD%fp7#0_Av$cvP{^ofn+e1+5DFuWLVE$R6CoZs#*| zu-rF`Ote|K;a0F?lG3(qSljYs4y7z2srL2T_Iu~Ia?useVmnqJ%S(UDX_^KdQn0&} zK6DC-S0cqpa_x#*r#ks=N(?KWW6Lp7eH!Ijfy&A4OBG!OLI;zmBwaU+%ArYyk^JSN z$Wh}?ONoNkJW)S{<28VGf|L>>oD)0$2$aFIYJBr7zZ;hP89KIEs>&C!7u%gVy_GWs zl%u&VkG!g%9bZhvM#K?N=c&>z^jy*We(H|!D14E!Oi#_3ROJ|_#@xaFr-li3WEHWo zW6}1bjqNIeG93`Jyx7BN->om(zsE!OF++O5sD(^4sM~UQ)^2#BiIU;b#Ao}K2#zPN z-v>o=-?E%eH5|^gJ~EWy+w5rN`qT}&(7EQUOC5C|m@l24oTSjHTYl!7E$n%Hfj5AN znw^o)wcoI;PTrKK*JiN=!=yomqp>|*q&DERuEIVbpo2->rbCp9&sE5`DrxV~$oq;u z&&)!QtnZ5oClMp~#wZ>uE4BQ$_ZSL`+tV5$&S_mih?M>KbJ&g8hjBMI%ZyTFMMu7U zHJefckJDWok|@>2GI}-Y_xWA4pE*blmunPt5Nh3;WrH;5^|5&7};R z2aOBX`lDn#JuVR&+Lo1-!`Yu`CO||=7bK>2Hpdn_!u}WBiWcxXt66*rJ4*Zk!&9&* zowry*%jVgdkkD@c_rjJ*kY*wH;;cHKNfB%@6u-0Bf#-)P05_mR^I_#ZH-8scTO}vcIb0nZ z8`I7Yc*_A#fQT05_`VXNxT>4=9ereWJ|(~3a;h`k5N=M7sgOQR7JM{vZw zz`+8o^2!S}C_C%V@%g{sgaohpFKrL{*wI|MG7gimJIy$35n*yTJau$%;Dc3}c#=G# z7{b~eAS1w3UcM`sQ30eQpMczle#NbQ73G^*m$0-`Hb?M%S0j(S=T@Mv)9+L4bxV#ibFKn@yK;VIM1@CHRh3qD zHjHv+W-8dZ!skD;B)EC@O^Acb4uVPX@zw3^Z0LO1`Dd_0cVH9?L}POuFUfrR3t>C3 zi1Yt9D#pcI|IuB}P@YH+{y*SIhXb-o1Zzr8jvQj10&3e}IR-SsE!gEF5Y*>t7x<N+)#vgALmqy>?lF#Et#k-QC^UF*7lt?MB$}zylIK z34S}6(QU41gBuB`fC7r4e*n9Vh$BUx9eRhY|6E@S?r6X+qq>bSEb#Fpfci3s(;y z@aSp;)_GMC-iPAh8monU61MwEExG`*p^b?q%?s|X*W&XLKZD60ixjl@P&s+-U#av7 zn9cKaT3W-z9x-j~I=c{_18-~?8X6iKI}IL!@(ffYzd} z*QSF2;HZrRaOf6z@USn|UxU`tT}Vd-3%-{Zj!bg(P$UCQ2O-3zfHwn550`0pzkmID z^WP6ZA4nN+jGb1obSv@V^dZqfIZZKPlNv%TXvE;}IIIC(Tn;x<#?Zf=Pafg|k4#_6 z%cDTuv9zufik47sp$fy1ORG7-ll8{#=PT(Km!O} zfz3N%VIviYhv+C;-@Fbp3l|(6B+|jPKol#ZyRYyAi-G&913}0euNI+bisjA<)v=nI zni&x>vClvdFFmU|%}c{f4Dm29st~^la@CMfw1cY*NWUL|&;r8gFf{Wa(1AvBridj@ zy|%s{AI6^h<+A}2$KxO#kFL#VFt&+(M_?f(9mBg1aO*#W?Z5C4QS1;(i|!h7 zab*K=4h>+R7m86mOU26iH%@O$89iy^T)apG^(Mrxk5HEZF7`dKE(j4pn{YamAR!a7 z7Xf=Xciw2Upgy~3HUZ#uL>mS*Zh+Hk*xHr=VF0yMafliiF*^lo(EvJL-UbiK_ptpb zk&~Cd4M%lw;M;(oNlst)P_*9ti?CudGYt0G)>MA3(#$bV^2tupx!} z42&OF;H%E_5=X{iEW!X-u))E>IXm}3@MbPB)>+gLKJ&Un41*c&HOMi%jEj3eToNFB z{~$^~Xl3635;LX0f%^X#67V3A zL;lV+iO`7hj6KL%P`s>lzNNhE~e}j(;3z6Rr)>Bv#19%ud zp&jbMP+NvPZ#?t_`k!+N@0xev_3-+3M-(5(HlQ1;f=mH8_XmLV!b3!xz=rTQwE@IZ zz>TGG9=E4C2L;y*WhB$Rd*DQaiwgM@CllT`K9sqa1GZ&iY6^P$yMS;AY77nvx&x$T zKN0q~f3Bo-JDepy{OUc!Hgwod6E#D0=RGnX^zXO5AK3~Y6x#3tn@xOKP z;=^_*kpLCPlOWUV78$b1K}Dw`fE2(0yhu-o$m3L!{VNp74z~mfdJ*ln3pX6|Y=IlQ zucM>bjZM!aPhGzJgm5#Z-@e_tf;;mQ+gq~MOSuWh@m%PL=^H;rX>H7sT{+7>f3LJG zFLf97eDL@&68NUSWINS;c3M_2R2T->=Y~=)s;MeOR^@7hf z7higIGP(KB@81|RM2d(7o_%rLaoPy#OGv}8P3R>BDu6d-ru;ZQQX772H685E4#Khn z4SHa!p*7bPEJTL1=C+MaCAsOY%mm#-%5Z_?g!O6+l zcVJ4M7(9p?+q0En2k6E{PaBn+LJ1iY2#i2x7umH5_3_-5H`l_@%SppAT;9zT zNfW%Yii*nQQ2_yELeg7SMkWhNcPy0g^L`X-v-?X|mjpCa%^!t?;NrJZBv&K{@Swq= zn3@Gg3de>$gya-FxFCd0A@+w)`Yk?D{0|F&USCpt{5fLq0#N3V&=t#)Mv+x?tDZ^F zT^z;`9)tUaSf@=_1jB#Ck}*qd9f}s9OC~DC!^c2Qie**wDGV9l+;IrT7eE-pjCN$(=fcRbXjgr(bSSP$Xg#lasP+~7R(`1J+?ibzv zf|}hYrsOeiS~3oUe@K~k7RVTrqB}(GO>)4z<;swE8;1E(Ud40Q@LM(;YBsA;v_M2K zKc&M4!iGY}z|t>Vf&*SHxgOebq&J4T(RqyxMD5C`bPjnpT*Bdvmr{1KC#A)hizHGe&z1( zU|<;QD%2$gH$a6Jr_;frrJ;!BG))8%G?WlwvfA=@61?bwf4? zO&_!1xWc&NHGKU1wpT)4?ccrHZGljC(ohFh0~hHIqA@CR0kYAkiOK5Z^5LwIqKsMQ zSkYk1(99d;$bCYK;eWBDjbQjF@&LsC35_`W8l+$X*y1LcZG{WNT&ucssMfBfWf6S4 zf7rf6dNrTDf?G!p|g2K@(m}*aSmKaH6-oA^`fs7oKj;Qa~sJ`vP?X zo)!1SOtB%Gf*}^c+uPdS9=Az_G{Lx`S0{7Ks@b)C9VjL4~Vq3a|nIl${GBwgD1Ya$kylHy~&xG$IFj1 zTW%n6k^A>c%yADPJ@YBbjI!^$5OCL8|DK;Gl_MZo0A(qOTfd^=6>mE`1OA~pgRnXD z4B}48A2AJi`2c)VbUBlV%^>#h$4A77q?ELqX&Gtrxc&zp@2T2uw1Kk1{QLLs|Fq^E z{0PW?_LvV|bZjG}Bvvlg&-qLQc1@IXsl_|b|91J0>#_wp<;nGt76sOiCTctj8TjL8 zv!<#=u?S{@ZP4lFCbYKy7b{5>owwqk?5!#an`Q-ow- z&&Y_VY13g6T+V?m@v82y4{+~uZ?}`TmT*&Dd6dxp{{HrMt?Y+EL5RU%%g|u`BUT}9 zJ5h_N?P0oqzcX5~adEAv7bodJ%N!Dv7`5-6r0pc#{US6LtW?Y`}iOhBL$}LN}2oetbDCFrVM$((3_Foq{;WFW;YlQg{!h&2}a`5H| zK2avJ(#MPfyfwqhzWIU3fl9{w+$cG197|Sx5=+MW{K7$~)(!RcUQ$-J`PF@LRkuk* zxb1UNP?4~tSSB~q3Ci>Y*wv4~KCo)h+duX0B$3%Gy6LA02`_wwBF@#SRjXj0@EyLSfF_)A&`YRO2C-D# zNwC7YY_xy4ZWU3qV&;;{Fr|TAK1|@&FFnc&0GT8lkp-negkd|7vSM%RjsgJS^a5FBgTJtuDAIN`jBTXwr4e58Z|MU6Lw zcaf(gvIZIw6-59^iTnVaX4Jvsn-@pmp3oZxbc#v^W zF2{K&l#nM4ZisjP<}c`axfF8&%!riNuVo0M_4AtSI5FnGv>|Rtv1XR^F#Bx04fXY8 zP$fc+TOjptl;X+s9m*z%5yQYbZvU2J%as|vX0-3A?N!QQS(zX(3NdP?ojcYxlH;UneV-nJ|B}`=N)hKc;0BzB*pHX4jHh!!F8Q#^O>Aq%oW`ZZ}q^mJITxQ64A z)$iX!LT<2YY~RWAD`8s1PjE% z$|+KVK+tc@ETSk|;riGk_|EMZ1H3>vkIr{d#~?L|ma9}RTqv>#Euu$u1id9S6_v*3 zW*S0Lg7VPa)QDT=M%y@q{!R00=SdX+hvWDbhdwB%=pL~dtcC20iuvCW-Q}SD=ZGwF zh9UBOl&^@vxs}&AX=|I_VZND`mX?MFL^5LV_N1cjq*9_P4Kb*R*&V{QX%U5A`N>PQpGN=RA z#?j$g0XBu(Z3D&%`dnSW&n1~MG3R4h z*-cS)Kv)wjTmGrNiPs)%!x$12Ix7vV4Gj7*9fp_Cl0so^r4Ir!*xyGg9xxEo(tk%c z!{(fupacXDjL;n2v)AjU-G;H8fRgWd9x44i7Y~%Lt9&ZFn zorSK;e?My=BIx4EhrY4O3A|X3s&^3>=y*yVihyi@HMRImd?L1JdgS^rK%m}|E z3kTwBCF|G;=O-G81rYA};q{j*j-!k!TK?IGz+$~EgqsUxzK&$u{&SoM4{o7-Xa5)) zaggX*DJQzQ1fp+kdOcWYIL`2oe<3uq)=lo$?F$!FRBS2o&4aF#BaoJuk^=h+)Zz;( z-$onp$Fh&Z!bGlq*Z|wOkdFk01V9$<;KhC~EQFnZ1+A%7(cH4J^})u`GHZxwMT%Q# zWk9dZa!JI$-rSR!nHlF3*=FA=T1+=;MQ0Kp1OT(&q0$Q+l$tQ}4Gato{WCWn>MBLP z4uw1=*@ko$OuiHR$FTd+6(m3_VDtcg{m-F8#-wRgh-}TVO(1F+@^wHU#nHrUB~X!~ zC{uvk%-LP{zq?v?4}&@peYvIHSx1InuD95q0p&RM^!#c`&28JalTe9dAtm4LzDgpc zgHpgzY=n0qM<(qgUT9i^9G~N{aeLoK#6c7dtF)sAl}r{E7QvrC#H}Z4yH1G zmgEBEigX&$%M&Ej5Lq*1BY5nXZ)ef9qja{H5*oEH3^yREh*!SnQi+Kq%k!>FQLs(w z<{R*@L94CRH5Leijj#nfFI>oC2^?JXZ^Qj&OGE#d0lSOZ*+luU_ibdpd9A{Ee5Pm*i zhj;nj0wh&638P|6{krt>&Ngn2o#|8>IXzx1(~gUP zVep(d>&X;=tsA=D^yZ{(?;z(&MFhjxU@#O;<#AH@%cz)t?tyE)`zd zGsj>bA2G-c^cby_eT6KAi?f=mx({QvTUM4Z#VxKU&z0L!5(3v~)7?PGzzb9WG4I9I3#%hb1Z^KkAVIAcSu;5yKm>`;k&=a9u>h+2cKuc*SWu#xn%0OufLAoqN-k(Z;)5rN;NJh5kL)I;< zq2n1^DNE8^`UX5MC)6cp$&x|^f*JDok;LQLR^ZCB<6W-cnwOMF)o2e99=S55h+O){ zNe@=*JyC%mWAJFup(({=bH1#g)o*^KoeK`qpJ9I7m(o(%PaV8BS<2z!QOTo*`PtW6 z%kL zJ#Rm0%)V?{Xd1EBZ+jj8YpREC{08q*aXj{FC zT*~(A-c>g~X@skow9VB8#c{eUH+}!ieThEkfN6_gg;;)*M_jBtGiyS6NZ+T$gB`#6 z!|5G!-ERFVUXLbq9A@%DldO74Q%Frh2d^vZroNq>otD-v*K@1)08#ABFbn^?v2+Cy z{Rh>v45_CY_AzyHTK{?`${2mmweLq)O@M^x zstLrg^_@2I;8fl|5RdWP;F|AzJm4~flsPDGrNjsK6v&}-1vuIvia+jYarho&o~pa zL9Y6ZilO7;;E3xJTZ2H{Zwb}>tZRueQbhstfHNh-$+rx@s;^%N;YPqByG8BDjB&M_ z=m?vu=?;4OowQyQn>6T0hFFrJ4|W-(sw^e%vr~Fs5FoPAB(mU4%$m>doo0rW)Q)&> z<&QFbXdo;d@8=^{$Wk$OOnpq)oyU2i*RX-saF#`MB%7mNi;6KKmZgdL?Q^;}X&ekS zCMtQwVVRCI;@<7OPo$q|^6A)LS=?YY`M2EAvx2U?V5Y6gQ0|P(t5Fs!t()uQTc4Nh zWc+z?x2WsH^LLy$t!@>hy$B6@_Kb%Y@jC>$IEbvW^74uba z*P{3SoZ4lz_d7ENj(a3Tl)m+EGWPNV8=5{Tz0BuagMRBLaesNmrYwz&`n+MmSGQ{e zYM7iZ>J+Cp)c-tTAMZ(W9$RJgPM;6k(QERAj^0o~<|vC@xYi+Bc`L^eG0qv|p@~JY zx^5d0#Ry?s5Q2w9^t8U9ROt6Ehxd0&DL0U_T#F1 zg>JOCI^OskZu*|q;LicIjBq;oBt17D89$5N$FiPW<2si6e7zJSOR_1O<3DLzKlhC7 zzmwkDE^BxiWQr4W;mP?|!^0-4ibfIVlCZT~fgiES$6xbwM9$PjxnSUCeu_zM{F$?>yIYgriN2cqxL`ldUbzwj?!fWCRtOezLIee!>neu) z+EZnjXdQB`t*mBML_Mx|wMsk=ix|&~NK(sjp;vL*wmGOn#p$H8r!!~Ayi>TPY=+#m zE&IJohtySi6=YELt#BU+?TD9(4QaO7(D<#4t5c@g*G{<0%=@dI^#0sZo`qxtzNr0D zQN=f^Im*KWxsu>0k7CRPe^v=pw3h5)*`R4jGJ6;Df%as-emw|ORZ&i^H+(g*z(jHu z+{~Xhl?T>bPWc_Md3j>_Pd*24H{Z2>^%EzF#tr!cd#&SVzHB67HIi)^^-`WxnI&V-nEa<4m>+vYAz`HQf|6*?*eb6!ra$O zidDc*u!LF+r%>G8>FetYJmh`LjdkuPi7oqtg@yU}YI`i7oM%V4s!0he#$J65h&nFT z-e}cYX()U6we9D3W2aBejJ>_b^>O2)uwRynOEEU~@#!;DL1igzeMz;^fpxi;s1aFH zdyZdULnrHORr!?X2Ctdn^OK2zo9!EoEW_+r4kvD>3p#F*sXpxlTm+F-_qM*MhQ8~7 zvhZ+s|Em3<N=?RQtJBaI4n`gaUjFgt>U;N_p6efeORaW^ zG5D?McB3uB-_&W^`7a8~Ni6h+o6VS?nR7)MrXcoR*uzkB3+X8V3(t z%eu6sk)=~B#5m_Qt2#rS>u&Me-+BgXwwvRa*~OrxHmw08lzDg|RZOyi++1)}+W{Dp zvr}sa$QBv5r0=fJ>G!F+ZNErBr_BgNUP8` z_MkurVJyXzi>2G(p9$b~>Mo>9XeCoc3AVW$@3R7*# z311t9h{blK3R=MA`IyXeM+IvXcOYVPE373W#)~H3d-VmLWi+U2_w=0OKX~@M<8)#s zLqM}mmamB2*vR_v-Q9g3SQKiC>2JFn_I!6nDW9*a5IU*9A! zF%(#^)KrnW4}NsMbONwJhEveRSeMD~_DeOad(H(mAJUlaseDEs_Um24K+U~It*;{z zc{4^~NoV4;O-DXf&zFh@DmH1I@%0gZnUpjJQFcJul-dA!5NjT72fXAqo7HN+Rsd_HHq5MO+w2|h|Jn-Ns z(kaL>Z`%G0391jGPAyCNIy}>}wI_7wj+1Rd{PVPuYZ(>!8v+9gKOAy&axzWVsB6ZI zJ-h|m(W6#PZ`H}0Uf&nBvoC5F_8wzK@Sv zyt|BKkE(Xuq5n!x>IPzvWhJ@1TT<5X9R*ZCBf0kAw&4TaPj|m1Yr3d-#B}}OIQ}0N zaG~9<&CIG7<_(YH$IIIrP>P$~ zueHn|=Ob-aV}k&}p5pYD8>CLlU>AruC(bZaQ&R&u3#DY`3!Wg^2nh)}W&0My#vAg7 zS78AFmdJPDncOiF6v|yt0(*6lmV9CNzknVV~%S9(tf&0#rrp_fAD2xu0 zQXlaF6eGqyv(1qp(d~)FzklDVCJbs&dtR-LK8gLA<|OyE`S6h=H-Qc@_?6ud9GP)k zHLcJ2m{UX~=k42{MR(Rbv&5K=(C6gffACk1CuhQEw4-6Sbirrci(ww}jA$h@UIoh737~ zMRH0ENT@%?J<$z|<)yG0+?MX{gC!)yy0U{bDsRyd%|&2)h>PoxH3dy#k*p_TkSShA zLs^@I$4)Nk<&axJZ!s>;!e?e;5>&poFZ3DtxZQ7PcM1Oj^8xLn6fN!~B@>f3CvIzM zXoz5ZH`sL>qgmx-WWGU;6_CtWR_`WsJm5(v9n)?}|G08;=X^#IhaL@35Sjv<2cwu- zSj2=>w;ACP)yisWYPK6)NN5BV<8fc%36w&$!7&OC7d2~z?yJ6REGen5WQVc)GIDY# zChL%p2HdPYJnng+J4uKQjq5y(*O{Jc3RB@yOZ?3F4hN|=0Vvkz3IHD)K?s|mZAUp{ zndEN_b^(#=8yEm50XYFLB`27^iNW7w^So9x4OGhQ&7x1f-`V7vW>?zU_%v4V9C?~5 zFKx}C9QGvNvoCyod>(jP-~@2FEt#R53iI5HfB^734ksg}+PxsG7=3c-pma11j2GQ5 z;R|~L9?OGD_x9lKi1pGktKy3T!)r%ph-;g+((9%@Y@&-cE;(u+qVm@(F?j7d% zwUI4*_K08m^Z+i6QX5*-R8k5G4OK(?Ja{g6akZEYeAt;NbZ4}*WwLPIUo?u5KIZ1oZTraPj8S~)~{?NAfkQi^ld8*E02tT#?k(~4hOFwY6KcF zwf;V&FJf(LdZSPtGcIU;D1K^V!F;vT6g&nHBF%=0|z8|C7D zJZJJS%4WQ=aaI1pQA72h;LNsv#;f`@BphxpTvuBUDb`I045DB7vj~~@ehkRY&PJ(W z(kO{o;Cd{347}d@`g+6+iWA$g0;rc_FVsQ5hX}w!SKZx5u6hhky**$WvN-X=IYU+P zM$odv)oNX}!#??ZxAcy^Ri8B&p?hEE6&O-}Vlp$c|7_ajF7q!7dM!@s^wm?Fcg|j~ z$W`@!l4dmVV&}WokUzrCdBeFgBle4j1?RqqhsE@ELFTAh%BSuTR2h`E;NP`K%^ z3oK4mViEk{puYhiB9Kn)!T2gS9n!%NgQ;MuO-@by=(_mF1*iQQw*;ng`!k6;+_t)% zrEp-EV0fd}MGX@^nZ%|V^`#Qln!eSPX4XlD|5q<-Mu;X_Hz^Q`{gadQ&jgK?tnySj zt3FqEYMWTA8uGe+x#qZ-$PrXy$jr=AQ(0pogyz~eF(vV}WgompY{$fX`O9{^br{e8 z{rwe!V4|qD0kYpJy&~kpXVqne{!Sd3dSM|Z#P$685S3}?msI0*D!;#-3;tMc za^R8V0d$qACO!LGy?tWLuKU#nK$&XgHU2Jqw^__C zm9PFJqdQg3C~;JHHc{i*&aUSdJj5(!zQ`LoIhzLpVCv8MbCHvLZ1~n)o5-Ohpa*~m zDMAMEEiEnJiv1@*TpJhwz-g+#3I-9c8qK)34|L~+SEw+cO!6tsFOlN*`bUFGQ*lFJ+4ujT_)dBigz2%GN4NZhNul57>y?rz8XY7cYVQv%a5$lOw?t0p1 z9UrS5Z@c74Z7>{qRau`J`|7~d*i6fq@TtFd{bmo&_4<}`7y}?26os+I;2;{JXa-dy zUq8QK7Lw~X*o9;tgBq#~thKd4!Tom=yGXt}SiZ&$ukfgrUk4c~+&4;p zoPYI?xuW0|)wBLNGbbM}ObjKRc29W}5+Sx%b7)pcjrtUQlz73+9^215jnz(oGzrD4 z)pT~b+F*0G!Jz-Re)Hx87_?Nifh7C=_EeP>8(Aq%5^ZJ4LIY+r_uH|1#DmFjzqWt6}$bPOKu6u25A>u@MD|$x8imt5)7+EN@5~l5PCTJMs-&_tc z+gd%`!4fjs`A^{Ak*lbRal{#(zgHNYa`#$l_T5#JQ(HA=9?#v#{HBem69Ed%b~=)Y zJ7cz%uzQpm1Z(j-`&j?j>2bkC;7yY+)n|i@Y6L0b|1}XC}}-JS{H7fgV^VZY%J<0%Erd=e@7osYPe#1 zVBc}LiCFQ)?83#-IcxT^K;d9fOAXn%>xtSKeQJdolRB!O=kL)ecn>hFja7=tnshlZ z7wY7E)7NT~Y>f?LiL)t4SE#F2e{5hBx35z;V3<38_d$eVjE!& zXg7u-I7A-ef%^D7V;btC6kb2YFXpvpT0{*b(GrrIrBT)+YS%}eaF{Gy{!?fx(t-H< zbYaZ>Cwbl3pJ}vI=w)fwozhX59qXVMyuyD&BjaVsLiLC>n-=}f9h_?HMSJ&$m+sv- zN-sso#(uo?vDn!D0CnqS&G%7HcSyx6ad{d_sq!Xs{Zx*zVzl*fm|o61CfAj;>uyQ9 zy;|m|A=@U#xza0fq#Td)dSo1fm(dCV-VOo}TqOiZ^pX;QN`GMh-2~!$_TWQDk1FAW z3Q{mB@z>B^3EH2`<)1L8ZleA3_jI;mUfy@ssC;q$_8Fg40j{waws}~={PFJ#p zzW8b+Fc!4A$s{$g>OJ))zn1=mWT`mTw??yTnQ!Oi^Ax^_W~@JstyAq!EsDWrRo2wx z;zjvqZ~)D#J153|x>Z@R2S$G_`w(a;5fpko%WAFU?ybr8qeaGF1TVMB2wawZPdD~u zL(VUsVrsScvmhUFoj%vcj&v9-D;E4F+rs*K~)d*t&kkls(q4B zAfM~P%T*D1=> zKW@5^8+XGpb6-9GRjr*=ZDWn5$xk>g3k_iTUNcoOs|+|`t@X;aQh)1nc1oJUq8MKwB=ytd@sm!7`yr?+l7@M9Tc^h zBn#dq%fn->gF1%3fZ9=BJ^)ZTgavjG7de%Z`8n6}f-UDXmdycD?=(%2@nVptB9CU2 z{(Bv!8!6%M)sl~8Q8OA{4U&FcbK=ccMUDUKx&4wI{p;?vBi9P%hQmkX>*v!4YN*w+ znq92b4L;BN-06r&|Ep-qrB$`Hf$?SevzEflxyP<0%EJvcqtj(^I@-bZ7bn(#4$8!H zjX8Ih9iGbO_&ypc;4|E=G&UMi+;NPtuA$hyy005H8_Fsf&$RBU>t ztIbl_on6k&%hP3M{U~X%H@WGJ4Vz7pS+&{GtNjsErV083hjsKEC)}L{fiWp|WPZuF z^2_JC+JE9^@mSlb-4e?)p|KJLIWb?2TNdm5awSCC)@N@Q^egLsD!%X45B4uU{^g82 zL$Vyrej9z}ZFUs2yv(V!c%dwgLo@Y#0!liATxgjf)Arq8K#$rPo&y~hbxrc4rmFm2eA-?t)5I_z2xH`1Fw%ex~bWLM4F>A<6%oM##A z<3G^pu=r45ZvUo|n+Dv z%KZ7_qt?DN7S6vHdK=2`Uq~Nx&ofjDI~lWcYt+)Gh-=yvE?hG^{d9FB#GhY}`{81A z_0{rHWLFKX8hIDI$$VO*^#9}8NT>0+0@LS^^A+bMjaeI(51woM4-3FfP4i90{RQ38 z>eTx%dHSSf^&6PRAREAQLQwp;PI5teQ&7j%B)+Fr3=UlXHILecwx50)Rax}H);jB5Wu zH0a3TT?$qW^}MuS*W%c6O=qFxG24E+yk++XYc~p&o(ZBZW|r(z^;e1Ud~?~tv#u_F z_G(=Ez^=CA`hTkPYz??a7m{1kmlBtcHB{&6|4FiMU3M}N)i*UuxFoH~-RKxYe}RoE z=AhmIz3mn9*(R$8Th0+%?^-eb`lFsS%D&%8!eKU*_h*-dM2+hPqNscRSLtSX_6AIq zLyW%G(*6&ojUfaN;nfu<=f$<}MSgz*+&A?8UQ$w8>t@bGLR|jG;QF|@xgnd`)G?d+ zT8npxYBw*g^}LOI>DXU}SaPX@B5=h-iThrW@{O((evcC}dVA8hT}!(I{ON8~}<1Far9 z(!6~ByEf7(eLac)n>GkE&?|*G@WsX3QHcqyK>T(6dMq(!UO(17HfAosj44}Ggx=|W z_%1wrD6cNVD3l6rX%)J3R8>_$bGWSR4CR}UEz36eG4fJlZ5}S!fxZ!BKtdsw_Q7`E z5N}otBWIA|OJw$wrIB}D9$q~IcfP&~VxlIJq9T^heD00KQ=f}o+y2yeN6OF_(f-UZj0yCjkd*<|MnFh4hQFkYS zDiYUdlB|=jK|u!DXg8897N!SHO-%`QI0t$uM~2rDv|As`9-z7zEAfK_8Z4W0QBIB> zLWcpC++#0g$q&KaBM5mXK+nXJn3!1qwXsnJ-$a{%ojXasFP(@nA#(oN3H4YdaQi|Y zJ;LFXk)pigB~S}NRBtrFXDN#wb80gk!DmxaS{iGN*)&j!t9;5II|M8TKe+$jAH>D3 z2M>yxHuPri$Ly>unM>s6e8c9zD}v{O!*Y@yjozc99pAp)>M5aG)josnfc~DIi%q&)C=)$a?=C7Qv_I1wm0g$$_U4*oWC{FBAwnv|%D6 zfHXDb^*@lOm+(KtrIvJiw#Cr>?pQC_| zf-nNgoEU+^tCM$-*l-iQFLIaMi;JdM2)=DAw`5a7qlS*on9qBCCY${ND5AQ>^vM|M zecebfkt`y5n|dORX49rkw6t#%6G36cdx!>*@4;u6mXRTQ32^M4D?I}PBc~jD*E}! z!TdhvQd^6A|M}WcX;*Jjozg8*|En)LF%lY)_}gzFEri}rU>ys;djsvqJGuEjXdoiy z;)>l_Zq$S?JB4FCN>0-@Kvf^lvPznR4okfoIQlCPQtRu3(025{Z#sV{b34;j*wZNt zj74V+*mL9qG5C*fB2LMoBWoM8TR5gM%7u~4=^VO!_~h!9XL#`WT76n<>=`{D0PZ=i8Zl{~{hfz6-6#K(eKP_}vOCcXs!E4=|@R?G!7k|3#lPlZ5_O zaz7v7fE3Hzmo&s)qwhD%W=M%NUj(fv<;pzSb_20G!N+iX8#CrR;(wp|2LW*y8q?9$ zOR?c^-yX&<+L0#{^_Jn05l{!n9t!fo@}y3k0z%EBllKjjM}lyde^Hl}lcQ5`JDcQk z8Zpw-$0~Eia&1}QUFFszO&%?vEi09Ql?ew#sT_Oo@toePx9tfgY5}do2HSN!ttJl zh-^pA9E_^J^HUh9fWPtocZEoQe);mJu8>pr$op!!-gdtLd=sFGb(-k%#AXbBtg1rq ziX1-xv;X_xJ@`w3>nSyRA3u2_Y|)t)9ZiN;d_O@7#!-5!Izsl}k;|~4smLA-7Z1Ne zO+*gPI0^}6@*7sC*b_yTD**wtR6FY*NgVcmq?+(}$qU-Z*tBkiCxT{S0xPzpxc~ci zDGoZfOK!hneg`>*KAqQd>*y9RF~fDXbfggNQtA{Cn3+y8>z56G*Q zZ6%CCnhv^ej74P(hG!VPyG(XCki=>{L&K5b59nqfEd|;mH{g=)L0_Fr18`87V+2=Q z{{4=6x=`%J-{}U$)$Ua{`Q>un{rLVJmzWqkSpIrD?j1dXItI8Md6*C?H*>US!xn~z zg#jdnxA%ESiJ_QL`!&|Mbcv1c8@uAOzkR`q;_zCOP4q)A8+qevqrKO4HyPAChiDQu&;9#%s%G^m$rp9xC;Cmk4(iZ3t>hy#PNy z$Sz*nD2D3oO>67$Bl;i7X)<{6k7&zyb*=%cqLphWk6WOkrx#7$0(H#U$;`}53taIS zM^~(X8h8W{@_z3$XHFjfqcX)Am>+<6104*Ap`xw)2QHJ*jBxMw??;_xZy2+a{b+<2 zC6w#tnKTr=({@wiBJL!E-;afv5xl$J2OnSGhp%u@;YCmW7>v2&v96l|b0g3IrK_rH zw_^90N#3zu-ewDS8}-8R6mrDB$z>G9J6#4z%|*HSk(&9y?@&0p|ot zEWtkY_4Mq;vU=?0aBH)&wM8BgVsj`cGBO`Fk@SY-yScr4VI8+xiCC(Sl`H97r8h8l z@UiTjwrq3=kP{~Bbe!(rSSM{&nrU~xdHI*O1Q`^F%Enw6w8_CNX_R2)wo>7{_T$Rd z>((OXM%X_)Geb{A0Ix+TKmJngAzHgJs)n@UlwZ79Gjwp9_9T-C9QcNI1!<{MhW9x~ zUw=dVvpl5eYGu>;t6XAVTsaS^gryD zaJg+}Ho9_wevy?i#Ea&JkeM|%H%I<}1{&DGZxjRiqLVG~M16Ql&h#ENwPyHrY^S|F zdXRPrjR`j2yKfSB?;bh%!2Z6?Nz~ies!(IZ9Y%7)_vjVMi>oDTUgjPoZJK;gWhGqP ze@np&RyVZbo5mk5y?InJRa{)0NmpNb<;$Z(&759xqx17^!1CAW58NOJ#y%u1Ke)0D*zx@1si-~pR-5c&d+)>LS5;?|D zXggi?Pw;1#`3hZGGE}#e#(0TzjSkG<>aW~M?juMgO8yy1N%w7zbT>VcjeQa9h4x95 zQKMI^I2Fpk_uh78%z0|AAF4y;@qxFuwBi~E;cchMJ)qi5TBZ%%gD=TJR`SyMUjJbM zSWB6g5{tlXVxLk{zsyId#dkcx(191hk$&YKy$QTMQ zhii^I{?ETD&(uH-_WkGN_nDFA*rEKnPOe%;G+H4eiPU{)qO?$f%ehF-UNKw$8^6G0 z0Acnph8;V!@*P!RXWU7Lnn)IK%`sg_6tscZ*aM=Xn)s`YnXvdpPrmo}!_*wj*6RFW zJM{QmF#E=7PYkCv5`bHa)Xj5QT)yExvN85t<(z~JjyPS+xyt(7C~ zU%(@D-6G?2E0$U@oali>jq|iqIqkL~-g9s^2lxUzYfaLdM|Vlc+Cngwaoe_i#>Oh@*Uc=$9&`Uu6wa?^-p~UhAjf2GL23SNEs52thIU6X4Sb zimcc64k<^wA`IMRpX}i*97u|fV^%u8&tLDp0y9QUBQki>t0;Adr;WT2(j;(eUvR?R#qbSRdI3lT; z#WaD1bnP1ikALH3S8^I{O^@OY<5(9D52%M=l;!73zC?_Rg0QnBLNYA7VO3aJ@LqyX z()qm6>-qTfGQ{S@SzqIx;9dNo318MOl+{T04{W~?W%eBGjH1_DcG?9BR8{<0x`K15 zSA?U5a2WCBIU~aHT&p?)b4Nrk@ZJRMgVby?_83eDX8=wO)2}D7D{%LTa4Ug5`!gADQNF{0*~Hb{ zEWfF>)e*BM_V}&E^oGgFpjWto&L9D??%AWGtbC1+iS*;>6FH`^@;O4k4nYkBt+D+C z%;LfTG&CQ)g1;ItSxc-V(>%%_|C3n3hR>wJbx`N^gTj7xS-ScyT}XTJ-q#OX9s1T1{?~EmbZ<|bQB_sNa4tVTKV`!LC&a}&DG1Gi507h|lZ|>*!N5)Rh=_>YqT1zmgVvyL*ojc(P5rx9+0?sKQK`fbRkC%tMuYGhdjC4muS=n{838Bvl z*eMzPE>uw>^9jg$;3O?Rc=F^jI=&u1Mwt^SF$z57`8vy&t-;c&|M9L`R+Lt|*!c=#sz56cdhn7X1E$~3fX zL<&D~9*&(+8l5$bdg$bN_Ov&bhwfmc=)T5KhMo8Bd1#rx*NAYT<1PI4#L_fhEo(s` z_T(X{fwwZZgqD4dFS<_~PAk@;?5TJ8jI^{X&?kh!z=+XxjP)!C{x{7gAks=H;k-na z3JxG>T}Uv=t5@_XR-t2h>8x1CmT6D5N`8xiW&fgYR}YCzp3G6d7dfK#(lb@s|M0z9 z`sBhwDSspCs*-^5!#nIBNd>1|y}$5DC-usJ^t|z|mQ_UQm&Qh0SRxGASU?1atM0aR z(>m328+qwRi`TRM9efdZl%HQ8GqFT>fbI4XP#WQXZRU)V9rIhMTf>*Oe?8H1|Kb-3 zcJ7nhix%g*9Nx;3$uqu89HZM*RAwlciewjJ7(PjubKGep*g4?yO&<7EKHtUD7wmUw#<&`r^0HP?tWNi?;4@ zB3cq0wkM@LT%CNmRSfB*>o>F9dzl_zx_uiK`bA+Qdpxh0-R~!(nmeoLtR&2Cr6q0> z<*qq*Mme(T#MNY9lZ3RT565Zt$PSjxy`w+eZ~-$ba=y0z^Y?2^9Y{7!*WyK-Z&^4R zJdDREAe~$D=f1{lRT_H!;?$y3%g2`tE?F`Qx_gdr^sBP@PRHEP&OFmNRB02RBPW+I zl0R*#>#+CY&0HZHWtC-3<@}-Pchl47hk_d!3aNO_9QXBSY#!bGeQBGySe|uU%)!F1 zCQelm*^At!3jy{vKfh%1rpA6rvH#Okuery~I5e)%NDF?;FhEMvkZ3yoBlo{yXSCdba23mYkwdTehqa#L| zFD5?m=*IYh)HR6ro^XAZY@@w#bRASGa-Srz zPXlKfgLB%q`A{_pzV?H@J)b&9ExaX`K8A#DTjY+_HFFvIkuOvbwby!lw(z%D+Lek| z^a|h0W(?=@=7;<9Z7t;vvmWE$%><(g!vBo1rbHX>v`KD{-5kNCH@E=HxV3_oxe z6t$rLy5On#&ouUjPYdjG%}mX1{fLD&`29Hx1tAgbC%VHSHG&6bU$r^=eSTo9d)CP2 zR^U9X^I^>avnw6*p;eI?o^iVAXF3NA*`~gj=adv%Tb> zao(85``WRiyUz^e-ZRfD&Pc9)H=2-tFg4vw#cCk8aM6#|X}M8Nv^iK=u5B6m^c^e~ zC##-BFMJ7o@Z-OAOq&FbaT zdlzeS`+)_0-Kg|Eg^EYBUTUJ_9nTYoHWYf;CKpcL3cV^LojWJ-%{-(aiG6RH(Lh4D zXrqsr&{BoLfRlTq>ud2=>+y$P!++W4RcHnt)FDpOfg_Y~n7jtm=i6P}LM+ZFv% z&v@Hk4>NnW;H|X%SmZO|=06?mb0t?7h8qWT6f(Wtf0*PaO5DBRIodKjef~vyafO_0UN{NBP=lk;79{Xbi{$8Vg^J7lS0aVt|S{A5xi z?}OFlG0z3~#l$)+uT6b4$!{M{`x@EQ7pExiEW?`gklOHzM4t7)R{gq8wdO}51v5=F z3oNk-BD4JqRrjgYMlLaalZZ}H4zGMw-!T7lUGU%mzZT1^uY$dd%r9n{Bi{Rtwe1sO z-={n}aNtb>J*t7FZW>=TS?K0>bAKtg5b)iu<&=9~;ivrm6N_nT7hM)~G?Tne+&(!R z5GgapxVz(e;IQt~(Wt{6GsBLj9x8iow65=$8BP7v^ldWXOnT2KW0O;U>h&`p2(PUZ z_hg*=jnmB@W>g>i`}`8)2b-~&z261@yo$9MSTGb6-L|i<)q~2qfAAD{?9+_Si`OV8~!2EMD{j?nSDIJT$UqbKGWN1B6)19Gl9-uPZZT{P{ z+~XeG{PTtBp63h4mz+hd;CUsL4)Pp1GJzpQXtCHpLt}mOCfXTv^gnIXyKn+Y7fBaH zEp5gzUq5p-(ar$K%9J<;=<Y%&*$=ITn+oSk7PQY!b=un})fR5yzjh?EfWAn#@H>@rquMJI?%de_uz>0F zno+-wa5WV?`sO_lo^Z5lNlw#JY$n}aFm3(82GzWjv2P>11z8C)0@n&=`gQVCiL@5I zEv;0^dD(jtUqtUNzM;P`TUs;Nuy9n;+vt;!z-z`ExAaQ`mKYGgF17pR`agFZ|HxRM zSZ|^0pnpnZpexDYW@t-T;o_Q8!y}Pg8nZ{E&!=@8w$Ik#>2*U##JnwIZr zAGIwCa`_B;1M~`}pJwiAp&kAz)Y-xp_kz7CX>qubQAa^YnA`esPlx=h|EfZd<=ee8 z3i@n^&&$bmWDk{9NxzwrxV6~vVNTaXGdNaYagCE_TT{Il??)vS6NY?K^%+CD*^68I zViqiuPae|b$S~X6*wK`Rm7@DQ(=Rt6_fe;pZUX2g;ieV6s84`5r5}UVv+rfA|27crqEf9b{LCz=$wa z%m>pBKJlw)4cJCBy+DB29y>H!TN_SatCrL4xb?9^LgW}32#VzoT#k>exQZa^o!Cq6 zzB(EP5Bqyj4rMRJ9`;|$viCHoZe?Wnea`xf_uP&~zrFmosZt{3U6)hx3;+Cz({&x{ zFHc%+y`0edA%fpJ1^MY+`)h;IKp;s?_vD;o_ZrdC#M4NJEqwTO4opTj! zR)gpb*DpW+3Do)XB`19H(2Y@R_xwt+=DOMh-q!htVv#44mTP#Zl#Lk7zd;j%Yizi* zt)WIE;&%I`Wu;8#gSH*~_c!mPJsa00o8F%y~4fpqWKQDhO!#Q)#-fOS;#yc@_@z8WYN*98ZBfKt5{nK>uG@4aJSny=5 zo#76SB1&S%#ha+vXroq5!$F{ee$Ie_OWjrEq?Bj`Qki{`d*pO4GSY3`TeWX;Q7T$q zbGnu?QJQtQFm-U7wo0hxw%bJaj?23P6M4aLU#cLj5aQ)N^S2WVJL7h=DUFQ1>%#{q z)ZJ}aOCE#5=HBC5N{?(-$s4s3*_#s*e!V$Zo8M7!Q#^TfXqXzS@~f7jCN=2k9qA#t z*vf&Biy;T4V`9Yb8^?q`vdX(2sEFVzPxKKW@9T{8t7)2?h6=3As{DBno`<)6&(PMs z8uPe6tzpq_xz1@%A-#2dmfd5fpU{53=X8W8Pimz%a*x~2P~eat?5oadRTB%Dt@Li) zt(sKVM`(6kdah8*6|FqCPeJ^>%cizVKey9qGPK3Xc5O;04;H+=xvkz{Ax5^L=LB3G zS6A0Ffx<0fL1~01&{|}w9)L$$XKO1AF0^!hLId^ZbWEW|&*T%W&XhbW{ygmr(%Qxz z0wa4@9eg3@9fG`%6*t{E?(Ezv5h>ixvhM99dBbrB)dOXM%JXA|Ay3Md%L;ys{=mHE6f~E4`_(0lFXeHw4Kmnc2hS|N;Vr%T9c;S z-fu#7j6QaJo*!tZM0S<29+|I1{Ncw?dN6VE)@#qwyoDA+7$4lk>;vAdU-cjp)ZX3> ziYZO`*dAK1;n@3!7s3@Z$~a+u510^WVgVzr95Z3knc9$Rp2OR&_1|>#Vidi0`g{4| zE1sGRBXh~&=G3`aEzdj|+pZ4^t^=3tj{h$SF5Rh7$eAXnzgIXxz4z(Jg!-zg`7(i*!d^?f4bFqwu^nU`pXLg+^3yA;;*97G^3>DmPJV;aywc)8 z58K+7o1Hs7MoyQf553TW`^k0aosqgI`c{+TD5VwBg;^Umt3weBX`|rm0yN*8;P%+I z-;S?(c7bI9zxy#dDst9xL{YCSmc{;9=zx-*N26e%OU#vrO}1w-+PQhhUdi~FZWu?) z?WBC{>*xAc9`vnU#)f$1wj%a}Zjq{gvXl-vM@Oz^?R*cxekIXfc>V2II9?<`QJl>k zvjlnL^GDhkHOL9&TnC1)4eCocJGR8Tio3-zK3twPzddBEjg8Q80Oe@FI6`Zc*^Ujd zl|X*MA4X8mp8%x^EJz_SHZ(BkjNa6Z?+Qx&V^JTyRjaBP zU5Kyc`bE%}IO4%w`)!9K@v9?+Z-B;F|FM7Cn3IUWA&2+EQqf0q#~1gVe?7=8txGZ| z-ICiX2NW1l1qkrF9IZD)1dir&VgW+{b*tf7Y`zZr zPOW`-AKTP+X6z{YJLgHGoRe9XvHF^1d@!-Ab^U&1$vkN)w|DYhm;IPSvZ?zjMUUN+ zkDo7I@F_vaP5ruwXIU7awrV!wc@X#(r^dT}_RuY~ZJ6s);ST-{ByyYl!Bm2Ih^|*- zTTj2CXP>GimYtlGN_D}{b>Zbn^>Ozs8Nmxw0<%0N#+@vvnS3tOpY82DAg^IiH-vs! zy35!V&tAWIB2m8T@nT}ukg>Q>fqOsp0@d)d>+XKUWB2V|iqW?^>YA=Hz7VzTZTofd zfP&z%SB-hkQ(?BZ{WpEe%q4NXz&>o6MD#Er#U7WxemK1ca#kA{L4yefpxqqkZh=hZ z2W*`;AS&40)P#}k1>DT&d#r2FGBU&@9J*ct@xj?q@9UEf21z_y*k8EJXdMa;qd$~J zGwfgQ+mA^@-dXmwH6H!eu-zw=UU%`v@OxLJQA${I_pk^SQRDfM*eJDSL&n7U+|4xc z9NIF%?zXNj)6$D~(nRa&dl}6wbh9=st}aNW1ztK}TfQLq^H|_`*?Q(-SWgJ^LWfY_ zsM4dpL|vh@wC&@i27w9Ekm#Lu3Cp_r*_^z^t(}KNxT~C+wudsU%cCd5{T5boQ(vjn z`L8?s5%(YK2dhd685nw%^u{gSx4+CEP}LKR8<fZ6d?(%bUFKx*RUMu5H9*wCfH*;u9OCB!peA`Y+ z$slwnxGZ9vXV`<&dA>NkXMSg0_yZSDavtdgFKq{#O!IW~n~I-e?9Ub5s*(S0M~p7AlB-=|12~5?mS`1 z-4=}c=M|4#Zm-y5#k;63biI5gU-IPMBg?kaOm=QD?V@mEW{LQ`=h=jZ?n*gjxaJMC ztQ5UtxBQrK(mI)od~a~?VAp>n+SZxogcfa#`KhzFPY#S?5F@H4WiS}Y0{kL3Mk*XJ z5}4wHK`EWn38so@#SFD!>JAGK!#%=Gu3%SBAW zy%A5Bmx#3{*Zg5Pd7%Hu{&ZtUIYX)zN6uM@zr^kxZx6ZkW>Y!$k6L=g<}vi1Nfo{D zQkM4N*5|Lj>NlUwPXQ_>~4^1u#RO<8nz+s!EU9Q=j&+o-r!|Mt)%W9 zhDXHA<6}oB6D5M-eJ|td_WO2mIOMRo%DZ|qNEd88Gfi0J zSXn0PO%AT}=ABP-T2*)7pBo&?)^fu+<&@z@OK*=!G1@bHakU&YTmAigHL0xY1+J3z zRRVe0y2(Pivia^XZY;pg0$u=LDA)kRpn<;=2`Jz*2rzaR_*?+ag`op8-xzrxhoHLj z99mxgY6l{8QbS++FS0nFlo>XflGc{J?3q?9i1xBBtwxsf7t&*O-ipWS#J_3G$Wy!P zh3lK!7h}o4>sLMvg;hok+NjhgHYF;924#Q8iC?YOy^&>+q}63x6f#nmwT&>+%lvu$ zJ!u{toEvK%lYu$`XQ18SZ4V- z==iJMqPo>eU5^><4vYnl@f+ndP0a2-60H55hLd$Oh5q~8?ag-DMPIEW2?N{hkPg$C zy0m_i8vID7u-ZRfc@GoED%jW4qpK9~7>{S)Cr9;k^dBuV+GSmFbXYhaQdgd-?lMCA z=hMJgaothlus62xwSv>WOIFRZkK$=$hxe_OtYr^|88hnX%M_f8?yb^;A|BCNFIdog ziHaw5_>oHgYvrIp#ay(jAhw5I=nJX--HYEwlD$pbWj`=D7{?G--mvgMb;Y!#tRvj-H+1pmswtay+*rJSEy;?*|gQs zm1Y^4Q)c4m+1+oEx>`B1zq@r+#+l;PY)%Xg%I9R`^?^28Ye#(z?Nh*5BP9Ll;X~VR z65fPbQXCvaW>`ze+huJWX3;E0qevBT7TjUfQyx5FMQtVk9GQa>2$;aRf|vv#ZGi3% zVYK1N;0N-&a}5td^MyeHLJ{0Uf`fn0ir9N2eqY3Zft!d6*08J_huznD8Aa#FXy()0 zU$t@Hm0Ok(l<1x{Dd{G;x9z8O(%R3FP{fZMx#at+R7zb-tgm(0E%ITW+)K@TSEKmS zP%M-EFSi>c>R0mU7v;9mbtmsq))GtBeo9XPbJHCgS&N*uJj`8p4xlwZ3|qt`mc?0w<#MD)NSj+(Te?~^>PFiPuT&MOo5bnfuNVNk$`}+=^OmEb*Ta--g)Qp`q!~8l2MJG$MTpr?ZU4`B=$ecTPoH$ zD6jO|jZW^nk(7gcgO0RLCZtYB5qYa-pC|c^)QjqQti;Qs?{{8_pP-mb zq5-Vst1Oz1E`C>JsaDdBfWU5%>q}o>Nw^w^Ol%o9qqYt{YF)f=sT@%&YsqDPwa$Fjm%Oz#f0X=CzHIrVVnoKt`OoL?bpsVYN)3OT1-`G|k_ciV;4>M|1FaZT1JRJWV;te#iOE40; zU?T)PHwvH<07mC@Kk*biG%y7z6aZ9jb~(Y$00e<6?qNBdp>!I4D!U&X&-T}7J_^z* z-_A39_z_#T^oi>>3XhT0cn$Vj{{SKCIh3h$+tf)(RaFNt^~zI=}c|jdYgkAKz!f>H@i2c23qRKD?zk$IhitRNX?xr%KW06C1#DGVXejtq#DP{1 z2K523D`;X}0U+EgwP;YL2eWEOdCtVe8Ui7=Jo{-7_U+?6TKk4-hp{Hi%CQib|I-<4hax)G z?QprAo}T_Z;*7|E`TbvHbaav#Jz`1KXYZ^K?vjXIts4CN;4T5ci2=l7ex4a#BcKa@ zZwDMc5NtpD157nDo$<71;9ad+8pQAXest`$wr&(cw zS^Jqd_sEa41&w(|pWj$pOGrqB859>6W8n922WbGbH0Ht((R3UDec3iTI%)(uyB0zi zlpW^hH^_{=#H{E|P(7{&as&WVGJr`P3nD|8|Hc333&JA43xACE?^h>*`^LU_QKQZ! z=0@VVk$8x+7%#Pk!Z<$Hs*DUnU{qoB4Lz#Q0L_@|JPQIaIJj|ANMW=^!EP4*UQAmQ zql7#N#+=g`6jOfC($azn&`rSWxQG~W{rj~nFv>eOB?f_k)|aCUeN3sjq;nM|z|F$o z(2Zr40WxF2+8i zLDTzxEBE&fD9u17FD7=yQ{_TM4y>1vw*bopA!g@%t=ic5czVQdhBKNS<}z=bfg}(( zOWxDd1AbL-TUZ!p5GY?EURCsfC^n^35GjAB&}ENmoc}GdX9Zx^B9*{1WKy;=R3ZXb zZ|LTQ#cG3@ajkp?#sVN6(dYwcg zcKdn63i1D<3Bnc3?cnL1V1)C**Z`~%&m)=`U~%7^1~AP%4hszsiZ~-u{9gHQ6LY}3 z1HdzYue9Cv!Oy424;J@}^^J|#C%WFO))>bUg#D$z%ct%Rhs8*-ew!Y%D&UN@v=U%~ zx3DqY2G}5-M<6rt#@_yPe>U-qvhbVXzsnfPIjjOCBGTXnXrIla9-jjq4TMed01XEB zv)BkaA!iu*29>|D;1?+v`vEo#NQM}5vNK4aKRs$a4-CO#0Vzs=5cc2O+XHqbP!R82 z$A-TN0q);t_>7-kqvBU{b4FEFY*Owyf8M}#3NQtOcyQ697ykbqO2e}jz2e)nN}?D>!sn-N1rQB$yBL-mW|2rT zL{U;wO0WPND>;Y?p?mx9ryhW3JUEY`+9$sGxSuJkrKrMp>@%^ow%$B@4cTB$*U-=a z@%N0?qx+Z3+*yzs?pEYR_{#xO9i@OV(jXrOv1&TZF>tmG{W<@BJ ze_&vBq;eAoHs=w-WFoj2T3SluNF`{wXu{&uWsk!`_|yM)JuzRZqXp%9IhqJUMy-EC zkL(=y!G)%SliB%(kF)dhoM(&+US3Tg?wuu3m{-sA6UV^%YU52theZ@?pRfxU-$wh%qfY)(D8Ct*!VVN$(w#y zB`5Xn4stq$&?Ep)l1uZN~z)Qn_@ninz4GSx;s6^n8{1Tf?6JP8TepeU$rT^;l#3F>p z-CSj2Vn*dF*VBIZ5f1eBYScKamvr%|*E*HKrGUo@mTc<~W>8%w2$k6k$*Y_j$%p_SQUVcBs zGXX(CV5_k(GsBe){=2f_*Y{Y!6YTGGGk}f4ht&T2 zkjPbfI_hdz_rRd={P!{2G(*Iq~9M)F-(KI964J71n z@HxXhdf;yZccBXMEPHzhh5VHo37duqP|qMqE@rKJj(ugM8+Y2oR> zYdoXw3g7+jLNj1X>kOjwz*GZc-b7Gnh8$HeU!(HRT+#&87q<~j!5F+P4NW$@E--(( zg@kdEf# zz!8CwD){jl4v8(uyZkxM`8z_!1Jx==j3gAX@WBgzKX~&6LsWrUnF%Z`pqM7QY=~J~ zmjEMBK}*XG*v0nt_CWGR_#312n?PPEH3o`}Fpeua3V)v&k!lkZtb@8ikUMDBIJ||( zeYWbn{ZpL1KYqM??l%JyNKS%c{B&ysWIPad*8ipg!uMokx|&Ix+rNEFsi|=R<;Q^n zZ9rhsaF&{yC#Jx9nzPA&)>E*cOsOb250`G!#tJ}^Hy z+DEm+=KlAPSNOYw$<^MxS%X{>5_`B$4A;iv`9q;6?oj{CuQ6jJOMWw?^vnYb6j<{| zARP&hE9<|<#YNq}2OM2d7P+5f{QTNLOQNEo2~8&>BeSa1KvW#kK?3-%6a$9Zju3wI z_dUx#U}oMp%SNG_5_NVYH(+dsVDsBQ0mAP0?=OrEz|7{Xul#@79+-#he*@OMHpskU zBAFYITFxOZyaCMsp!9Zpz-Y_b=RlN#@(X--0Q`FmLFD+~?FInbiBYxq1!k}5V2EQL zCvf6~rGwZ2WF7FROn{UM4%49qfO&^a1)-4X@1sf)0}o)(Ly76ksL?db=qQsvu{EY!E7H zSQxdiWW+Ch0MZ#S%Rz=g?7!Pi1cOmX8G(I1Gc|=Nfk8_C4;;HF8!*I1y?Oz~Hg!bn zS^EKGY0o+Y!q@+M($uW1Td)-&Q>3M$!qkChCE}I|zG1Q|bs96x$Dj+m=P+#7cO%0|U*k*4+7GM}@8pJTR3zv=Q6 zG3#Sitqc_QoI<}{;*z(2c$Xv^M%*xcpgr?MinQzn>AiGCuf0Qu7Oz6BynOTQ ztUR{J@TD>BH`YYYk|d(*=ulD$RWo{@X>)3i|4j3Jl_2=xbwzf>#%>%crhzX%D`rWr zDScE^R28#RJS2>yZVbKU<)nEWt1%w*4#0eEA8PR)IvfOdu*#fK}Y9tXoA1WYhT_ zN*_#)Xj&hR2}XoQ);xNqaurwqZaYrx&!^{@ZoN$o?+ld}kiTV9Qu@xsJ*`#5n49-g z8x zPLDrLwBL1Cd3sL_CxJ|sk&nau)6Psu#>gn+vxAJQJyHq$mJ8$T%SWYVkND9>ix>Y4 zu~|noWz|?XDQ5MFoxTr~zvYL!;;uhD#Qg9XLO7uM=mquELZBEz%(9THs#aTwIEVvbVFG&GJzUEusdu|nf8=_Qz^$JppO%q96?+==$q zUd2muoarvvV+vL#v=L>-T;S{!=QZ)YO!?P}uXEojE>hU*v54*#w6#YxwKcLlyPa0z zGj)ODbOx99=)`n3vL*?q*Fid_ZhqI=@DS|>yS9^sFxtc<_Jep%APU=C>Wad!#ci4+ zJvrG!cgs<mm1tt}Az>6Q^W#(#U6*nK<)Sd!?n;O{X|a z)NlSYMljKg+ETXOH8X6`31dcq|Bd+Vg0#37#r$8_JDV!T{1VF}#5!WCR%2!e!J)X)|E$Pdyv1S`}%R>pZ zqeM^36*#K)ht3nxzxD?(It}xU8Mg#QHB^@F8=38J!$kE`xmkv(cM?&HxZ!$7Og7Xw z#aG1I@Bo6`M_KPSKU#CQtI}XsCcLwpXLOmv<;Rvi27k07>Vo=wm6`S>XA z$iGp7gNv+udx+0|QeM7KV6Q{eMM?X>N$EZ>{RJY$6%<}tS^rT>f4n;_9pgNgi}!1E z&3*x&DR*IxVV8!>TBFY_-FYIfsdFZI$IQ8OmKSk-&-oKUTTX^=e+pAu$;z%nn1>xU z32xosK@ZN9a1w}$c>ntK@*IL!1^VWvC#$E>e-#!Ig1#K|1LH9|g~EPKu}eJaHp2J0 z*~Z4p7A^J1*2TioK6)`l7{nEzwTbVlno11~^Eqt=Rgdj0Xhc1g^4uR8ALsnsXB4OW zLDEW&(ouZ|dN5U#8>NyN3=n=9h21U}(xa=Zr`R*GMlc zAB*&Eqt{bEUpp~BzYnY6V1FM3h~}Xol8e#R{zdw!Oa(9`SA zeC0%>+=A{|1J!Hl&Z?L9xLMjcSRbRf$tUGz4=b%sfBy&(pXakInvi_^GgX~IC!D+#B_rJBdxb{V`kQ#WaBmN9+m-vWbby*t;jc=WiB1rDqW5 zs)?GiXv;JUc9iT9q3^!rgGIGKOKW|#%g`{McD{%v)%pX2vG5xgQIABq?5bA(>*`&C zt|#xAgH$&SDCQO9o;=c8$TS?dIQCRlDbjtTVVvp3Ob^~Ib*t$dc)Hf!lfz0ZsNQ+g!pVh8wnjcSf3n+2T_Y_#tm z+O~z-$P(jrtH^6OkO^iRT6f{daB+6_Kb~h*8$Mk0-RyKRmhies9xV{XY8@j@wmUE= z;`jENSahY9?pp;v&aSnriO`Devaq}h*)i0_969qxl;fYPe`eBF!~tZ9GQk1k+I0Vz z6%0`^StaxXA%&5Rt%M#4%mq+@fCcO!$SDNfvPqpR*R7^Tbxe;zEO>={l1Jc{VIY^8t*lDr}qs>weAd6Hp_M_-RY(!m7w8s*eko= zc|fujvo(e89sfSMC#sNY(eiNkupED<;vrdFaIsVO0e*mT#=sJ(q#N?rT9fmgJDxpZ z&rJnObUD9;Q{=9#?}Ur3lRbH6l_=JgFQK`7Iq&|+shF7hcYzhW8a-K!=zo4l8SYMx ze!5ATOOQifB(SAO8W#zX)kI(X%Rg-WjlN$LaqFYM$d#T-D`o{tM9|d`p6+~#G8`G< z&R4=?{9asINzfBRzj|+Qxn|XddM{*aW1*t3le7;_u_W_rQC*G6 zGPv!QDje4TfHy&fNgoZyBI5T-WR&4u4sd*BCc*)EnzwHqq2+V3-{qdbWeaW^PEZoT z5al8iEIJCt`7KXPwDPX)`1;4DF;LN(c3oL}*&+T?V#4>J@x7U~gbQwBWR5OJ?a`a! zlS^yYsqMA}_6CI;`f$3u?;OyVsf z`JU40p9q`BZ#IC9UyAAKe9MJhy+wsUFgm=w42G(OB=kiNX-*K6sfx;@M9T2yV; zK1E+0rbw+e_~U>a)z|P8-d-iwOtP&?|X5*`*01rl(G&E>=*s`fcr&2{D1X#9l>+MVGkoJj{!yfC`NQaJk_4tQfB+80-staox zdb9iCcsQSK?21oPTcw}k8^1ZTy7dZ~U7_Z(VP^Nrt4K_(+NpmjtK@>hW};da&fqaB zP|gY_4ThzkVI3O6Z_H3~cW z{e#`aQU|*)wPrK^jbvS6i@RUt^FJMYv-JEHj(S_2Tv43=oT?1%QHZclx~@@dP}g0$ z_$`ikFx9%5JdVKl9_6P$VUJP13m4uSF50e#97vf;66z;25?con^Vnvq9<-QbO+`sv z+Mway{iS0TXO9!hdzxq4ZE}d)*R(n0n1&TqJ83AHa64-wwA{VgJk7N=Blu7x_f&V` zmNT}HZA*+NWm;Ig&`Of8_(!#5dGE)JPP7fWliD&S-X609j z?HMm+a88hv8J+m0`v#1Uj0k?>dhP1W?)gI@A;g|91I0_j&PvVj=A2i*9 zc19;|kw|k%tFz07Bs-bDVlug?F5mvPxBVuXWbbh1`*@o3NPYRJjy^FQ@3hBYoYE*u2*5^DJv2?kPC3NWURJngcz7GtH z-F3BIu>N+`VVY7_7pCn(ll{dWw{^Fj-R!w9FSZWuOPmQRM8DeR*T$x)i}MoHVTs4K zDgrB%m%8+3W7=X5YEaT$BK;%|mb-T^v;`39k-U?TTk2+$4GVd-VlJA_FO}oTEJ=<`Ufg~tv0nb>4byf6P81+OOYS;6!bnkqIcS?%rem)T@d;@q;7fjzN2vG(Ki}r zcEObSF#Gzw>LhX}`_>aI%I}LK7KgP{C-Mo}y0|~tA}tRKYFBUDZSBSH&Y#p!4ltZb zeRdv9-tQuG+p2n%Da?89mv+jZ7t{(NSKL)BxX0A|E8|X_s8V!G-}}o~<2G3w)!M|a zsTOl_HV^We=O|B+;3dDN%xDi3kG*Nc{YZ_L&waOnZ}ZoUD3)N)3<11*6!*(dt4>$F zcbID0b^Eh#7cgE`_{r8LN3jh0H}m$A5Z|8a_3+_L8EiC4c(# z0+hCdtmSu#x7q|r;;p>4)`}2E^38;ZNY-)83fSA17)Eo^1Ya%i ztvT%*f84yA*XkXX_K18k;Cd2@(LYE&OGrX>W}Tz1YTgQF%fFF&%tI+bd{x%7(ok_j zN9lMY3M2wVRxEIJM2|u01P0hvl{{dt159TO*A&KGdD+>3(h--Ggwe;+&$RO%vQ27= zb&LhFm&-E;KQiGqoB9@BjHh>dQ@N$ng{$qfkgVKWB$s(Ik;&}IKvFPy1zKr zpZ*Rq_`^eHi%-&I*bM|~Dtvu6Glpsh#0FY)=6RfULuRV{HSdVf z4!g;TG#xirDU#M6R$Goco_rfns;gr+ew%8~xROzwof_Rtx;%8f-&f}3pk^ z`s2s+eM6!5b7V+PQ&Y>64o+C}IvJwO99MgBM=z~e#Clj~I%CT#17$4y)v!b>{aU%?Dn0 zZrc@$<>1V;#65R`=mz=e3+ns1a*-mO=TF_;DJO;5Jv9YYzee6S%ueGR|E2GwnE1YI zBlvLKaW&1dymqUjx~s0tVue|qUC{5zz*=j%%ox*FnlpAxbFSBB8s2mhn;5mpz?iFn z$EB_uRc9f-}gG9Moq4Uh3AxDF(Jqny=I z*?kV2>X-5k3thwF*;pP=I-#$uTz>%gQ56`0!bA>(*#YE}XwTlBLqZYuFK2+A zN<`9yU@$!(p~}@ninxBBK?9*vYMkHJ!etq0P#PwGuiJst)^}_Nk;%C#DXl1jt6haH z6@&je$QS?a)vHxlL3hPFT^u_}5SM?oqe`dmIAl|DCk*D;^`AZ$RBd?|nj@{0>i4mn zcg%^+(B0x)ShR!ub0x7N^CoVZ@71TceU6!=t`;H3NllUO3vQUvE`4*rvkLOq7weV? zO^JFuUrJKz-y_U2^!nVnD6>*I{? zem&mUuRnI2CHNL>e>_T345!O|WUnaLo>n~5?79I!|jqmI~ujclXr-xpZA zWEW!W$u#It%Gv0=qWL0pIpk~XQtmb0jGr{U$X5}M%douIq@A#g@@c8EZidKFHIlzr zCSxpB@yf{%zO8qtDWUo*Anb4nC6fBZN+U{r$-*SQmCxY0=BKO{;}=6tmFZQoAE-wa z)Y3E*b|@E?lH9h+CEW3hgRu2U9L%E}9S%4P`(yc!3-S(+E+pHS#7o~NrT$SAeqURg zmwZ3(XUplPb~V3EsOAJo@mvUD4@Hjo1?AcL^#QWg-Rd9>&n;_?g0kJzQ6u!uN|L85 zl@@=jk|DD9gSt6i44595y(qY`A$vf5JAuSJiopuka#(!&s?2o2qgxRaPk)jW)6&+w z>>n?BFvXY6*UgdEemY-&?hlyR1*?r(1p_G&G*sGnbs;?e5fWKllB2VlI zI6eH^3oxH`s?V$={{6K_PM@8d)1sVR16ePX9&E1`(x1@v(!55;W+%jNww=(+Aox%& z%25PsFrVwYzI9-WSlx#*pReL3QZ;jAlX$%D?$>IboN zIeb-N}*Ef!&$s8H3e}AwruUGX^S3<|y@YB4l zbEX}$t8VLBue7bs9rTrQ74>f2+m;m?t9LVhZPh@}pgh83$y2?wFuyus99L<=&c4Il zEVPniwY9dwLB3s*msRdkK>qcQl6nuzBfI00gRk>HlPaG$51iBoTDyN6NkBQ4O!+o^rvWCLnW|cSp(_Gny z9>(iO;3o7pDN$lAy2M%(T&hG3_tYhOXRBpI{nOvZ#-@~1J5M)9%*<%Jxgn-mRbAMx z<*-`dY`;t4(eo?VuOmkIs&(IWuIVJ3Dk z(gg?$AU31g4#(e;cytS($J0PAGKSX~38G)}RQNFX8wbZtm@5K|9OyL&j|mLI$zM3~ zPnh9V0&L-ZZwBdn`4{3*LG+Jf9zQO>A;@hZw@q|tS%y4rmCDE8iJL+B#~q*8O9a`+ z#6P-N;neEvYBevyfhIW@l&?AQ(|{@Q$mv_P<6~F%f5uH1R8BSC5nb<3m*7;Tw7QOU zNWkE9{A8GE;p%9wp6wX!Y}NO6HM&Nnn5*VS;;A&;DOOaLBeXoOE=N@-QWGRoE^lxK z9=M`r#+!z`t;%Eh_6tVcyG$3XhYm(cEl@oRQwxIu^->Auck5oL&`5DUj?ix+)-~4^ z4C{R`3i*Jxy*NLl$(a~y3p8zMVqt(m)G%fve?8h{QY@c4@$d@BxPhyry1R;5jqu^S!#-$pz8cP>rb*FDNS`D8gigzp>Q-L-(PY9roB(>{ zreE1n7vt&(Z^Z@Lr^R;+u2aRm0^{;&m#|9iR~sh7qpqv}1PIuO8D;WpHZimn_Z?8B zbsfz{w(3`FpQc-7dbb#5^HeMb7pXgZNy$CHMPt)q9>F2+`D8bMKf-y3+`jo! zzX^F);rUv&_cG!46RVuImfbDdv+)FTzpALc;Z$3u#XmWxhAXQ?>)rU9gvYWn8JXml zDpZ;Z@GE5Y06SBh4tp2x=X+P!BL>rX&dl%a+;dQ zUG66syax=$0P~_>rUtWbtWg#OQQ>2W$fMQM6H9Qn1dS&0hY!Jj74W*;am}L3Ig>kp^+{;)oqMr&10s~Y`ujCYG%D;?eZG(} z>V9)Ct2?NUSr<>7GE}}igemZF#cDDt8Mr4Ec9zEt?;8K-9IClf%!CmEBL=u-nOcZcy7T@%^8ZZH}Lpy2f%vC-bb> z-$WQbd-k%NE;QEKke`?GpdhPTW%%Lwkq<#M-<#Po4m0tX^`p~WReaj$Yt@XcHJh2s zNb>nvafaR~Ad#bX<_4z2qXiv$t^Y`AJNex=&S1~MPc71PYtipSF>`tEAbte^5kg*u zw+vDkz%9W6i*y#mq%fQ~7oV(~FJ104j&vywW^B>{%w9*+YPE2~1% zocXCIsD}bKTt~sAHd;B@dg#e`beOE$I(vvGfwh=at-0`>H07MC(=o~i)Tpt=FJkwEUlp{>5^+9##z!4%75Z$2tVk3SZA$j-JZ_SpCH|bTn;O@R7n$$+J*49$keR=^ z2NjxOzq*EPbOBA;VOIQhkkNYXUWdW?ONMF~?wAt_8 z(AisnYH&D=`siLz6}xJ7wIyW2Nc}NMlOn90c!B$OZcmSQG34ec&fF`ZCrFFF^+cjW zTfbqB?X^sJa+Au!o^ zeXRDJ{3=tEkI{66hh;t*_j$PW@c9>X%(HFvupAE#ZB{u1(&F!tK0s`NSn^_4M#9w_ zH=v$60de^=Syq2e0OxRny9X7 zc6S3*%L9bV`@^RU70@%iPOjpjQDATT(apfOiCtOSVe*cWsp9_E&>WP%0AV>Ds=7_& z{p(4&&%6!t@>cBM^t8K|zW0UjVw+5n)ZVX9`$<8kRs9E-#q8B1zwW~MWqF|4vwOmf zqD8T&q?fnpr{_am`O+PoXSzaxEvTMh;?8*VRz^V(+M3vazs$vY;Za1z1%;4ha;c9^ zDXYt_t8=zNbS=QnCtxSK}u#o12QeN!mb zlsPbR_Tn zgy~uitGQ~(4fx#&+6%APOKQ&`GU-PX>qeM2$qdjp3lRi|5+o5IEQFF?=4wd%|b6$lRDyy;PCIWeaN2 zu)U4gtj4YJt~({H739aV8Xc=_1!~-{J}GBOS#p(7o+Z%hb?UN` zOOy|Nc)>}v5VLztpt>u?y_59$XsP6r*fQ>6>oFz0W{Hh_Y`Tw+dz{sdA$n72q>+Gr zw8o@W3RlwkVrs6_RNlDM!03!~S>&C-Rip^ju>sFff7>RZ%z6d&FY>tH@PnWI<%#a5 zj%)XqbT5EoR2`nawdPL3Z{$;Km|-)iD@qN!nKsFBp-_K%*dih`JT`T+U)ffCn-gE_xh-Ip%E8% zP?K_ep&|{OrH=m))<ou5;`RPOfSEmP~`O*y0?NfzZDm)C z;A}jPeXA^RDc=34R+sOX*R3?)Lvv#U6B@R-vUE9lb1^SqOxcz26o*XM@%H*b;|9)c zx9dyx&6;*@ozsezYR%&=)Bd@E{lrOoog(*q)N%_;OAN_pNAB}fWGQ*cTbjYHX!=J@ z(fe|%>GIK)51v_LmuO7uXLGPskn!^GUO>e=8h??2RQ4YgS*AjVtMBd<%T%1V-QL>$ z<#5MZI-+V@(>cE*j)n!*GPtw1bj!(5_Tw#sjYX$0jeHch#?a}!(DH0g&#LQ4-hdOu zhwJT=yFVjw3_YXB44=)~pV=%+RH?5QQZsgO zi#c3KA;Gn^7uneP<%d+Cz;I*~mh$ttXsb6|Wb6X*QoTzu8IK<$ul4Y+xtH_}JKn}s z!wTWSeP>%|0&-GfDv|CVi2Gr`$@NQjx$l+!;eS&6_1xX&{Qc_h>HM&z5z^)`p&48+ z{Z#iWH$nu3F6cK|**i?5N!v~qr6LZg-WZo6CF@t+eIj!3b|GwFCkZo9S~+a@%#f}& z>-Y384L$5=lB{)&qwNx7XSLT>>sKvoV;H7+UT6Z48`9~=#1g({M=iek^{;g*g$f>tQZWz&Wu+q*PvZE@$;rpfs*XAsp6 z0QKZ*7JlvPE2*kFOu`+SdU@e4WFjQrL-{-M2b3xYQ%l`|y6Mnh<`AKcH{1?z8lIia z429#(czsN@&#s>Xqn38g?ucN$yks!1@3T<1#pPlSlwDZm!Krn#z8m&xey2JE&Ufus zXvE18`lX4EtY*oVe?O(0${?Gu5te*uuBG@K$2PABo64(s@u`ec77-dj>9IwZ9bS~1 zJNmy6u5ubizd);jX2Zx+nXlff!R8w(9+JAf1*gs*C8>%`})`811S9NQB`40aBg=ZsqZ`agreSVJ>2cpe6P|?s@#0C-+ZPNdS(3A z&8hkHvFCrf3p0F^B3epixU=|Fz1wuk*G1)6R_T;De#ya&zt#^d{3+i2tG|C%M#ZG8!tNZ} z2Ap8{f>Vf(Rm)O624syGf8Z<0$n>quX-VBr{Fe`I9EXXc*5x@#=?M#!rsn1tIQxY~ z&dkrRxv!4}|4jk_Vmc9Df#V2JtnUXr0xqT~Z8m=eU%9RDq05(@UBBiKK1WuCi2%yd zdH{mgKv4QY=sh@1Vqmq(NKX%i;sn4gf^Qm!?dsx$5^bP&^*`1^eXDg)VmiM19A2M) zZEt%mW@P{(IoQ)v1V`0S340Ea5=5pi1n*U5H?`$n4Uomnux7m4rQH)ibHFx$_JGok zf8PRO4OpoU6x;#|z}mhxUW>!CuTW=q_}k8$>dKWVyn)D|>DBAk5b9BqmS+5i2>HK% z54MvXwEoc|fH+~Zz#Hx^4*|$=8Ly9k8hpr!i-*mUH|@aBumhbF1@sRT@Bc9u{rLvA zKcN=U8!A;(;i2ikvtwp>ph~g|&yn!)DJBT?4Gd_235#0P`w$?Pv=|!zuVD@g90m2? zUl9*VLKDyeaBl*0X){{UXT$|?Xf&BTH|6;rIc4Qg@Q7U^C+9Kmxf9r6KT&rAW*mPR zQ-p9VkQYA`MOPLUm4d>6z^#P<>j(@yhu@blf$#x5(cK2dNDx3V zBMjF6mgGApx;JHX^0qt}lAyL9%(EaIfqe{)YCWG$;azSX zo_p)3e82pq)_t*{b{K(7H9Ug<^QS#ncz|Q5O->evAP|spfS@()i2L3>C8+0o4~*;V ztU4Iw>aHM&{z9b)cNTD`y=6?K6!oqm6CN!^Y2$MeI0+y>WEvl$2vvSIDV{zF&8e6mU~K0jHNw3MYr|+l$bA z@B{w!yy9=;eGYzpe3XFlNODODkFc=U+E{hZI|L^y@wszwMFEcnuNVLnl)*0;0f+g+ zaQ{)T{QDGg-bhSLgz{V?nAa5eYv6u`_y9w47V&5qt~pS{p(eEM^EV;J1HMEf7*wFb zFEu8=_Bi$hBjWsDwhKXo=UiaJ8P~w!JqfL=rd9$S7v73@J`DhQUS3`zA|iY^4?vao z;fRbV5y5Hu_c~Hu2a67VGt|q|6TE=1!?v-48P~%*1}O=Y-*JMBE%s@F!3!XivJ5BQ z{-OGg{?fX>4 zz5%1<&h|VD@|3`~+al&B0sZa&Ey8rES`}ttx+Nl}O3ks{%EkBQ6O<@^=CoG#)qeS_ zjqw_VMQnrF&p(BvEswav@r9(4lV1r*Dat3~-NQFQeAKYJR`%wGymZ0uNIhfcWHx`Xv0&EQ$*xrh?x?OC;XgvBA{Ry;r!{r(Yp7$7i~O1C+1zA&3&?0&c%!Lb-~DH1GAuk9lb<$= zWKv*|yizV@&{I`rl$N=B&nHoBNbF9HkE8?qIO6~QxU+gicRQRTZ3HSqU{tCzvMuN2aysz7ru~hqqRAiL^~eM{g54Q=~Zj4wbVWq%&~cUzv!KYgENTRUxar@;6aXs zSW#H~(FnQPLRYEI^3<(PEFvl@Nzla_t}oD-K$BI$Ljnn8*IWt2^zj^B-MU37%w8l_ z8c5R{AC?6%?)ERLYRMZ!H|iQdqA@SXn1gx~djB8?_hRK5TdoGvijnH5E&6f>G0QFG zhL#f}ZbP1gCy+dl;^1wrNBjkUlJ`eH-K7sGAPzdDv=|_*V&s4U=Z*#|9eC0DEJptX zWz^u)e-W=W!uINM`*Z#TF8783_{D#umU@A?$_qe($*7>`jep=EE-nt^xQe#+0rYrq zE&1>~(R?L*YE4Cjz>OO>@Zdq+d^bD;q@ss-VDSW^8t-ZT_w{-NvVfK-y`dA?2ILFQ zjZA)@{g@9pYbZqPw8MF&G}OTNz+b}xr5$>1_YhhrucT}iN^e*UHGJtrWy8^Jga>-I ztQ2A-GXF29!ThJJ9#dbstW?GXh!ZoC?C(wOC$ z>_+cXt?$p}AGmb{F#dgr3MayB+IU8@l z(@$WrR6ixL7)ig7%uSnIm|nIn{`%9M?;}$h>#B;fJxf2GUTEnn=|2rKyvuLKsZVx9 zzZpF6pE_}GSu6okwxu|YK(%DY_$JM^?QGb)b3E<9j-6sv2n^h~NJ3wZZwp#^j9vKekez zr`f)paKlmN!(ze32(1MzpYrsaZ>+kRMo~OY=RW>FIKeAZX$S1Ypna+uH~=Y z%{52gb81P>d}?7Ms-KpTlQucTHW~Wgy#RX^CpA=I;i3Q6bA%>Ho1gYa7RYKgy(D&; zgYoP%bW2ln@vT0s63np1B$E~zMcvDqa9|(byFwzY6DOnXiSPf1V)AyQF6Q0rX-E}$ zf-VWEpF!(YS$KP@cr00zFJ8LCPjb^?eC^hGJ+0M*JVXHw0mlr8 ztRE{Wb%Pxi*7rSt@Ds!O4w$LH1}@a_XGos`{<`qRpMN<)X7dj4-Q-7xTW2WV`{zw0fdY;n^_?Dx$_1ymBKu{F-FFSlCw+02( zm@CKa`jMq46`iS!1L?;MdD;M@X82?ukl7H+lytWWO%(m zYxG?oEc9vRn;jcF*jZmtC8ZF!EI3X3LT2-4Xc{6UWX}vWUu+p_dfG8Z+K^>0JAebT zGJcD2dU%YE)90Cv=K*H3XED;uCMR`l{Ln43{EN~ZrWO&K)mC)+SSB&n zUho`dV($0IZaNqk&wGzD8VmF}_||$qoX<-S{g&eS(_x*!g|bAJil2d9mXQBs-F2wzG3+uzOi zwRYlG^7TF4^R$`e=$5kcwR3}tc_*^E*|f=WuVPf0itcwrIE@L;w);43%cqf7z%+e- zoSy}Cz^rA@jAU1Eb3L+8sF-?5%goWJ+o0KN&Y0U|9qjHMB$7S##AsYbgM-g&W4scV zI50E4bSO4z&d6NNZ@l05o#DXA^DzbMihCz}i_g)@$Ts!Y*J&!ym!!_FUq!@_^At1-z2tGNVM;G)pnnVnn??HSe%f5 z{Z8!A{nWtI*zHbBL1OEU2M#I{lDTqc$&wqjG%tN#oTL!c9m`}Uw_o5NTFz-)o*Um2 z=b&oERhiGN>gmL7quI98UA<}?mrc<-F-5;-jFH@Y)mLW+KZ-{?gw!0bX2q{&iT9`@ zV{waArUNF};K1oJV*yN_X+@2$_#it&6#ao?xV`f~iWwn&nk0k}^<{KLS)D^GDS1Iv z2}M!ZxwB10EjE7d@Kl$L@)=H1(Q#nqhu^{+NrC31A{_fYmXLrrCmQ@yr0IH!Rkza7 zDRo7?a>ialLC@S-CPj;aijMkvq>xWay29$A$|9pr?IGWe;#o^n@TWWIf(|70^?cHx z4W~_fRsYajr_XR@We&xzHi{66bI8fE3uJ#AU#zp2S*@K7DqCn>E$=B!JXihI&P={Yqj1p9d`f4Fd+6GxRq}IN;>Aw% zcQwOLs^l{!`rA2!&BbXX6#pspKM3P1+P_+~XnYe}?#{fjWRt^m_Z~S}Y@lV@u!_V2 z1EWMXS!|Zi@9MbilInwa$qT6~HM|0{#4N+59zo@$60Ub8io^JNHDZb11&3zVr~1a| zlvr(#{4*T-)Dksxew#t%Mb-}+FGk$=;gu&xbrzP0jL zrZk~ROZ&ZsJ9oag86)HO*OG)~OrN`{ea45^hd(XHJu0tvx@<=#H^oTvMYgR`HgJ8t zf04TfqiW`eD@^Zq?n18|9(&VC?&simx|WK?cKMq63gh&=`G*F0ZOW}$cy*fYIqPJW z`KPhEBd31#=qvIs#YZpq|}2=*n`*_QBQ?o@VJ z)L! zUNF+zOLys>nt` z<6_0VleR3ulZ?)5V{C+Gu#53TUA+{dM9Qmbec}jQ{ErTHkv~q;x{Cc-6eL_fsR-)? z=?X-&kXwjBuMWJ9e`!+*(J1x^ImT}j{UA>jn|1pjOl$%+2DNKQemm-Y|R7u zT~e>(U8KgMYI0KeGc6^3*DXn*iMT618o-QrEnW_K!gxnG&e*%2Q<~I;`Q%hY&>v<8 zeS2b(ireEjZYjCstFQ>YR#b%DZHCGF=G)o)2Q-_*C)bueg;UrlHy6gvk5)<8h@b-( zTj`GNaZH{0h5pZ|S}{ihW4pwHwA1Yejy_{ymQgr{4q>bA!?|;!?xUP%MjcqVzdQ>{ zS3Leh5m1Ka2zRj$up(nVuZ=vwPJJzz9TjXleS7Y4`xA4C)`qGnn`KUye2dPen9j~( zsrg*;q{D%Zos9RJBBP()DC+K4um#T)8>xv)HotgRb3VQsD_&~OCg`?kbI`fminf~X zRTm|{c2g#B%+or1SOep(WV_Sf6T51uOeD#qO}w)<)Cxb5r6{-TC<_V}8Y3772~dTYOel>JJ2VR)f!AOcqlL zGl;pY+)xcmoU9FTPD#{bB8)mE|FC zx6PfmbQU&Rvpt{BK*7A>-uUt7N^L&TyzKX4ox?SKYDvRAr~pP?fsQ3ynqiXcbCh*k z2fzEe(L>3(yaujs{zYx6eMr8vQ|=NSIyBlF?sH^fblbAuONQD@Fdq4i$JMy@ae662 z{@hO@y5dmxtWfoV@7cnd_>|XZgYjk!&FKpQ(!$~ExN&4<~6Gq_i%xAm%g zsqnF8Os(-id7L{doV^Wxu>06El;Qc020L--cXfH}*aJ@$#aj7@gVI&ph*8{jU_#B~ zjEfjHv4Cns&>W`9CC7oZu0OkHH&ug50;kJ9v@w>|p~$DzZi=hvdtz1jVkN9dnu5#q zXfQ5>)*YoWXZ!yzk6`hbB8V4AoVT@OuDmHo~S#e*vX|`fHpG$;ige;)IlC8DHbuBgb{P5tLyL5Aw z*&J=ZH}kZG%dAB9w}Y&RzAs*wy*h|Ye^McXngE_+3UfeH>lD&+I$MBuNNmYsy;RGT zc2O!RZf*XDHgQE&5n4F<&N=C)v@>rm93LgTV?TS9#F$XqIX)|L*7H%+0}j;69JYob z?u1KBp(>j(x zeh3S@jc_z9)C{H-W^Ts^*$&&mI;3U^DBB3 z-5tvIJFQVP zwy8N*kbr6|OxjCt?91xF`3RFKW9&Oq-=5+lU(ULEX}a^DiL5>FYl?x7Okv{7QBn;= zj2ikR^rkgJrgUC6#O@=frk4!?IdKF%{Lu3fI{#{Ji`>o|x?@{4Y@>OuC#xNoHDuz= zvfJ(rZ=A3Cyhz>b$z`mXb2OY&cqq0TmgA_`B6#pf-(#!di2jvI-ouJB2fkz)4m)#K z`OUTD37^u(WVRXpcQ2r!^x9hHMH_*d)31XJf4_x|!v6k#Cayyb#uY zCO|SJ2A@;r+ueC@{S`h(^$W<2Ah-goDWF%+V=ieCNv>COc>CiE1*=JlBVWJN^1 z1(KOVz>T^4IV*SE4}!k0Y|`kBaJ8Mo>Eu69hrTMF>A>&|kV|7;PH`W(6Rl(y7W6c4 zo2ZPE(mRP~T$WwSL^N1$8oW~A7$DaWxE5RSsmCsU)S5h?xXM0XIo+M#Uhto|&QBx(g2$>i>1qsv^? zf86%Gde5`4_#YlhWFUtW+RIPFcUaLGTZkl5m;Qh04vL#VcW7h7ScD=v{&^I{6I>=8 zlKNT+oBNDgo9cRlWj52&7*Q-6&(1irkt_$3>>Qc&O4W&f0DgKau&+#Qyy~m7&{^!E$(=nX)mv9dCic z%dF33cG*eF!?~#mX`es2_&;=zDerDpxRVgel1-6y_(~bA=5Er*bN)bPI=MXkK%+qR zwD9*Ac|QuxeEmO+QJayB?66K8k`R~l+Zn5|=uY#OKl#A!Eo<@E=;KA&n--*QpJrxR zxV2M9hGyhZ&d=IqgJe%Vwi!zmc&}8zQG6#O$Na`4%k96np`wfsLt89O!sRUGJ?1JQ zfgWQW;iHwPXIu;93vH|Xr0!DY%ZAT-`@05^ig|>+(gXfS87JobI_ed}!^5rMObO~&>FX5{7a}4e@LS<LB%83=3#F^aLeM9_H5PfP26n$DMeZAF)YQu6(OS{sxNDGw|iA*x_ zPaAe&{UT>}|B1`jic5IyW>vyWek@t1Is2(d^paUtZk97u2`aBrljSkT6)I1);F{7S zlT`0gp6$prr;%rLX5ueP3j1W7cw7R8cK2nTmMbrqb9lbY>M(B!EXi7Z9#+jPL{`eF ztYUL+d_X+=)*1vR+MX$JZ0h$+utmuwdi+*Nnn3P zq^a?;tZ!N8QRu9VTye%VwaBVp3X(hWlXM#98Kt7KtN%hDi;>7sCNf=(QDv1OX;&hb zxc!z`i@8eY6x4t^v0n%#Np6chAD!0NJH0h|X=x^aYDrbp?R$(%wBju`k|^zXy1)wW z{1dd2`vxC*Y!8Zki}{+!P)9#C6HFfs^F7zCKC^Y%T90;h)Y~#E%V}TJR;42`CCN-h zV{znFzKM72*4ndfP3>v3QyCJq- zLa+7y{d)*E?Hw4vZ6@I$Z9Nf+-Vm$q_*B51rlgh+U(O&5CA74L@rVBE>dyOH3(?V( zClDR@o94D}*g>kb_<~6-mRUkZ+x@qyc9%GN?e>=u`LWW>NZLoy?QF?OTDk4Q>^~As zC(_@O^uNDSO&?WWGHPpM?Q|RwC!zhhQW05sr0o|)`BZ?sW=_4gl+b@YR4gxe^fz{T zXy^Q8WExXOYEhe}VGhcwAnkWq)^pRE=*xqKBJwzDQ@V2xxE5O;lE>gnI4#Wcx<9Xs z=I=&0z3nS+_8N7MdvT{B42o#!d0$p4DVn}gsC1_Ly3;{Q){_kL_MV#!V<+0Ah2Ngv z;_UG_nfn5pNa*hV(lA$;c2$7gAhq>QxA10B29m+DubjxgHcTkVIzid(_ z|APKjN^9rZ*oR+w>c(^)vMZ=RGa2uN!Gq8;qV~ki#p0dKaI-Al$hq)dot&Ke>O8!I zy#5}C$q!g=k4hvxDCec_qJQO(m$;#q^!+r~AhGWWe32_Y*b|!M5Xzh3G*Y*97>PC&MUR^TAzW+qLfGNF@oUrv~$+bdZkd_29vhi#wSb|B&Sb_8)&z@Am z$B-N#kd96A;iw_y@!sK9(=Vqj!=GC^1Wr8_@Jrd9TK#f9=4)pP##GGmVAN-`JHc{{ zjYfB>{AnCpFu`uwTVXO3DbLQ}iN-pnWaX`$Cy$X|5)-JMb<6fnsB0wn#8aM-l++zM z_c2%thmBDx74P*fcfxu2jaFV;7|hD2^*TtRE?zw4i~Yb-nPjnIL{nMX(1X>X<5Sf4 z`_bVquHgtf~qHT`z60ipUx8&s9<&jadYLS2`7RwUPRdezv%_&LGF8s8jBAYR(*6AxIMc2Vo zQi7@aF>0=)l~!yU9H+JPV?0AV!otUxRKMkTFxjlaHAQ23r;RA1(|v6Iglk>wqE)>& z(o@yKbnmJ?UD8FmZY02HCI^EffMax-c|7U7cjj|J$yMx``O#)!RV@!S%~*@JM_au$ zWNY0+4rrEhV_2V0+Abr+qPe>_P`^)HyX0hS8ULv>KQ77Y+?Y=L{v3sN5&HwWjPz2& zjGY}agtg-2m5I5@X{UvE6N4p790Ob4Rfve%kYdhqbVhAR^|^^iDi!S{-0kgu+fn5t ze#WuWN?NX{>H8z>o%TUrRD+J?!rWbr;u-B=m{D8c=nxh@;(m?*X3HE8Abct%&ieImbC3rfQa*^P(D|S; zl-Si}WnM5Yrv5v#LMmcfWbqD4)Sb~ey*%%mna^OaMndY6NNF)zHX~3uK(9zW#-XiL zdp&^ZQgQF?okmRD$U+lA2@kzm={1Qr?0TcZ-V;=0V0wKLRG6MFbN+4cF=s4m->61+ zx9aNWOG+BFCjtjpwX!0z(P_UYjt6S*?QY$52rBF?RPkR(prJ!JQe2FO-BwK5Yw60; zmH8oMb z?E|2R6W;D zi2BSQjS5VMw49!~sKe^+N};=RT{}jd{70xs`YHZ{@V&z*KaG{7*!@gz@hIw|#G98` zws$5hjGJq|LLsJUF0-iSbM0_EHd55eZL z)JlkM3lB?_)T=GWDed$pdu0L7Orb`wz{XqACoSwWOC0kBqgbn7C$rWA#VaKpJ^zVf zi?zxRh<*3&+@o!bcyYsskpH2Dd<@*T)h1iPT?SaK`l@({uwqO!Xw;` zvE|y|)YB5&^%K2pdL9c4Dy|G|a;c>XB2b>nR~IX2>(g`v=?AYGV$`pCqnrx3dHQ#=D3Zo7Y1nJ?X@+aa9wR$X zsR;E^9+m#vkXbRv$>??}m;SYoM~DBxvc&#VK7G=rBPX;<{B~qlE%ME52!Rp#nZ)vE$Z&9^#;T?{^B_hpK?U zh9am{0#ViF#ky!02vdg&Zpg6y1FX2bz}qlBJprMJiLvon#C)paB^rd7=u>nD%5Cnw z#puY;36bUyHl+7*V?Jx+2-|7$PB;ou{H$qRb8^g}(?YYYgFm&)0cl#+31SG$UK^dBDoj$~&| z54E}!o3pSjQ?T>Z!^Y>P{%(_OLS+*B>Ws!zzhJ3ffpE?>vX711 zb#K3iF*5Y!caZ^(u;ka?AMW$v+|ttuOVibf-wvHK$mM_E7aA_a)kC#vq|4MR5MfBC zI&V_5J$^r-7R|MrCV#rf5>-rMoS}!a8NhSuD@(MILc6)ssrxMgQW+N#TppdM%d44| zd(qiD6d38no4=IDCb#D`QlcH87};XzgzOiND;l=5sIgOB-~dl!n-cp5mS_*B)Wyfl4Cp1Q05 z*=PByLQ~$8S}k9r*%C?7B%x1JY?V%6;xx>pcgH@wO5FqZDBZY*A)id z|GF`zABv>k<>#+~sRuk2fV@NUdKS#Ijawss4GJ`7gK7&8x5Pdzv;7oDy}N#5gdB39 zpuBh^-{)v)u+R{a=ue!uovN*YaG}MgrgEC|yI362;cy&Pru-H;CBob$)fVhbd(+i4 ziq-<*qsu0>BxA!}M&)}~i_@m#syPm~=cDLLdfEqYQPrX+EH~OHa5!r5%Oq!!h?|90 zAb3_k#k?Ku1W(e!== z?c4*&i0qGd$t50tw4xG|*1X(xiU7xW_Ac_t6V=GwrJrBprk}2w$|Od2&h3rJijLb> z_ismh?t8M=aDevNu|D>-VXR|SUzPsCIqrGX`b3cs2uv5>#ctGh!{HS7Wh0DcV!HP(Ajh+2gx}%>*^-qaRCr=L) z+Frjf=W^5@B;il$^WMq*iIV@#TYjheMR$+8b2es=pL3=EbF$Xe?&$P`3>qoiNZ$M5 z_Vk?gINZ|AyS9plqftU;!LJgdG;p~~Wh~WRP8}DbTSM}+N_Jj8Un&{2S=HV2;<>AA zljQMiJDa9KBG@CRS||UD1IHuF`h}d)gMPj8Vd}c3o+mrAXnjxHvhLxdv=wPb{@;Nb z7}A|I6w@Fsuy%P2WwunoUHP*HB~i(TTVAtfZ?E6E#q}F=Nvdo>P3lv!!Ns&QdF`}j zx#L@_wYa6D`tsMM4%%}gpG&iUrucX`ZJ!ggJRRnF!-_4aoRFb-DWu~{vmN_`;%gmO zSFlGT6C77w3P-We_J|5x$YzZEry71RwsE7wlOl*a5YGI5F)f*LtT|s`n^dq=nXO#o z@K_D3UCCp$vGo)$=8d*@SvzibSQEkgxTD%poLBNankH)J(${+f=$^ijCw;8@<4SQp zzvJ5JmhY=RxOz{4qSuKVA{0=Pk7jl{FU;~@NC^}{P=G-~+q!(__+TxHfEFkU2wq*q zyXH8lP-9i)xw&j0ZNZnt)c%ngZY+bNk5Ok;UC|LTkO$!I&ErKFQ*3v_DGx#x=(d}kI?u#yWou#fveBl z>EvH#Wo9$i4mwT-(MYO_ye+(>lKeT=By4J)Vn933;29BLs%{h|8rkbXZK+5>5k8|q zlA`*0EN{Dp>P`$VD>ps!T^-R5D}G%aZ-Lt9(Jrn{+}A9%I(ch1&{B~jk*v1&ROXJN zM#on31Gl!eE*ieP#kuN7L6Pou&Y@#g>Ucm2BYDYnH2!x-ptzl#ns?ugyo}+6bH{60 zSpwaW2gg#PSmmX=o19Fy=w<0vT}rd7m}@^~yCNn#C*3by8^pdDF&G{L~%%)^hV4Geu3aVnz;7;I zwmW41GVA$!n{z3RL$m3#uPcl(8Hc+Zv{zFantvWSdn`GHg&OO5`tcA0J7O}`7z9qC z++MsG2YgsU_>HJ{SHXq`OR4gxr~y*%^^#RTl#2-{ya^hy z+qZ8sF`2K9loN>2g`)|4E!+9btg9R@E;jzyb2@6=Mmed)#Hhvmv>(!145114{&vw4 z#njQt(Lp^wU0N}0wJ>M#?zKsN9zWC3^2JAuPORUgdwaq?I5}LJLXLf9D%k~#nYG`Y z8gSPq2_02cD$PokLzGc!k2cc?`Y$FM@I{1MF*9B31yTGht^&*@wI{M=Q<8^=?fyj^ zJd6`s%P($vg;}Ap?%w_2&HMJIBhy@upYEd6_lMix3JqPPMSpTNi$6E~8J~73J5Zj% z<*XZA%g5-FF_9536deOhMayq{<3E|Rbt?e}jaJgek!__8hgQ`Y*c!pD3vQ0XOE9ibBJXyg%>bGd){I?fQ796v=(s;-9NZ?som5})e`LXz0mkRuCU zw0{F#Xx+N;{ps3ci?nuIt|T(K^NF`O7k`;r6I1rU7|36eij)Cf?RBs-Gd&F`42N!oD)?y(jIy(`K6Dsm6+$f!X-FZL z!ga)!6BdPuh>VPFp9P{7GOz&iz~V^-bhDs<+kTi=qZ@`Qxc1lw>(cv+5UTnqYpjAw zhE7`&+BTDkUn8sd-O<%hon+QATD8)^X0%ICMyA~50@d%J6Z%{8>o zd-4)b1y!A{vOFm)YmmIe11Zs=r^*i(SfLz0Hn*g_;yuQIn&TV@f`D9Dm`m(U-pb{NNHj&m+H!3m5-FnxtH>Yvg;s!0#DUY zkumbwv&sHk?ZAKlnBIU?5}wNIkek3q6$HuoppgTUVHMN{GcYpF{VQt;5#EUw|8x=2 z$PC-V6U>Dry>zVtH;%y0YU{)uSzVJYcm5Kx`Ndq~oIvnRPS1`sH1Y9;94A(e zVc{pcZa-u>w5a@t8mMdHGj0jTFGn(dlzje-#~iR28ym}%&zB{AgQ)pRF&L@Gk4J3l zNeB=Vcfy(@*>#zDz$^qH61)xGc<}El<<-A;5Fl1^SqX7*2O&O(DPTQC)DMESRv}Sn zVP>Ypb`rdY^9*%Ye4kYymu@AIHrz)X~5R{_()0l zBVN-ni9`X)z>^xLE?@|P-wdGshzJRv=Ju0N^TOH4Uo|R1GHxlaaupiqZE&!)jmIOanKRLCjTWwBP}$_zpr=VaHaP) zX7K)>))_{{L_ClLS?Q-Zt)#y^7J;Q_VS^vn!B0zq+G z_X}uP{qEn)0beG9^8Wwch~d?%R}c?k53_w`53n)e4(CMiUL;$m5)8%$|CoLQMegIr zXApIDckbMQ9h%{l9jB#!Cr3wUqmzHJ%SPgIof@<5WH2{tXQ}r0^&J7f+0@=nUZ)nG zkbvT*A}4MDSdtK7zzK5^g@vQy^mGt1BN3F;|NAnfa+$-s}P9uuQ8ZicqZ6+{$*}h5bsX>?>=-$ebeMp zdoT(H+|e40-{(xL_CW6jrqxSzPb{siJDQr_Q}K;^7U$JIc<3A4=7S7cv^gW-I zXc5|GadC0Ma0{n$0R5W4S&qP`PoHFDWR#~Ma63G2n4r$=@qeAoh^>jKsp97i0YFfD zxVn;&lUKum!^r+yq#-Be1v(K{{F4E_ALL0%kJ`7gfie;ZYv?G!WqAAVMgGVK0!mP^ z3vYEj0(rA?^fX=*wTR9i#$pLM4V){zU;?LX2!by7CipTuDA4=UpKfe!f@PY}R|)Lw z|LHkw0Sf(%@4!MrqK1JN{|)1zIX4v{C#T2mYHof$9q8uV-7DZ4U+ZZQ;NelTfDv*M z6rgaC|8kXxEhKvshi1V9BveLz?QwE)!auB}v_f}%13z>`E-zbUf!574O_>RS&}f4{ ze)-Qo9q|u}&3`bQFmU`UD@3qA?@5(`7;`hQ6OLMsgOv;x{oto0?5c$!!r9UB@#DvG zc-IRe?+z?=Kw%EhEExp_@Yw((hO1!ilhjXNvY0|uI!jTV-#F%}Gnciejaaeg8=IiX1>Cnxc< z1~eL+n4NfwwacmwlQo+!J1C#P*QY!cscqeD#!Y~*RU2{qArHZeJUl$$Rh~IXDm39< z;PLGB{=U0iZZ*uhXlZG^=tN!DCt!IdARy2oyksG|*><0R5}OI9?Ek)AUc7j5y?V{d z1_lu2gQgF+VL4~*?pvsRiYC_x-p}DJk_v zTyHJgQM~o#-z!B~N+8?u6zA+;=^x_d)WU)rOxZ=&>rPsfRDggDW@dwxF8O3kmnb)3 z>}_djNxj<}N{>8+m{`Q$JBo@?RPe}XmsqInB8#^=JfROME#jB<{i|u#+)Xdr%2iz69K#u!|*+vP& zU);#c0-^8=IfiYW{pg2&L}GvOb3_z~qd+q1348a-DHRp_FgXM0`%RrDKEe(9zWsv( z2Sql7&J}dLQ2dC(6McP-ysQ`cH}K|g#Q%94^OkTXeCCB-z6}aB3KJO9Brg+qnj53x zM;{(~{+Wa-D%NEmXy{$ieaW*m<9-;cVb*0)()0awpaQRXPbwa~^8NcPm3hL5 zhx}(mYnJi zOy0O+a%#VwqhLNl%>D0+J`xM~n3|prezC9M$SW`33sV#sDtE{jwWgpviC><<5C~RB zXhpaU)IMAEWxtS{m34eShHfZH?=}Ng-TlS4K!9Nr6c`cw|7-h44Di_F=L;Z^?-haD z5Hxe70AX`x_x1KpPftq+5uUr+gjaTz1thx7UK^M{4|gmDyT=V-!RyT2(m>jdGH{pe zi{jJQKA!tK*m8zLUt?W+1a)TRrf}j0=cN!lG6yX6DM==#-=Ck=6D_#N$;qjzhC`R) z1m`Kc_#T&zj!q*|%*ULhrx?aLm@E9$3ag-^pkQ=&RCSMwpiE z&BhV>CPIH=V6;AQju#FDD%`&e1WbXAIEu&ALxP2vbUqN(`PQh1;jjrl4YPJ zl|FGlMh7Nn6_~Tt05{_KQM%m>ZaMKW*hk zl4)K>#2ay0yPtsde)z2;4Qm($pq~Yu$)y^98iJ)yEw@%U~lYT$Zfsjq{P_JU_^ym8#;mG+p5IqHjgFvE>3a7gAsYS z?176_gTkn}gJ8pEwyyH$&yRpi1c>eS#IhV69TAe412NoB*VC>u$z_sQ-j-j;Q62)0&Sr*-z;6qW_b}1YT8< z{;@0Ez7vlRUSd9i(FnW(!RcFhdHJ_*ACd?FUjzW1AX&kXWwa{fwp5Nk6WlH|#j)c6 zs)Q4Q{*9G#_vyk6J_3+NCg|Uu|tTfz`V8E}G0w^!4Wf#l;=0Ht8Thl7rcrmWQ4dvCmE2__w7k zOz-dSzkT=49f)|M(P(WgEui-nwKG#)U5QsLIB7TnBy8cLHtZxA9*~Z7$&iMI8pDTJ zSPR|%C&G+X1PB6u?)hAu11dHpZQNMQEz2n^B&r$L16!@0IUevIfMg_^OgZcQU5s`& zu=!vA@1|Jx_VifJHMoNe3n~N4D=zlgYeZrs2#0D>`OkfOFqV^S*$npRwP`@EGCV{I zVDr}1v$Wtmb}RqSg|>;ixt#)z7f|b(nVE6jnLY;nh8}}b9k*CO-l6(2C!x<|k>cU+ z<2A-?(4%OCAZ2_2+TInURhXbc9YbRH?_k$?8*zx?IKb={|I$A|f+L^Izvv>P?jBfG zRTUa~|MRmI8MF2E^@qCVcuMD2=;kXrblBIg@j#&$oTWwe_5a0~_u0Ty3&yhoh18Gb zQ{zRt^}y}s4tagq^^;p(UDd1jfqe^>oBPfwe{jK~{*sxAi4dHkMhkO~KEBwvG{y!0 zN1>w%*yxi%Mks+V5}laHs9O%^V@|*&LIDvCyu9e|4OS3oVyJc`vN0;8RJ$+;FW?}+ zDx$}@X_9xsB9s$08})!s3X3&nO9n(IfQg|k3E|TMOvDpeKYx61P^!U9?S&ZHovjD= z5P~xb8-(cTrCLVcynol+Rv$>pO)usrgO6bqvHv?M*j}!=M0Bo`XPO=`jRRPGNWoT7 z33z=lVW;ekW4*-_Z~&{icOOvQj_YB2BK}UF4J*0LG<|VWIAbaAbk0s>yrUD z@Z-&ho)sXntXE}~VAnAOG3KhdWM*aE6{W>tO=veDU5Fk|&HjK5?l;kOhC%a<2_X*r z9u{lRmds`z|L*oLK;#Z6cnC0e+u#&+ z3+RCQW1@^tL2q!w9E>Kp$G{H z|NRC=ZmfR_>sT!-j=JyOfu|SBOx$nsQ1`|u#$b9zKpg^=JGij1aVq5ne5%-KD?T88 z+C^a=xB3Ijcapr*Nh6-{f`Jt_TvL>JUjVp#d)6O)cKyIB;8Tkx|M>CLE>}n5L(S|A zNJgsR4zf;872xA7s;v`z;0s>>+z-kcL{8u*sT48gZsmbR|J82j@ILD>Hz14wSp*R5 zW~}_mn!uE{NHLKL_Eni{|3>mw3u7&TxR`Yi`hqb=1_e~gpMz7w9CY95s;h$6nt^cW zHya#m9f@Z(c=@vUG6OOPr(lPIyDfeHFD?@z?q?|mAblY9Xt~}GfJepn3Ezm{Gc($# zSvVXN;lSx$a@Q2)fn3~YK=>#WaSpPjSZ|Okf;(FsRr`q2pX*jL(2HSE&Z2*e`?(eD zJA!dMj~+g>17VevloVJESUUGGnIL;42#LT3cyi*BKPHoA1)kh@CY~R-cv#SV!{h(0 z1w0iPfonKe6}P-)caF>#_Fbx;wXwCdbVR%G^c!;}JO15&@L85q0Iq$<7VM{>j=q5{ z;W&Gc;oOO1M{4kBcqJHMr?Bo{`-f0=5SatC3Jex-Z76u)p)DkzoK(fG-ug_?BTI{R z!(E(0GK6ivRED^!R;`y*a)Z+1l%U`==83$xzUp5&DtTF1S*xoiK0p%<3+q-mxZ=SO zaRG0G)5qX}TM*5!PR;y#g#SKJV?*aWs|9{I>*p=bydg7S*IgsZyn5~R&#~@{fFJ;~ zX0A&tq3j^$9p3BckiwjJz@W6Ex;kHl5rob<|58kn@oTr@4>WV5y~)4-dz&*J-km$Z ztz!<7xqol4$N7l)3&sx9vR70i#eqJle!NI zH~xU?DYj0BthN$wE>0<1suK+Ks*S%E%h!T?H#EUaP~+Aa6ZiOF%u)(?igEubc4JzpA&4DUkTo1++;)jQoHe5ksef?#j3T0Swy?!@7 zC%PQXjkuGV?)(=g{$!EZrx+?Vq*e{fV$XJt-Q1IQ>PJY+C>d3@;ntz)@&~54j?NY)S*Dqmj~OFctqA_~TtVb4 z^=|8ME7A9f!v(!zfvm5Cw4xhuSU#C-_+*zEiQu)hDgmhjAbt+mzx=P)NFHF=s;mTz zYvC*QU@G zJ*A~TF*o42!vbqF*71Gz^T&C++qtEGnZhpYp@E8#c__2Xk*47yX+0pNKt_|3HXxS1 z&~)CjtcOZ{jQ;1hWO7w76|bC}^(5P-+Lc&sU`y-LmX?t{Xs#C! zNo!H_B?>gNpkO}+c}@zqIp93O+NBWn)KEyp{H7MOXZJLIhRDW#@!UBk9qT-q_Q01+ zQ}CZ7e?+>-U^O1hGiGc|+3(KHfq(0!#dIG;e$6O#Dk1?@kS^%FnYu-7Lk zrNy()xEhqX`s$_O{$L82)p4n6D#>+_&q(Xtwef>&bJUw8i;A#Y#CEeO6>;-XpE|p1 zrZfPJ0&EjYI<|ehL5O=op|iko=foIPv|1LBcp=NkweCY;-n?~t+4?5Eu;h3<<;G|dQ4E60***hZi6RwzvRmU@cOMa9ok zn8lS5YzZI6X=UtRwhO-tS1T{d+kdOhciVj(f^;5ai}ilH*1|kLXcW`z-5@hEXTYxw z9Z^Xf+IjIAa=kddfO7nMP=821z7QGFYp(TN3#rV8NTMOzr9VbZX>VHA zrIt;r22*KkDdxQO7WF(lV^O_5@Zc*1?y#Plp4@01y!>P4QiyxoQ*3Pbfjke1tH`nS zfZZ3zh z=)-0e2G*Q++*AYnkkt!p2eco-hWzO{eLZrV4W;LmH!~+VF2cGQH?$$!2|11ztuU8g z*>!FAb*DG{@9*C6hSpd0j~)uuZ8KQ{4t1wrXT8mXE`{ZTbxT5C*IQ%dFIRY(aVn^bmarZAnX^}D1;E9mD07gH$(%3P+BWZngvB_|JPdfh_+I*l zpt9p{gFlw$1f0!po%flqzD1;7-lHLA5E>WNp7`k4{Cthdu_$MhQ8G2=M!KoYXTh;r z7)w8kGr}lQkbO{@e(YUyL;vYkY`nTe8@E^L*zI2LCfIqY`#-=<~Ot8I8J|KO|Bazw?L@#ut8@t}gmp3cTY3e@K&4b)R~k15tg&RnB-Zb&B0^F)md;$KHFQx#f;dc}Ms_kkBk5%sDA zhA_*`%bUt?FK+PyG<-0YLBb*!y2W_%A{^SRkQ)kh>sg9D&hIs)XeYfe5iZ7e@%JBC zF3zKOCT>yd$ca&tol}<7{j@qwW1EWW#W_t7JAT^d_Pk|ZrFx;LL+CO=!IZZxVX2vQ zx_o!L)-=q^t*+M{+Gyc&x((|i7q!~(s`vQm`qND4SZlG$HE96h_DhM1pu(;yoCXkJ z#i4JHz;GO5I7GYavLxhVmXkx`gvxFhtg=zEaf{l^s^tTuvH+_g6u@9oT@ zG-)0$8F&*r&WOVpb%~+yni;S3SH&XZD>|)#ubh9DJVB?KA+vRnQ%ow#dMTbgR%~hb zGFSQxR7^XKkw>nJt|D7RV$WD>cbz(FV?4}EdFw>a2$+fQ5Co(jkuWr7 z3+FP4D1Bl_?yP-u`ASZSN2&_pM2YT(cFu40nv`e-S8MbqLxju^xD1Yt2#T{1cbdbe`Jc{H%-JK<(i% z>}KP=-iYaK8d@~h=AS;INfka#BGA4C4J+A$dgtT*thlnq)izs5 zD8olf2Cs=_PqkbqOt4w0{q*Z?>|`1_jgZTI>$wOUrs%LCOPD5Pn&MvQP)9bCQCzje ztq&Y5FVJyCoBx@HUDffg=?v;-q#nBc$f$NchFlJ^B(-kzMr;;LBvP#KA};OA>pk*G zh)+KLbQ}%fI`he)-%hnNv|-Rub%8$(gg>N4dFwNyBZ7*O*WceQqw*=;o+N-2H&JS+ z>Ml*n%_++AEaAYfBR>K*R@M)NYTC@P08M@e3u_PtgRSf@wka8c5r*rqO^UR>bg?C5 z%sI>~$KmjZAXi?hS8u(hdA2UwVFIUnBB?I`(vR>Shbz_flX^c{YBTx1b>EXIy7eA+ z*ym@xa_*MQHc!j|XCJo^nTNCQT*knNsm#PEA5h>x1$6vstkP4ofXpTc`q4~o6Y=Zu zlq8{doLREFlMgmXDm!$w0&^^dV}~5lc1pCuiUB$J_AOb7l&z_(r7e#Eo2XRhSPca! z#%NvlTV=0QHA1;uEX|mEOONliKpC`6ZR{%^-zZnHi6jg(3v!|xu{6D6a7s6d};VoNPul1A?53-eP{G#C^QV7Ci{c)LlZU08Vac4N9=<7$m_HN&$hqWF1C_+OfZX!y1M5}3Hc(YUsWneN0Zhjwul7p-) z=-C>osi}#J|A+054H7|6Fs`1FkdVOE<@3dOzZSdkNX1&y*#uss{B)nYO?sNC!-pLY zWv|XC3*(>q7k>Rgo1qZ36y8Ljt*C9ToYE?HI;CS~Q-Z=_5>btifu&E;O4`SCFMMk8 zk@(+Qz*|%Lp+O~EC-q+ait6Y}ElR>}Z0yXpo&yiYD@nSGKHfz?0Z{8{L-CG@Cy zO)kI$M2Qs>YZVp3{CY0~=U|km*9{t|w84CAN+E@c=*k-T;L^&(txV!Z@tcg4LKRo) zyChPWd^O`WXlDRA;;qfi%s+jguK@xRLp?7qFHq|nt@)SbV9GOUez50l!heEqos3Y& z8o!ipGOl7;bzMCTiD@^cB#v)2F>`LlEwAii{t_cj@N};MszVpkUhKFh05k;}UqMcs>sd$M3NJJ}2 zoFI|JAg#03_d1S2fZBpq_Q-XjPtso3X|O|l@2lY7EgQd;u#)PiVdd5PVnXQ+*iG5< zx_?S_?-NsbBj5D<3=fbp*xGgjq6J|~W}y)!)9UHz0CB1+PGQ!pc{iV#cE|UDMzSLQ zzseoWS05l@ELOUb?9s!&|H?#eQ8$bB@qBrg`fLf`xnoYKRXK~qEYM$ljdXKyj3>nD zQC-R~>vkCPk5>HeL(M(H2yE|CM7axFz(h;NL0KAnf9N_Tm-<#xQt~c}HYPe6B!T;Z0RI|JPR zqp?hJfBj|O4)4St9Nv1zg}Xzs%5|%qm3V#ylpLXEqxSfM64PouY0$?j%Ls}FiL_xj zQ{d+?UdN>JSifxMK}GhTwjcbq;Tm_<>e0AV5Ad5|1V`$5l%_FhF34he(onB#^J-Vx zmm1+$r)C{%OhfKEMuZ&;$kkhN*w^q11e1Lwc$Gn>|W2A`ua5+7ZB)x07o=9e7 z4@$c1p)k-IbaMkl3@R!rz=e0@g@uLnCGp9A&&+%QI-LKa@954`%{23(7RTI!!8#M^ zbrtF#sP`XVT~+TfVRxFt+e3{#Vk~{XIWAF7_Kn91O1H@tpNxEskj+xREhOD*rsc|5 zto~LzJ>vFekmT=C6JbaPDW$q(sB*2)mofmfJ$j|_-BhdKZ?|2t$`KUl`r<}Bd6Rcl zI+Wo3w6+WohyKG8Av4wT#G1>=`eUZjh6%Qa3d&wWXzsT$73y=d2Fg&Zxoas#y(*`J z{@>aS0vc)G`R3?(KqFOno@!x*iSIGB8lJvO8B(&B+TSkZ}tEgn?#% zUReQ!w2YsH%lh{1#&f(>snb9y1m3Hj3zJ&$3Tr%fQz40p)p8=~WVuFpqXtjk*IPGA zogd^r9ntx&uu^+|X}vn$G_-cWap}=W@?hJU<+xw#PPg0D$3|)Bnz476sCPde$695D zPC^ozh~snh1?vk1t&mEAs|hf!h*0-PYxDdKU$`>{xy80m`LQWvc2LRGj$38q!Vmkr zNSi9TsOk5Acrx;Auq4L~hkt9bgp`@5NbbY<<}8KsKf>}*M$F8UcpB)*3V9%Mz+SQq zsPZvGkCGESikRQsvYx5XE=)+UbEmUPwLvd7$*?*pDIA@aKB}nPYWa7GfXk$7b|424 z=An?WaW?-F)X=A-+Yd%m9>Z=Ps?kBi;tKY`DJTYYp zkAlzJEQrLeUa)+Oht-$1qkbar-gnx|gO)|5=>ao_Uj8LBGta+qd645#`_NrywIxjK zsT1LOOVihObMF@oW$vx2VbzxqPAxX}NdLx)%NnlF9ZaHb9bfFxDOjA}JuhlmBKnIO z`|<80pk+T@32-HQ{cAM!4D6;0Z$XkQMRq^@>JdTsxs@W?;7ZGkDd`8YIve>LdzQEX zRtA znwuds<)Z$p*&5~1agrrt9Ug}rZML9mInC2;BEqu$!u8*p!BiZE0iM5xzrx6~i%+A} zxN`g*j;6@`xSmjPEw0ji5jPC1xR9WB@mo)j19X@lVBhjCGHBGmlPes-SGhV08;?A7d9LxpM1;ER$ll3W3aYg?Vx`Sy)Un~urtBMvT*o621!1V!pLkLr*$c7tAW~d zR^$+ocli#a56Gz|BiuS^mWtQ&KfWNhN(9*($*u_)zz4lW&_gcPtv{)sM;8U%PDsJ> zEqyK|biUjb#hG9QW_fjPVBAnBdmdu&Yy&N3T{K_SkTc2g;`6mz3Ky~+XA@~$b2e-{ zt3MX%C8cw1&d@3wo?&wMwN1y2*IZ<|?#|{T8&)d%{ONYrI$)S zz!+rCN_fP}VrJ?Wafs1#RBV7u@!P9Wz6<5P@vJ%*>74rW&|4h*yp}6L$8ue-MSn4! z&ixVE-SKFs@Yrc1x>ZhMuo`P4wRi+|qR}3#S5&@5Sv!m^skX8$4wKqS^qeVX#>00cF);AW|3@|pg z%%9~)&)NF~hr4%J@XC_YPpmgKyn%R0WyG)SiN#k0atUM~Iact|LGCvC3_~xX1gUNE zPmhQCwyaKigb;oZuxxCsCXYs||H@Bvw!5Kv#|NN=^2Lrel@)Q(4l#ypl$$qE5}5X( z6glWBt@MEA?;#j7thKPg#$tb>qONXgVj}Z>#-_e+U_c(Tidu4BjGPZ#uH2L!oWQtQ zeM-o8mzN83sq?fy2W2syaBso85fE zYMPr3ih8J9s=4Z5hqZRzlHn6pRW-CBZVSlB-=aR}wH+YsOx+Na3X6Bub=(kYw}j?# zlC+l$eVWrQy-Xw2FE-C{9oJRZNLiCwc2no|8ynOq?NHE}^g_CD7o5AGi!Ha3Z_hCe zR({sg8|Q)*6WMr=aIU9uTd{PevY7Q{Z9b)QW4G#-@9sNQ;)1hU~B+1RIEbgK^+!REjtP+4l8%%CK z;{l_JO<;5dv{`b%;1@>C0%(;;f)RAUdh=h)q)VAw^X#0*Y^G5HvHmCBH|10DK6YC* z2b)`lkvLL!d%o>#Rj(wyISs8R?i)OKUc;?6k|qb;VI;m`Up>deCPNc=ccQiO?QP^| z$-6fuHW##*TncABa%9#Y*)}t=ShX~=Q_eOabGowb%BIGV@w$BIB|JNy)+*%d*U3`I zyIDlCmC%9jowC01?C*20n<$!)J9=e;p6N>5g+d2Y#pL~W(!T!Ci`>L&EP(YYMVa!( zPO*{6$C!8E3=H4w3{{UG;Eqly>?-FbDQ=m(t^Om(Br#l6Z;5!&g4K#+yOKLT$o8s! zZzqxFdmzQkPd~JJOogr2@xc;0lPj+Rw)hp6HR_T^#q{{b!RCTFRJnVYDFK@A+2GF%0FnXo zSF zPxWcYIXxZ1?b$D45%FMXXv?6JmX`Z~JHq(;tx2wMgtn53j$Qh+i|Vc1+h7+lP%#60 zHTMw`2^I6XTzfQ`-2d;1U*$EUo#6b@ZpL zn5`=TLWkk{yzA9qkk_wuaC8)b0sW>0Q4O0(t<#r*{?|UBbBK+J0S!_wfY}1$9bh0j zE8&6l^wu9eh2N|RQFgMEq1BwtoSF|jQ*|S%gx^nHS*aPmR$E;xfPZm_lB48gaPBXx z&U}MI&q7yKJRfUIu*gNF5&dw%#5;-O>Xn*z>ri@ko<+!Dpj2l&Evz?{v}l(LW~oSP zuIT#yjknGJ)&kfP)0C$}C2v<*!y!58(^(JXWRfGeO*=+?cMZuj!fc|m*!Y(zSlg5; zjteJn`&m?mY9Dft5bG|vuR3%V+1aR6wlj~`4er-W=Gx!QwSTrdK4m&`h;KE+TA1_8 zOpUJ4bCoVPSdUO;WiNu3A5ik=NTd^*vLpyC1v-k9CVp7nosD&`$@%pn6FY+QJXYW? z@wAd-dsA<+#M#RgqOm6dyuvb@rk;6AlgA0eTBVkSG?kKZZs)q@!(AU4dQcW^Vc3#m zypn%1Gowj&)s15Ma|T>uLVRnq_Y#R$Q2G^>n=OHC58`H6+3$Z9Upz_XP^akb7fRH( zOHZGEu`^LJS@YXJfuWRm@>PB*zwC4i{7d?-X%@+S$97)xwh)J8O_4&?8TDqx6|W+w zE~KqpDHd|7;QdYjZx!bjb>Pvhcejx?lk<)+0ld~)3@@m305 z0|?o3Ko$e0%&))z7tqTCBcK>lFDulpzOIe~VBZ@6Ps3^w^kL*Zr=kJ zZ8bM+g%mfON}0%22OXBVSh?w6u=4gu=bf!HU;vN9@G5Oh0V*;f6) zOT7bmul)r5!sMaaCk(t7?KAn|&T>jA<)7sG&jJD_P5LR_#Qpj#1TW_95o;ULjT8|> z-CUIA!f_~i2!|0$2%*gree_A!*<$10?pa=z{*$S#;zlUMMsn^3t$Foyo!nDVh?i>h z-NOgQHoEB5%yWH@ja-$&O5vpy{f3G(@#MX=tbwcal;a~YF-O;mdWwkeKIdpA?RrXn zd;0VI0ho8-ENW!U$H`S^0RL8Ow>5RaoXsPxk<3xwhH5=e>+>@rfv{8k(WpHiR08$D zR>G;l|BoAQW=d;S^IAaTi#5L;3!hsHEwnl(Z}u>jh=9#-PeJ$!BEnV+pjRge^-Cz5 zpH6^D;fPq&)yL#PGR&y$c7A!?xTRp#@rC++>9{`g=|-$STpv@sug&-wzS`ZGaGw|?9oYS*c_ly7!7gh*OecREzb zhA6$)Iz)_!=~A?(kb-SY^SU}idwPEE17_FIU_g|NoE%WdE&-(t`p59UzM{8c!zu8i~| zL)j#=|6BOOm>^E%Zui{zq3!#3!D`5SPO8(9fz)$2DS8aXTxaO#O3;z0y#(ztSR6HTLR&9qqv{V!c`Rc0$Ke@TKn>SFH1TY?T4_ikQB1?5|ucL zzP_ngO#_BGR6g`>yFRp5a?M6ED-!8beykvwFKBf*y3U?5t{uNgWKm7{n@%#w%paDw zzMmFs@`*_kJv{TyC%N*RkIg#g$yzE4lVCENh~@Nooa$Eg<=N*{p%Dj!pZ>*Z-%`Iu zvnwK5doBCL%}HDr>njUnsfTvEI@?uA=6Z~Js+@BasAkTi&Y>&I+RJqecZU{R$+W#p z`ItM^n-8_z5Y&4Qn*a2x`uDo6zYdte#jDy(#!HCtN-aq_{mUoj{j#&vbS!F>rA~x_ zmA|F*On`jcTUDhIxk4?fh!|!X$U0sxJn98axj$`$BO|MV{k>j7P1C86R<3fkfytx4SzM198GA8-q`2rBoi2(2(S%l)e}u%2bzT zOC@SLIdV`egFw34?7hU(R=*(iw0FLeTAHFaiwD>?f{oZTFDKLL{izzv4^@pk^ z^c_vOZVo_pVfrkhKJr%=E3V9E3w&Z;TmzBlap-tQbLsOu*NX`!;htsa$!}X9XZ`Kk zevgZMpg!j(rorhcCoSi#Q)(E3Z+UDxb_bFmzCRog2z$ZT)^px=yA$y68xK zIo8-^p8IL`L=_czPUY=}Bp36tk#IjAuWqWL*^n-^`O+P`6vTX(*{Odu60A3%=9=`> zZEl23xc&m}b&}^PENmRQv*P5a8 z_xQ_6A0LQaZq<@ug?Xo*{*^S{e6_Mx+{`DooBhs9(1NMIOTAMJ(R~ySvlzgu98944 z)TNDq&(aoP)UZP7L3?=?;8RB5zJK3vJgyCjgW?(?NE`mu8f)Dohpo|~mJ-_SlL~t2 z3}MGDMZ-PKyfU$wg+H?UAG|K4k38n$ioZ3X*TOsN8;jwd+vU|R{I`tnG$RvCiIKs0 z`Pp;l%ZA?^o@pI5Yd~#XxFDxxrm2m^(tpIB9)G4DI3^Bj*16Ks{IM!h!6j>F%DQn}k=}p{aKz{p3)=>1JcEg=^zJ6p7#x69DZ; z_@;3WP*Gdnh5Vep;{s-@qv!DS}E`El# z5uXdhPM8YONa_F=Q-9%9d_Br=t&4dl3hH7EUVpwSa&}~#X7I;6Jr?dNU2JuW&U!ew za~C&?jOUz}h-7wvIb8b;5e``v6Ia~|9A=1vzGDW(B(Jx#s++fO&jY{<03Q_;Yy-gJHkdF6uRa9xdy^IBQKMgh z-Z-CaD)1bwk6>@0-}!XTWwa5pK1-M#D3-**l_uWjiIW%;B< zcc5t6r=0Pe#%IQnj`cDhYM?JqV~x3-h-(P+InIjO8F!`p(xRRc^ zF!2tby2;A85QKZFD7xwPholNOwVI3}Pwo+2EYa=pdS!7Bv7Ec9WIGgJTb0Zk%=TyT zMw%w{-302VN>KvxSCAn$+jK4ZM~VU* zso@%S(^8@i&^o-ZjZR%q$(v(iN^*B+H1NlQt!;{B>Ir}RGwBm|rrw+d%)ABY#?#r~aA{pBPc z{9PKUQ=z71GGF^xrf^9ysYEkS7`mu5ZkFHOH(4dL_olMlfP=KJFeN;Zkz!;zF`wcH zp_%2Oly{DHv-wi8-z1{wwxts!E5z`tul%ye)H<@LvmmS6VU0kc^QH978qIK8s}7Z& zwPEhqyhj-VB0o)~d<8ut`{k&Dwen-f=a}mX&03i0L^Y4|E5kBpps2WxUW5EpX^@Iox^yWUZ>F0n|e14W6lGi?1o)~6g zGaIXI&YkeS!^3f{sYH`88C}YqG`pKpLOQgI%Ki4?tD0?2*&Rds07;!3ry%wi3J04o zUP0a#w3CJIS0|UTd&=J{UfD0c1~J?J0Co)(0l=foxJUHr;s}Up0_LgI$XTG#+y9Z> z6F{sfDJe@!OJ!r3L89)&v1&WJcK@r?z&rds(ruN%7Xz}<)ZI3fMzSNB^Ip5G?%IuA z2?aj~bI!kULvH6#`bi_)2G#U!97L54&J06DiHQvyqnLljGR+#lt{Dg=dE}>)0X8P6 z_ff}5u-Z_ux51^w6;>sQr~G4Fm?It2W3*G3^Brb!tz0I>7Ns5B{M-KiHTr4O^AVb# zBC`tkrcp)mJKmB-$M~XYX%@)&R4gM#+Q3# z@ytcmPMM+nkjFsRsj3)v?bZW(alh+8+{akr5(ighlGz5Tvh@WUuC}V+&yO-Ecw*Q% z#mK7A*v3f7u$C^Dq^D8#PUlCvRk`kv`h5@#r(RDphXw}+Yc+YEfts8BLd$0W(PRa~ z`fY8XL(YDI(Rv=n>9MJnC1xuFF}8@RskPujgSoy<{^a4fC23Xj%&mmWJ4RPW^0*Mq7C7PPiAU8AG=yrwcf9EfZ;QOI}v+*C$?SalDq!577!d3)^yY_gh3er zID-arEI`J0Vc?c6c9C&$)y6$fatR0t%Yks~4Xjsu7}{M=Gk&-_lmEhKG;|mah`l5l`VCsJleft5JR07yGq{|7BIM3k}_F3ozVAG z-sA6pPhByRl04t7! z7ho+hnoU4|J2uuEF(eI0ltGjD6AfUliRrUj`py^G2c{X@)KVK)a*ndBCEuDkjlQTB zTP)IWJW!truFqSwys#Z~7a|+OS)p~kSEcg!*h^=Tn)2yh^45RtXVTB}H_TiDG6-Bc z!-Ykw_Y-rO4SGID?m(l6dG*jH*M^+7V5-}fthu6M5WH$hM z5>@bX+ya~8YCMi8sQ_E2?3Y->5d~AO10+n-;zRszg|f#+D?*$vDjcc3YW4y(SX5Py ziQ{L~RFzhCdrMp>?-a6XIJS;b!vr-Z+}|@XW6#QgI-3;+`tDoG%FXQr=rZQ!=Cb9X z-heOx;7CE6bwW#8x)aFT0&?lhlfvpbK>(o#*1)zJX-%H8^fV_5irz}vyJ|xZX#42= zEofCJhZue?b^KrrPktBXR=u)cb+IYhQyF7yD>BNLTyQ@j@zJVT`~%w_%IO>j{S-q+ zeyOK5WQueB=a)pA5BZQINlewo>cNYzL>L%)3HGXqj4AS52;UZWUv)b#I4MT`SEO(k z&+@oRUC(nVzM_8F9$)k-zx@-sP?|7>VyE}vpZQc}n`Fsk8obzs8XZs)CjO|fQq*3X|U zBO{KW#|5$-5W^rovGz~yhF0Yf+R6_fL&TbBEnX*bN$JEiUMiSM`8_x2pEVEr=yIRWeb#MaPoBx6yrH@1p>s zKcsnHk*3lqB=xMV+y7qS%cqM)(I2HVPM7#xR1isV29@yxV%cG-u9g)UisjaQb7gu* z8cbD%Q(WsRWR=a92s}gg?@7#*0nwX8g$YW;I8vt92^{Ao$@hD^X)oIDK#pu(XO+yr_nlH0y7KT z*bm$;+TQUc4HpsZA&_FZyjUA(tjzzzL+{(C5q+@9rgzcsOMh**$-($^Cwz8`c}L;k zOj+!r5286H(pFrRS1{mxhF+2NuBzPT>|vy@92)A^aBD}2HP)^u)NXXgnk!gjyX>Yn z7B?Lpq!lUNEa>=a?Q`BIuRlP}8(}2*8y>V5?y#&EUF?2>6J5mFq|RZT$w-N%L;l(Y-%EPPG^SH(d;iULsC(3pm=Vl#25oI{1#A- z4+juTP|!{VkU9~O*^wxOG=SJDD=C$Eqdk)PR`>RN21JmjXJ?>-3Fmy`nYvS>sAp!{ zD?h4dr?jTKM;-h4+p0v8L-*<&!Qxj5rsq$@&Kznt)~Knfmy5hlBF1L3@Kce9@T^_8 zAp&8w93qkb$~H4&*w9@kNv0DN)v9}5OB;>0Ieu-r8Qhc19h+oPOow8UI=nlh=eu8L z4qTcFE6sM8#Pnv4?fyja&7sYQ2Z^Z_;3ey&LA;lL(SOyHPiDlLkk$db8rfp`k_+nQYLj#ww#(DGTZo;E^r(K^USYCShg|9(3dY~wo;AzoK zSyh|Q`LDjbMB3uBp_0yXf9?8gT~Q@%3QKkG<20ijxAuGvrgt}2FDi?L8%}tvH8gzl zPXcr6;G^5gt<9a=-9IYmyt#|lVLM}mXO^k;mW4iU@oLYkhD(3+vr3Nc97%Qr{%y_V zF{5SCjJPRPq%Wl^8L( zntS|lZ8nW!7yFx+BV?oBWR<>?S2dp@B3-||QR*Fn51)b8>tGar4B8&T0FiIh6?t~F2J!Hq0=nmb zR>{KL9K@bFM39s1GeAgU#jE}1&2MBAQUnuA5r1SBwEIJ3+;4KNs_8vmi=eSx-ml1zH7DFS%U&R(<6rdDy-?+srSAm}R@t)arS$({5`|jhpBX zy*L{X)93E0L9Nf1LH!=S9eu9>AZC(TSruZ7I# zmH@HAT4!X@d%-r?(3|;#!z=U-5`T_tjlWY%j;OyJR`WZi@8^SrWS6}RS4re#ql_-9 z`10bHZaHm-N^U3&V!e~xHACJ2H0EY;QzXWe9iBNe8h1G7BsGsoGm zy5R(g1&vv4AGG2o&VX2eONaa=Xc4dZGg=i9i-U1&-h)f^tBc9r&(vRiCxi6eJBPaH zr9FcBR5VA=${b9%awMWw=dLFI4)%Df-l(zQp5u^zlPTh(s}lNN2b*ds8|MRLxh`7z zzvCNJ*w7x^&o(DV1m^NLKahMNj?9-!yW_>qymR(-3g3cTeU0XAGfLn&S378UiASYHgu;Lbt4!r9dkL%ubak4f2s{3kE&T+LcZof1((fQ(DvLXEFNLp{w z`q8u`TcNkNR5O85{>Gc#PYj7R;JFS(*?h$v(cLP(WP^@rL_jeu98H2pf;_B?w`SSr zp08bsX(<{izdaXw#=vS6-@MSbf)jKooA;jAe17hs2!pT=q;>gbQf=O{nfE*QN;yJa z23C$|2J1|Yh;2R9)-=#p3lS&gG`BD)uwllQNF;a4ZXOcU`rQR5EzBPW& zxn#?ji2c|6ANl1yOUueRe#Zpy+($JA1*T+Tr_SlwL0>}<1U2K!C{r7a!=Kga8ym;r zF{)(ks&q;FznAw45V4|b6dvntCj{p$tC7+dtTsbR0VTXUbgGlHeK&KoCB z-wvsnX`D}XjhNW)AY9)p6AZ{}Iu$EzE(UgfAS*J)UsQ4$_5>`6+g+8PC*%m#Z%KRr zPMHsEoPcZb5D*d618lAS6>0*oh}kwA0EBA`k%GE9@@x$X(3^i&4l%5G0DcX^S<5^| z*ZlGezl*?{!F%mTG1KHqBerjeGQM#>WoBSueiwR^#`?$m0_siJQva;-_?y;GMSCcV?bq8ZBTo?9JfmVa6@ZHHGz8px0^$9C*WD1q${ zMmW~CR=NI*E2umi*;KyQM%VHFt)4>l;z&7muSVzqD8DaTn$dH4RaKwYro<=E8k!;IeJuM&RPFnwo-?t?Y zhVGs}IQ#(;8SAG$m`wtVK7@21;H8G0;6K0<2MDXkNJzl!SdA6n&KW*_$;Kvc0Vo2Q zwX5EzEm%^5eb2D4-~BkqhY2^qdE!v>-KS?{u#&|5p7QFZYTg0Gl-GZv)I_7ZMeMi-pDdn%mFUvQ@3r^>_=Sj`W@sJ3d(R zsT2-91W7RX-ltw)^k{ezWvewvRBx#FC(v|s58u6v&`U#}-ZoN6;k_4XfvLp#hTMDb zK(E$e1<*s1C*QeLg}cE=HwnwH^SN*bh;`r zih<6_k_?a_u4zZ2lsUKei&gM=_O&B>vt@p`cnr+z!J!X=N=1&Qiue2 zwdYlt9;YD#-$=Q+u)JC2lO-=7WaAg#oh(Z37!LK2xN_!vOwgnVP*Vp>7ufz|gG!~a zA~*RSoDl`qV#siq{J0bg;$EmiPIMtrag52H^^VT&OA3OcsbC6A&&*+_4xjWUl+&}7 zS5nsy?-pOMPywMBm5_@S;GO{-jewV^A8=nifBqa~AlTz@cmp7WiK3I2*~wMT?gvtm zbxoiuS4037=_T%lG+Mfn*ZVFgg>0;m-z=nt)S^UPxJ4fHuhu*lLy2?)h>wz$$6W zy?OIH2%EG(Enh5678dBRK zv-JlpvpgMDo&P*^i$0nicd3D4+*?U^sbbDgQfOY0@|XJuZjH73h0JPkiXot^ijtYj z(&)AQhhJOF71q1nI1P6m>x^M-9`GMTd&I|ct^FLk{J?&r^8Bbupi;|2b84{rcsru; zm060`T4#Z=L#^DrQ`?B0n^a~7v>58oliLHJY$3}cXnD*mwvt)A*HNn-Y>Ea0FrowcMeJB_KU*%0~B?Uk{uR%=)K0cqjX z>e1`$QF8b>Lf+MOAl>9#<=aQjtz7Q9=?xlNikyc8BLk|fwZfKLB%Dpuf0A%S=81tD(DQQcw3I_*KY9OO7N{LIFUHQ!_UWBTThMc~A=)fRhZ zAm{}#diuCF4f^XDjpTi!fN@3TiT8L!@Vl{=z-m0(-QSm!l|2JR1~{MrAfx|y+G87- z7632>JMGh)f&x##-2!AiXjePn50ZQCh@wMxkQs7(*Ku)xsLkmaIThfYdBQM9+tTYS z*H;Fcd)2$sF!eU&BEIwCP*T3!)OJ%P{>bL(!K(Xmi8uuoR{f|C zC-h+pzu@~R<@$U^=dGa+(;nu?1bH=fqWje}1N^@FW(Ia)g*GgKg#}}#8;@?82x^^g z{WcI9xTC$1jZAzQMZKxi+qm!B=*?5Q)ofF1#3>DWL=_en`5y*ww9&HlWvtubG&p)i z8Jq_n#=F$_tl2$-U{l0w{{E@+Wvkp!W#jTEXailOcx+*j^vf;FtnDmi=2eKywab&` z0x<@_1Matx+LTRf#>K^4;>VAplRD;Mwg|-2*^zuzYw{e4Fp>ZI?mm?2S(EINb51a9>qM&NZVR7BUU!e7vv_G* znGlnA0tvQm+f z#@#TVC=Qc_`d*n36VkO)nr~%VJyup?tI+)@Z-M0jx|=$J)!j-4w#u)X)W%gz9{w|k zWp=a%!7}n9fBKbLe^V}LH=Y{OOwDcriNzfw1$+i$@F zT}OB%6sF9Qjzc;jWb7WZxkDWRruY|oN_QNc8?~1!ZWS6E%skY7CDj}gsae`S^BsoQ zx%=TLNLHlDJYR4?JJL?TM#rG}&zg|*U(c!_!oG8J(AAxON@&A7r7~F9E5dAn!KB2m ze5;^m>hnYAj2@;QDDl8noR-$)MrfA^)qt7{O?$^(?@2bQYgAZ}@BcV4WRaepmev_C z0nn^0L_BwJg58bx1@x?}Ge19wJ2;en`XnVI(+LPq)#!>zP%LRec!>B>Q} zXp=Mh`=nwYEuTBs<5Q}mRX3+d{ix@1SDu=sHw&-qKMrBKUc8jJs-ML1m%@?%gV|&0H@C#6$>hwok3&|=p)y!kL`49TwuEi=~`VYKQnz$pq;xP zLDdIOQSVTR7iO2W_{6Bo62%eMCSFB9@GP**2q9V9x zC&c)8sNbi-^{UTtrbDwHml_@*DRdII^2%#=T=kZmS;YEw(jPfO(|irps?Fx5WQX7T zgF!0~)U-bJEOp?0JIwChybV^9n`#a3ZK;;#eAG&1`47x*10NrMCN%;>{}yE29uTse zZ@Dx8!ArL-1s7SUDB|C;XZpA|eri4p{O zUwtvcr+ddYu%U?EXI!1{;kBcglEg$~%XV7%)v!Vb*sIFN^1$Cu#$oo4>;F7bwogT- zXKqzLSY(X;c-mFAFE zeT4Y-mrVkdywT>-4bh1uC}DJiUhRoc|Kpp5xnKV>yh(PLh|=2!!u?S@O;aaf$!{vSv_ z=<2OKrF=zy6drYS?`p*ZRgT=J2*>byrr+T7qldUc|M4n z$h)7iV|riWH7mjy`vA&0i!f z%O%Xp!K=mJbIk+cHVj*d4#N_lhtlfwX$$1ID=o+NfHI1#C@ZKnXNG8L0Xd;*(2EFa z(9t|=5Nd)$n9Iw{zcaOMw1>Y(J}jt-RdS40W%>hxUq^ez^n`UtqII8XHgz}oT-rB0 z!N{3wBbSq}J#R;C^~DK?8sPlIss8k#3jtzKob2!SYh@@n}KJDUZ;;#c3Uuk3%i+z=EX-Brm&(X zw=PVsSh&P|SU$ZX_{p_VK|36_Qd>=V)GS%LNi|g(%e`df{4h8lr)JSD>2k7)NJ;u~ zcYTn6p(WLHh;#aD!_n}%{&#$#5?(y}3pj3ei{M<*NK;#o1LTlJh7Ar<1~6j5wRD-Qf9%4|G0k8o7igtKHu8MAF1rst@%nUnq~9VjDwU6!5^W` z*l%{$!BpCDTRPK^cWo1m!z3LRV+eu-$v1;?#6YI})fJF`NYBVX%S(|7RscAHpPwHn zx_B-5P@b3u4z)9nm9dE2Gqysu47A4B;1AScEWx}J`nA=#jlvV^QAYe21bN98U9*~i_Q ztn?DQ{@?y@EZ)-H8a$|cBQjD!Bc7i>(b>2QW2?=v;M;NcYOdybkWjwYe44fJjA9nw z(`JF}$XQ4i0zaI+?_WK;_mMSLo( zn%`$Qw~tqh*1XP-B1F?Jyzuw?*h7B?rChwGc$0#ASpTdGowwt!O@mN4&;aj=S*TTD zR=QYiBb^kwxyhCr@aJC0#()!yTcE;U>G{T>qR7Qj!<7|B8~~j!m|9t(y9s98*9%KZ zIDmE?a39f(>ge-c@QD}=)mk$zP*Tyu7m{6Md1NJWG5aN)kZ|MlXFWm__Udae(okjF zooTyNHPjg#{h5_(^$3;Xc7x6#q59SGaT4)bJfzBM`}u?yRf0rOJ5PUiai37^{EDsrsGyLWKZZbEYZVeT-yYr57v9-bXrGDFkFSkH!>+Crcr}Xjk5~>q)9HI=WWT9}Gs^v!3!s(L>Zx(3o|g~t207qc zZy;-qB@4dr9-iG(L)K~~FdkUC06z@+Z>QlN(3~G2W&oNmAZ7qqIa)y1oaoPx0@^BQ z@?p76)oc*x9h1-K2ddFA3IyAg6Iw+HQB0^g#)4o|9#RyJH5o7H!j}VW5?JN`Z|z60 z7lBkuiptiq>Sg#OJTILG*mk5#|4p7!_Ea4fM`lm;uB62+<%g@E-5J&D1k9}XkT%g4 zzFJnP$7$bVaGz3E5<|@GNwAG#J%bzMa5EkzXWb8bR}&QRlQdp&(<+C_5Z^<8>Z6;sg?;$BLp&wLvmHJ!kz_Z~JVQXrvdPu> zmR`V2*kWDdaACZ0J?jgjKP4EBM%6`tI3JYf1=N#a?whTEwGK900QHyzx>AK+fbr@D zxg3!8RMXO8EH8&_HU&P&?Yy3LE6v7?UfEg3YNobwujK6I<=bA(pUxZH6idn&tiXeG zUu20og2o{{aZ9Yv3iVB9=@4j>9iIepu51zLaFC? z_zO~Tz-ef_>K<)o3m4=)OZXs%W)Na%--11+`~LlVh2;W~3XmX&#fx=e#50O5A7dg ziDxR_u*M(P!Ua`_c=ubGD<{84MU??SGBEkgpKxRwOn@+f*C~+W00D%$tee{ja0wxj z?K%%ZVg%%Uh?jhNPumyc;-F+=Od-qX z1$qJC6v@8&oQ9T_^|JEsbn&$aDarv0!z6b2V9RaBvZGi#J5d^re z!C4|;r6?$bO>O0D#6?8dg2w__&+8BYSc_ZNFKKoiY}^w84H)3vOG--{oZ@+`;_)n% z^M`=~0x^(DL2JFhYk?Sy+9?WfDcss>=kKz~gB^2Y>+Yta!A0H+HomM#PV_YuHhRaK*h4i66C zU=NTj7f1E?I=&x5&~X@B6Hwfm2GzC8R1QGHW2OP9E{>JX&&>h#q1!-K2n3$p-QA#L zaF)rH(_yK*rRDWrISK!DNj@nNJ4`hBJJv?9T1%W zs%%O?*9P%}otmFN?wk9+|6l1^LqPH;AV%lpi_bS1NpM+Kzt zdUp-Jne;W^g2)Oz0YUFPn1JA1-4uWlvlWs~kqwuCMduEtedGwn`zSMO>*-1xJ;1A( zvW~q>1kmk$-R1Rl#=l?B_uDlDmQCRI;vT?%fZ3Lqn7FaA0r>5p8&MAkkb=5Ln!+26 zbbsrVR}Zz##4cP_6s#_In_rv$l>AewYjF zMaXk<^g}(leyGs_#==>!->?`d0RCpcA51AMkU`vIu?{5AtTpifG3Tk_^|w)kdWbX$~Rgd`SM%6_l3(F$BPcpZWFc7w{Az zVPOP7rF#g8+y=@cz;D1*f&gUyjg2-y7GH1jKsNiY#*;y;HVZI>@(K#Vjw^k@oPh(# z`s2Ca#>PhAbJ|*3q-Vu%U?BbaO(%iH{O}=d=x~2O5m>wZ`qvEC**TD~`a8!pmKPVD zK*#|!1JJe{Na_H3$QfvS1}1rZXlSUaif@1fLtB1$co@u10OWr>r9=z>sefnEf1+Sd z!K4JDPLbcghdt&09v!W!r`G_U9SAA_H4vE_67&Uh7=L8)3JNB{taNj$9*X&G2)jX3 z2Z#SRoz$)kh;4Fqwh4e2dwP1JUwPpNkW>Y?g|B8MWVXZWjyG< zbv}bnfIEEu8w6os>_9=?!7Ce}{Q5xy6L?F2xeu-n9tB*X<%)= z1?(3mIMo8P1xV05z*xyBFIRQX;IsYn3S{I%iU<(2|dryJnBrk6zc$m|{i3ve|{uAJ`Rp{{ko&JrD*Q>}@ z6XwD}gar)DS%s{r=?b_)GZ1y~WpA_O*`8?-0edna&viNo++tROb-aL6-t|e{_4~bh zU7HDL$H?^tRMIw4Xd6fel8JcN4ZO)71c^>b35k0c2CobZGWH)vfTL@mM|nYvhZptz zJBQnzApoHNJ0|aUuHz*l4q(=RB!(1_)N*z{1e$(eO@j)22fz5n92b^2AOp3B_;y{pg+}}|rCbmlfG7m?awY(K4m2D9%n5Buz!-cD4?o(Q zGazCsQB_j|s|Q^zX%2#Npibom-r==rzJ9XCN3#izU==V*=c<6o4CGKQR{Ar5Xrqm# zB`9|Igm*iM-9Qo)7ihBpjj#CseV0mFG?BpN%e%*&9UuP<-Y_u4uZ0kO84+Dw=-5k6 zS{h8kDm+Y}Nri?}5B(7Ka<{iH0UL7+Nk8BXH^90E6Gura?d@|K8X49P(K2Y`iLNd7 z_{4W=l>uzC1X-9K`JBwPU;V?d3yiuG13mq31VR|NX)wRR97L}~5MVR|s6N<9qGK^K zLC{lg#1R+Z@G1_-1%b#VP>k~W%FUAP_z&}dl&CP;=3~Gr-Y9@`j&20v5=cj&ueH*b zc6@vs6@?cnNsWO;>RfvPB30O~ex?Wj!r9cUW&u0*+mPdWS`d*(&3jQ?iYH@8(yRvm znyPi7ef}IM@H}{LxTmE3>wcDjP}*#%{F&})iXwP_7p_<3+4>nU+^W>P!2=iJJA9fuX-cJ=c5w(3FJ zxde?J&m;2#vIhn&US_lZ8PQNQ{aIeYU6}|V^!wspxgKbT;C)|q7f^vrj7B!v4tL_u z4rp2x{^wVo2PN9R&8brrMxUly$FJp$Pgvf(wKPkQN_xfp0OJ*o3G;yov;2Sm{N)pp zr!NC&{VOlh+y2W1{JDwYzPs~?hzXXB;j8=dHlfP(0ApxVU61VFAJBpHwMqK#M+_q8 z>!t8NKQ})_hsOW&c|ZEUdTZ!E#1gzVB>(-r?*~xE_}>rUSGJ973-LdH5=8d@#hVk| z03mzk4A@VhU)cW62wWTaw!Y$fP)Se4o z*=S1#)q^jJRW>zh{pRDvC$`o}Iv_wwyo8;h2jXv9(6=}!AgEuCjG$ECI%oVqs%teg z0Tsr5$XjUMX*EdriBZQ#1^=mTf~v<0|H*0PfhRFiZeJkZk~Ee2>0n&{ii_jp!R!-J zs&_DvhQoTBY!#h8h{S;E-5OiCjYdReSy5q6Pgaa^vduG#3 z8(If}j!Y~htFesD%JVU_9>K=j=Va}{S{uvc;mMrqQeo}E`KV3?gb-{10q3-`i}*=k zXh=~|f+ZaJiYzBMc=$z-#jyP{YcQX6pfvn2xcu_Uf7NbUh+Ng(V?8q^^`70m$k=SM^CH5APO>c*jZU#S z6Cpbn6>T8L-hZ)MO4%yq5~U#pF>ZYHR#C=%D=olwcgz$;dChSTmvaL3^t`iQ2)~3FIo=Fh;OlPKv1!Pw!XQk zv>Lp2m&}l|*bc70nc+`G=crJkOABMx98C=Lxtujr=M7nej@>DjxEm=!*WB#%Znmb1 zPA_ynpi`ETFcwzcqvTZ6@rTizRK8GshezJZ^0iEq8hwy-etF%wqgy^3alDi{gKrb< zOBqs@am!O6q2gZ59g95Bg170LmPPbn92r7%-j+EUy*n{ol;C9of zUHRXYvja5e#POk70d}Ii2k8o|I#MG8$fZwJ?3K ze5pdy=Acf8o{7MNldFpXS+h^zw~a*C_q1S|(0vs5V}p#40{6m5)mrw5HoBiPBOPCi zD{)e{&r@$aFec-lp`I_{UxNoE*Hyoc#4z~NH>7fLZW}tuHk9T#A$#J;qlvgFCl!BK z$X8>cT%H!>5fDPLf*()Eo2@?aUbnOlKUnTpr0k+)qi5%I#M1Htpm+isej;XdET z1SiCSsj;+H(N1;oO;PGM=jP!`Sc~5amPI|t`$~F7qtsFM;jzPB* z(jmI@E3xuq8LpyMzkPWpN?AYP4#zL-8F158(Tz(`cP@Y}%Xm3;5wx|giitnys^_gbGS*LXbDbOht<_G451eSY%@v}AwTqtg z$V<4j^BPCvbAA^9IRr7Zoq(x2))&iee@t^QY0C^SbMn@0tLu(cQ)gh|5Y<$t|14TZ z!QdFVl|-5w2^zIF^|boZcd%vNgeKq$Qoiia8Y1`J-2N1H;3K*!RJ2d|i`^yyF=5Tc zH&R8tZs@fBp4Wk{LjRSJ-S_--eQITWr&$4tq!cA)p!kyDplU5@X5AVnY`>4%m5k@^+^mw*-h2TTGz~3w21k8puA#@&1t}n z?Xt9_ca6kRudL;2jS|**P0`+)knrz3%yQef&s;V>eVxmja%*ir)pcA@s?6@iR(E#N zB7Q=O7E(`~A248G;ESWz$;&I4dy*Mwd7oph0 zHPW>S@c1WGI7t-IdNAuJ}cI-W{aboe6$I>3qYe=6K0%?N(B0dWe{@tE(~! zz*Bg8f`}9JLVKGZJjTf{^`dpQL|j_Q6nm3WS9z?;P)GW9k>#F8IC^|2Zg}%J&(A*BnA?{XweMer2qkKWAy!$Dp!+>86d{IE}A>$UWl^*=$AW79L zQm2M&1*PSf!gYZ;;rtIqMtCd_&sKdkZUEcZeT7exB3EDHAVx@saR^k*yZ>Hh*G1|K z&hTl@;)Cq9*IL$Ab_ba!iZf(T!6WjV0rLgIHM>f;t~}ZoLPxI_PDGQ^#RdnminLw# zp;u$|8SNxR`)XKqv;$J9Z+e9qJd5kw4f_ z(LcF`+7`SbmK)zx-uC!odAYR73S%iClGDwC2t2xq+gf6%aGwZE4LDme58Ny>(z`qf z2>P`Nofhh4C9c$Ot!tgtol>tnM`&tN6#0C(`27pYgO2G&QTw2Xs`9!p@LD1~1V3x8 z$+CH9(cMZWPa(m@`=yPQirm-vT4Uu3>SL1e-pRbEvj)^a(>+8)^VptW$wp!)gNTld zr=fXOR0M-Dzx^&k$ow&rFE>Br;QnM=pbT8H37rRh(ULMqS1r!6mo!!DK$By_B4S|CEXm$CtNBk7`y2 z30u{?G;H~GV$J4FE_P{oTi?taPAK^7HCW8N9OWX}7F3~`hVN8YBT$87Q64|J@~bLp zRI?5ahF8n`-ytOq?&st<;s`q(h3hA|l#G)&;1HPLbD1@X<7P{>oVeX-QMj5%)#vN# zHdM^b@nhewR@xLW=*s{4HQ&BVo*?ahY&!j5#@u-wF5&d<#E8$8rIU$v%ZO5~SJy0q z&dQ{r6c<8UH&JsEkd0P7dBo2PeJ)oWO_E#2(i59D{00}9m5$|OzYFklOABtNoi+K7 zTv|G59_@S=BSbl(`IK+t;?(1{1X);w^%a@|+!(k2VFlZB)>WaMu(QlAi&FZACKXkW z5gLkYYs*q>LqAy`k^TE`6nrSn{Uo_!2~${#tSZXf%~VG|>gZ>>E%Pmvb1yANZz2Sl zo2UopsOS9ZqGpv-JUXr2*r_MvA&vOsUq$y{QywYhKgD(^q-?0FsH)1VlgP!!kqHTv z!KTQTmG)Q_*}OVfOG`G30wp!8pdL6(@pbhl#^9EKsD-}QtMffLr)RFz${vCu;%`W` zEIU{ugeRocZ&+!oEF@|aDv6xLW!G7Fk>6op*qgD1w~W=o=G9`O0OjHTzM+)aW9AB|YcJohFMkTrCwE`$a~s)1HF z#2})UHYMUD$#-6ei&-xhSUuZqEFnv(JUDxf_~PifsH0nxQPoPY<*nwlAs;@o6v84a zy5Deq*7H(*ce;){Kl+OclgH}Ex1l>ao>$%7bDMF?)}<)8b?e1FM9~>(<<(4sUgyU7 zUC*A=?wkpNOE)8x#L8N3{;e)W7h_JYZEv4VSuKS6hF&cSOoQ0EvbRNFqQ3+E)nS7$C=P>YL_SbAYMxR+L- zs2)h&)v7wDgGcDkXX-(LxiyC#&HXdtSPaPjashSjTX!eL^_}`kC_I zq?nECOh^6uNBG>IkLr8pu@_ua{WgiFIpbySXJIuT33TnFs%vHdZ-D1$zSq0B9+eU0 zc<0`i6qJ;dbCUYLm4n*T0(W?7@5jUBh?S|tv&HclPiK~P#5eNY<*P&zmbMp&BoY9d zw0uS(v>@!TmAqQ_m3#ZLVVKp?>mTh}2|7>e`CIwJnV~7DiL;C{d-i_{O;8P!=Q3(S{GdMqLp?NK?ACP;TFB$% zn+Ad9rh0bY_Eq!sS|&9p)Z+{#-mg$esm@lMdj(2dc!@K6F~Z-MtS{{q;P8O8=*Yt9 z(zS#<;uee-jU;QPbP;BG5S&uqLF~XhQ+#pIy*0A~3FhR~9h+KlRvkZ1P!o{FR^OaH z>_p^|{|X<;+UTu^C~s6TvV(5Gesyi!wVze(6q-AeI0P^4CY;7^AOo;yBz2XwU4O0}y~J$gh>dwh>GTi+4R z^u~NW=T%=`KKw#DMxyBbQ=mcG+--fS)f|)4@iT|@RzWK)6@fWX1Bgo8YK59faqF%( zP7u}19GQ%s>3tG~(oTtUBXbv1nAH(K?msL%O&z7f2*J8x)`j%iDJ`11dU=AJIG<`r z;lgXUM5GynUa%rJZ^xc_j~=&TPOwlW{Nmnj!>O=Z6D}%4|61V0m7$#7+yA>|7;V)v zcNfNoF`c8z_QS+0d5pX(Ts|2PY)1|CSr=Prj_xNA@DH#|*o$I2(TZ4)Jl;rdNh>($G5ow`Y}Cv$5` zr9&!mKWd*bwCHr#o4d3Hr#KZ685RjwZTmNyKx33#bhJ|~pC!Hh)H=yKRm|R&*p2iE zI_9?GW6Z9n2@L8>u#NiIzf_o_5ziW9i z`yhcg<|yZJ;XW!fEL6MB_A*6XXb*Wt8>s!>%swRw-{tBgEeB)u>~!v;7hdP3JPf^) zD0x>a-{Q~Rq6j*xj#lRq2*qp-*@*V{hGNC1h!KtQ`9Ee-vT5>(3!Fb*2TU^sl~769 zb}=%k34SK!RTXu}Qncm5YDYeGblA=+%(FE3Nc}dmUyrcRUWn#FaOhY2Fi$(nnccbR zFN}v@>oTgbZ#L4YdGFXR-Pxvb5&UITPXP&jz^iMYmzd_NoZ@EXM9WL&=zV)<5E|g2 zmsD6)mR!}Zl%TcsIn|=cX*Lgrxj1dNi$EK{i8F7i@6I1|XN!MLXU>dktGu_j%-#;D zYV5kSr1ruH#<6FCvI~}$)Yup|{_;fPKY#X;3Nq|+IVi}{<=#6yuC3d@u#(m!IBE0j zgNLSc)H=gR;(8WNm}nmdrl9RW>(Rj20L!RgIV2@jdCCOSl51^4IeC9IZ6H~v76ufn}JlkP?GGLlHs53oib=F&m zLiMDrdSu9y!P;?xqBa*cMiroxBj!m){p9nvW~Fae13)$lYw(5S>Rh_myuSFn9|C?_ zhr6wP`QiN6vM$1rv$IND@aS8D8e&L>R69CyH=%(-fX7^x)3Ib&brWV99K3g5TKCw8 z^|jTwo#Xtk5^sfjacbGa0wh+$t>wekl2Jcn2vt<4_(f^!#G1;@-8NrIp1!@U%j>eW zfR_|Lu0@ej7l8tWdsrSzeiubHJo}Lxro1t=9-$AUHV*xo8LE^1*w5g| zFGl0Xx_jj^z?f|8YTE|e0>?3C$;h_~5#}Y@Cxul}$vNlBO4*3eVe^Y9l4yX;PnRQG zIe$(aOaAQ*X4lx^mqEc94^nyjEtpJ2+u))T)jQV7NXCr0c>+ab;ITWVE8}n;BK*ev zXq1p%nu-Tb&gHK_X(258!=GskNS*S3az9n8TA6w4HmQ>mHc}P$eQ-LR13Bg!B=?&fAWhZ2LRy$Xgh7I!_EN#rAVh7aom4Ano zYkLmelPb(|LF)iv1o5JUtT969pL2!KAFlZ|O$K#vHQYlPD|B=R#}h6sR{5~x7t~zQ z6*Wz0ZHCN1u;i`%e^ta}nbW)ep$)tp752Gc->6s{>NI`V#EOe}N>U1vkVl$S>WJiy z{U`o(`6`KlrA0R5`$GR6&*8jI5 zu0SB(gLiq*&r^Zw%ttA;`m=)$2k{T&sYzdp`KOF)O*-#KaaE36~f6!lX40B#}hBcmUwY^*yZ@xUR6NE4}U!w z{9{}aPf{4m_G=MNQ4weEfp~Egch*rtJLiQb#B$$RcK6rJ`S!=;Tqi3U@!$z8EUjj< zNGp3^u^cbeulc3;{_9`tJFQU3)jsz7ck+0ow@;SaZYjCt>rmK0;=8R`+nGo5yjcdv zM~M}$-m5|S?g+bjeKPpzHSHMR05y1Wv4 zz9}%=bv%h=9QNiw)-Y{}+=j?AZ;8CtIn^~cmHFksEaKkA;P3 zwz^6-bTht8jA!Qf?-#M~(7Y(9N?*7?0)3mKLg_v49%~n3KY1so@|AEa#hyFLMCrn=vuyF~IkHy|M6)tv0 zB&E!nTO;18p6DK8w_vceIn(26Z=IzT_L7D*A!$<;*@@&MDr>)9Z6Qt9J8wGH5qC-y z9hET_ZJ)|3Cy(W%655$*j<_;LNcL`k_gH2`79iK?(n-9eWP#LaTjSQ zfG->~nY&!Mn#c-l=_W<3E^PukP`5gf8&@xlV`)?P47djqlR z=w$}c+o`M|lK3$aGg&Z@7pLrW8W4^6-l6Jta^ggIx^H7qUSsX7;yBGU_DN=pcxS7o zxOZgS*v?xA<``|gbrxNtO`N*7R+^5eetmLDQTKy?4O4Ap2tcOM6qL{WGzE_PCO%B&WN3Hw@D)6L+JK?+X63K+I8j8~wC z$~DXhjK)y`XDAce6StbTRjUNp{WPjQeq{eYjhNNF*{?rc{Q zI#m6pN~NtQX)9!~a?$5xThG(E#01W$sq)^Bh-%q`?r3Bmb)&%I5oRs);J(h}W{j~1 zvSw0NzSn0c(5Z~u zN_pM*V8!mF2C+6{u(8Ay4SO6v6{b zXnAS)0Y=8;+jPR#R-5j4b!qXte#uJ2fAD6eCOB3dtU_F zp(8&^kNoR|b0PNIJE_8p#La5nVgKa-4ovp``Eqv+)5{9j;{UL1(32R&Sw_5hP3W(gFigs5%g>=qg)F&C=l^ zFA-TQykGrQXGwQ2CM-5~B_WM57M6*aA05>YG8WbV*~ySQV6p!7(vN|c7Gf@+z)7xG z+^)i%g6P)4&#CU%Lv20mlB5p^I)CQ8o=Thhs(Ymx5!fOr!QT+e6P+M{$!2ZEz+&^{ z9b`5;0e2Ce0YkcsaagISXwUds$_^ImBejt|p$B8LEFtU+^eVyvNam12waBx7A`eDm zdPFRej@qLmou<4_#yT^9*?fD*o zk&Dw`2AvRsQwe*rcWc_ZCee93*btvAa#VZ%{z2x51a#$;30_WrwPR*cpu1X_fV*Cr zn}@{2*VORrj;dGx>Y$nn$y}FBG4Z19{^FFtVzT{|Qw;XBWe2mcr!~m<@gid;x#YoS zTA2>AdVue>j%t52LxW)2fx=R(V7;xV;Ij1-7`^9 z)gs0$2w~ic96LK{UMFVE19OJav!;A0Op%4#i@{)F!WgcrIoj|PKeD>OOS879E%F za~_@C2-)bfcRv&Mr^d?fo2g_In^@PJ>05k2og{iw*?8$UOdn_DykNjys@VfE`6Ea7 z&L>senBGuZ^)4i&Y}@qP{MjG7T$kynC!>1ryyC{y`8!zSizwrGcvg4sL^i2h$ft%R zBt)#pZU0nLJ*3e2Nw7V6UV6p9Hkv=|#L?{qFuii_aryi}be`u)+C4GPp zfSmq5QaIW!R~-7lek2d{PsZou?-16Dj9OdR$^RiEgQ^~RDx@XU>3_(XxH(j@L=)`5 z6*Q)`<wWcR-Ay673MEGKb`fv$c@T4YS%KQy1A~ zn1GXGnYl~rAy4gnHN(`FlQo)vkrJr>ay5-zqz~*m~L8${poh%@3PGNR9V} z^&)KqogrV%*QI!@qbgZd-ay?PMTSk{B~R~Tf(Wka890!`U@LrN+W51^Y65NhBp%H9f=MctWBP#ZxZu#WXk}e~@w1gk_2`Sdc zhIpI&VoP2@_))snPQo(sWipwIm%P}A4usnxwEgHp-)_M6&UU@#_mJQGaSrDqns=ic zUQ&h+8Xf4yMP`w`v|@WmRsP$8T=Og&>9g0_7RcX59Pcg5l9H=~S~(>|&3iSJ6B3S_ z2j#IiBJs%F4i7kLRT{sw7x~!t3|77vd={LXGE%$75xLyEPsq?@fc?cQpR0?H+CGz% z4nuz2WyW*$la^W!caBXC{tFw8@CaqmrOMD!eC13SLv5PcxW_flcFkxg-BuR*Rhnk$FG^Ebi zyR=t7Q+CeUj1IP|)`uVDk?heiL|TtE*bNF3BQ3mR$RFFTPKK3E9G&T#*X%PZPSxs% z`Rq152{B7Z`K?Ee9DB)TIBOPts5yO4wKGfpXu-6I;CPYKq-dHNVBc92d1mZrPZEjY$8Tm+on+7P)-Fj7tEuvT?T?0 zr^7_p4jIhSb-;sfSA7!_TYMa)Z#Jqo3oq^q-#Y8vW@g9)mHxG+5Xl-mxj=RGq@ZK=d(tWO&@Co zAIbOId!MaSbf%n6yNkarpu9NI%eHxVoC&K{M#g2NSS#us1+Y_fMx69E$j8QH_E5t= zb3t;s26bIpvLMMY0t_1Xx-;*sY70H@OMkC0yR(PZb;3AOx%qP{}pl(Kj=h?9HXOY<7d9Yl&ihKTX(Wa;-PvA4223AM?S5STTB9;$Jwl2`sB9hnG+w} zcA?FfoApI`XU5-L#uUjW&ewKD6-V~9=_dalvaT{9s^@F3iV6y1B7%U5pmc+@igYO{ z-JR04U=Y&XB`w`74T7}P(%sF{UHi`B&p+rppLDr*?#!9fb7r30rQAMWf`X#rtk&kyp!3ryDmk|JPW`k0GlkZLn|$ z?1L_XkXOupQp0RHGb1#f%ZM*Mp~Ip*L&^tTG-Fg0X-k-EiGvE{&CS<5@Uke{EMHhv zRnA$b$RC8fDXowpp&qDm*iXW4$UVYP)qjxPUfZdc4{l{K@g+!KT7v-u7`E7?)x60ao=^f^ySe9#79?ED;at0>rw-&?9|5Wdol2hCh->b=4!p$ zj}}~xb*Q{6nU+Tep!;RAFz%gRxs+dh{6D%zthx*q$i;;xcCeM7XBeyi9mrT)m&Sc$ zHHJR}8{Y@6{dmX|V3Wht!3}qbEq;DCFFf+QiHy{AZj*H!9hs@z%;KHMn_#cH; zKiN)Sr3iZG&oGI`_xeG!nn;XE$D2ZS3UUKaqx*xdWKzx~PxVBv80jm%F;|MNlZR?g zM`r~y(0jRG`9dmFFldBzZ--C85BnXZY&-4zlw{!KHu=PNgYT&Z-@liPV@8EoNb52-zpK#RC+^x@Nx4$Rjr0}At+~21$LLY` zG;)2M#&L@w!9rH$#B7Ow7H2F@Lw?esLVmDG-q13Msg^K#5@$hIuX4C|V=Gx&oFfk$ z!pSrnwaP0SVz6(E&sihexQFwRi^y?r0-Q*Z$#|cu`s3mS$jdc$aLqdMWDIHF8PAKl z;So~}Vv`pZZWb2iBl0 zhvX!co*#dWyc?n#Ww60~Wj+nYdMWu{Qi`OZ)m={ottZwx8aUxWQIXTf*o`D*wS>=A zKS&w~vRP0DID*_CB$=7*w{ANviiUW7Rm^OqZ~OYtw4%OvZNZXkw6mDHcZJGYY}c4A zZ+>5I5N*I)wnc~3#BQ1}(E7*|C%=T;hVLe7qhk8?f0myI{0~C=WFy?QAyy%(TSDQz zMNxn63Kn&vRI*4{1PPl-#48KA8?m(P$;Z2CLal?|*cno6o*tEq3eaP6&082(YMUSO zq`}->ilwabzB$liegrPX$8UEDSB}NX($j@O_o{fOE-%C@O!4{TZP_d_McXN&L1ApC zqZ)o6<_ko@r0g^-?5p+Z1X>Ns4lNQq_Nd}}uV9f`#Fg1~QxR}G1ufN0RP96c_YxX5 zie)jFuE{7<^`Be-3^qKw8E-rjP#{g{WF>4QW_%!&K$>SFa9KPhq3xZ`o~D{sg9VXc zO?Hyq&~+zzod_Aq&c)yg+Kqy)2FV+a2m#zR9L z8#!%1KHWf_Tj{K&rdJ6(dg|Xy-`6c=YMYZ>GL_uA|d?3n0nZ z7qM{Z!s&^hF^FfIvdAEgzvtBxfA*p`u3NMSlqisE0oH8DE;}T>&~_kyo32LLn5@96 z_;Pag2M4;wLUY1j!kx{fBqsBAYx7_QiMLV3C(M13+5>)^lxy!Ew?$$wikM98sekQ$ z&^t@h;gJPrYl*{vDv>eSrkl&wV}x!8PTuBM=pgZi3=0V%JbFk2;YiQ)f;COGR@FI% zZBk5!s^XGy0P1UEz88{9@8_i-X6qb$ATO{rwr+FByS?uVMpq-d;Hgl#dKez9kvDq%Q1| zSv220K>W;*xA#gzk2g-1N&qqv%bXE)fa{}P-B{cwl%!W%yaZAJeyAUD6wLDQ5Nu=M zi)_8*M*B9Anl7=qg2ezQ?R`lY8wcWsLCLBlG6qavjCB_9}h_QZSN*pI~Uexm!TsE953R0ZQ*A-w#dbH4MErdh5Y} zJH{n(BUyt@`ii*&7KL>3WQk|U8x1@6Pqu&D=d#Og{BXSXU}SqHwkmH**bUsXN+*|n zH&CoBzjzp51ASPQ2PfJ;@Ja9vuaaeMEi!mTN|4H$F~T$LYT|>ZSq!t{MbT?a`hE>M zOTm(4Hh@GQ)%MvL6)PQ3o9e1*2|SnTg_3D;=dyl@3HFfj+NW0}7wtJ`7AH^D=W}Q= zn5~Gndt5ikSFj?m#C#D)OZnJ*;()Pm_&0Cw8WsLg6$8!Jet8?w&m4(#HUIkA{h}_Z zzFxvW`op`K@h>E9B){mUDTRIscz8;eL@eR8&eF8@eoBe;2~s2o4Gxcr^A(5l@;4>2 zTgYeh%myq*`1jh}bzLqfhZKQfw-68V$(FqS*IOxe(8^Uzyb+~75MW#IHCPQSF_UbRdC^6=9~0)wm+l^bmo{iiN2PX z2vD9nCdz!PHH;i-4TtWID;4qd`p{KjI-BvyXbZVMrhb*_;r*_O6XCanGi9`*FR_XM9jcp%AG;s)zv{4rc~&AQESd42t? z!&b*j(H zOZBg_7>ovT9Fi(mw-Z_fVqsR>oV=?<{dVIH3dl%uCRW>UiYe!oMGjpn%OTF$`}G%r*!pYzbP0IS?xA`nKcQ}I}_Kvw!;AdKHojLT})`11uMUkyOqj{pj!WANAtTf7x;rRf+VdTR~ zX?&U8&J{iX#F@KP{L0(|FeYM4*TQ{N>`;(f>eyC8g`M(qr&V1bNTx6Oh(^WXSRju0 z&}rJ6QeZvRf-O?SZxOiaX=UT%wWfV&$+wzKt(YuP~AEWup5Y zKfE1itQa^(i~L>FB44G`=Wmc&Bm+t1kJadms#`(pOGHYl9(O_)(|(ckgm{umMm&C| zbcY1w^Rfp9a}8QL>(>v&)hsnteouH(!#ljnR(W$nm}sTUE=`MnA%=KpdntyNPuF}EIygH#2xHP} zj3#YKb>CHStlIu6xg|J(J20-rIiZNSRJ1iJYvtcW_xDX2IBgB5-o4 z?4mBb^01w|ee3dNHVU!mz@pF!o;QmHas!)rQ;aP{a@oE^yUT5hpGfEyTypmXqm`*E zXdmA5Mzr~430T@_R6teK_yb1$4E%Z(LMxUHs~oQ=-3!)eUvAOz5LkU}#hU1r&sv_1 z7T3LG`5^p?q!&4{Q%^Q05l0z=YUOlqX5f>18sHz7{*j7(a+CMXQ^iJ}-^g=e# zG_|5n&o9v;N+R{KgJC)}{#KLblDjHHEZfmbCS`+zq{u25}VYn{KQk1XiS*($Vrk1a(2x z6b+BoF3lz|t;Nh{jK)4a?U0Kt#xGP^`2GbW7FN944!>^cAaS>hw;7= zwqspFc@4#lVYXH4b!z*8URiFs-QfDxl-+yDMz(z9P2XjW8>tG3aCe%7coS+iLS{n9}mjWohQ&Z87gZbi6g;nE3q_45ESoT*H zaO1F z$l*6=EDh7`kRa&lEXS(fBK1&{>=m-_23YIyojWz@ZrD9rSq>&tKNY@8%)^OPpuG~N znTs^BtaIz$mCaClqaP;NhE>djZ}wMeXOPF8mg0&0iL*3L!v`t%g^hoGd~N~nwIFnW zu6ob-sV^N#M^Jv*qSvK4p6`-vHYuu{nX8KEA|FwJ_~2Q_Ls*s1y;66}{v%{pO3p)2mTELp%WLEFvy{>zK1@*@&xPmDqfB{iO}qki#PJF@lzH z$XWC*ULFaH{gVrbngPB$QYn(rvwkmH1;SJ^Ef#XYAb@NZtcm?2##Ph8F;Yw*}Um&8+d_U>_fdNUU6;#Ity>U5?z0 zgMmxXTq(W2*v(>^*72dphgEEh{heL#>*MP>y%{XJ7rDA$5Wc$n6eRGHlBTsH&2Z}M z%ZN(wcPcG4%Y9JyD%_;3si^_?OSR1Co0#~bF6RNty!7>DTD#DU+0flB8a?a2&{!K= z@l2Tht-w4kS$wj3VV%x4N4+cVDG4^n^%}m_#9qJeRaPsyDrE7;0zdqr$LuUN)OS}n z%3DTB_{4BRVQtgd6jkfV0}jv!WtsJPMsow&dgP?GM#KiFXtd`cN@$|}!NvlPKTQ;nyx>igRbp(-{%acfb3 z+sFZY^mn*jvUqo)PNu@*In}#p((aMWR84Q)cuAAEE7W3Oh8@b=^*ez{Uu5j$Z{RpM zd`If#fvu`*TwL=*%Bch4tb>(}BsMXG#r!fo|Ju99@V4XE|PjXM=30b z-XWimu8wRNGCr7YyA@EhhtE!xbjj7(8Wu&cjSJ zd`t+s<%%K+0(b7f$7FfFnwqG1r@OtZ$ZhoOart6R_LY(HeeA43!CTp#W}l4{9^qnw zxgx|1xrt~74q_8E*V0{VA;e($U6W9koiw8%pb}*7C)kx)h*r1pPdT7jdUGtWn;1Lo za9@w%Jm^jCiz&6vWNGctqM_LD=~c=O>WBfXq{_CzqNi&@WG{?@;z7F%G813g;2P3Q zEFlF-MYlbX>4?OFpr=W=MBYUsB$qsT3B-gYQ>0Hp&K9*88GuCgE2yM$>24d(JRe!^{zQ{v)iA0N7(d=EBdavwM1sroNd?6YTUALZ zBP{kdpI-&imio&;M4nOk)NEW?P%#=T_X-DiY*psR=~ub?RDvUL<~*?r?nYiMLYvR> z_qj@Km!iJiR6!)8Cdl{YM;~(hmwSGZ&xQOTr-j%VJB?9KZ!t}Y!P#f`59LW%cX1xl z7~cvEt;%*~oIs>cD3kzUN@nqxc+SmEYtu(Tsn)jqbnIH%^IS~K(!zl@!i=pMt|<$- z()3P00V9tpBffl^5wj^urXu2^wokr{EBtSHuJ!#)m7EIF7~KL(imre1l83+HsZkEd zo$0zbJEKEBfkXyhJA$(C>We#VBi(No=bM?GDZ248>6d@mcyS~2&0btQ%1$gp#ILq# zc~oZS%;f1( zDr&XlpYjiIWh%AdRh+!WV*xVIW9WDf2FNP2MzCh%E3ZLeT{2pFESl*x41~(aA2?&PoN4FxLNZ-rqOrMU=Re9zmv)M*mwq?+}R9onY zlfA8^L-R11>~pOQzV1xfaA;3(MIZsrwq`jhi!(7$#p#JrQd9-aI!RpK(`_u)KCuo2nx zAtSj(6Wv!S2stZL8&^Yv9U?U(uiJn%ekIG}jt*7Jq$pjJZM|-ONy4fklBaj)6*oew zGx>!{oxGw&Q19ih#U%>m4VOyjCSGR>Mx@@Pf*akHhZbautrYZMj20PtOdxMT?dx8A zG8ozK=^dtMFnII}TA7PG=sC){#7_X{(*jw?6unAhpWboL+~cnc@g-5>>UWXNEp!%rwB zQUuqijJ$gus0KaJ!%6Q7&MJ#zsNW-cwQa@{eK=wgoeOW1#qp<=Oz1Pnr@vFkR_m%0 zz;F%wNPm1!5oKr=k3P8;G25XaXGK1s{3n9Uw(7B+vPFEQnw!)?Lai?>bxmPc7p)Ns zEH-acXbhw+60>85d%JJ1UAIV`ov^rm{8eo7x~CKc6nYD+|9)pDziWA>NE~kGz6DyTUp}b$nY)V(g?&2 z&(@;VoTPrbJncL5F51ae9XG2c`oxWJH;T-{pGc5qDOghV3|;q$jA_}%ntm)jE?$|J z0%w&f{MI7ZcWaH?f5j6^Y`6SbFxU|LMQ_N=!gmuh+e-?^Bh$);Xw&jtAu7`}yH_r2Pll z{{B5YN9O)2B0bL1BP!_qD%YS@2lHO;rvlPrg3RH}d)^oL#{+g`GDF`u1fly%D34sL zD0?PrE?I3M49jmS`%=c$L3@k;Z3UZrdk}1=>UCCh*?zb*h2=zY`C&{ED=dO0#~LKp z!er)8OOHfB>CwDx>2S{A@VZ zg7rjC!i>T`gLzUmOZkh|z~3ezPXYBbKvowzsK@m5LIs5}JoGX9~5`|2(|* zdLsAWr{cmh6Ze^ZcZ7YjN2IG#t(yPFt*VAg_wOC_gz!m`TzgOI&+M^Q#atdKI9rik ziD2R=LBC17W-(S1??S&6CtcOHiFM^M8=`j;r>twVf6`EISEfxff(hR&U&A%*SOzQ} zYi2BWS>cD)QWuJ@+magPG+(|K?-?%W2+ook4D4_=V9T*TKZ4vO)LC~Q6s`|?o}g!B z#UHY&=#+2!)%-X~{98Ja40Y$ijdFwel@Tq0vc^&WxU3c5(yK#}XoJwPBW1(q60GGk zt#1!bVpIrmTLk#Tmew<~9a)MOi;JOhQUum`ZNK#_aDnYQ*ywaPGU_<*h~4WR3=HQT2`iZijP!^`N^;D zKT%5lil}McCvM8(NUiEO`z*qBd$yKHTv-^DRL(yrXIt-c&z1a zy~@J|uKqolZA~s>nJiIT+6EB%6;!OjKk=|+@D(;bNOLe_?{hCpwcAzbDXWO2rzQp= zC>03>)w}fDx%FFfj#>UyIVRSvjq*@PSaa!83sus6Niz+LC8C{^mAoOJezybD`NM9B zMyYpsFPY53t317S+lI30@0O079Q%tr#xAG1vBc?IY&>ABO~cHn5~0>3n*6}o+QGYi zGJx-c0XT}^s9e#1Sho8-UaisDx!~kj!_tNKM&${Von+x=V->?9JC4Mg=8Gpzk43Vi zFkg5?h7o`NQ8_gJqGCXqJ)P%YYK$!NA+W&b$q zYOL+tjAUiQTaqlr_y<*e%pV`k8mDqF5Pg>H}jkgl`0pFNvwU7jy*SsmA3S+B@nz8{Ukak=p%LCcpGthw8^<6Lw+3%);v( zdf2N6yDn1Gi`7di&Ia3fV?>HRI8Lt%h6cK=DHA;~IXvxNmVLuu`-~TKC?AS0F^D4n zjngy1B6omsyAqL607>c)FcPX4S^b&B=2o#H&(nLP4_vel8RV_V(524Pu*ef_X zskMhmMNgXi=*(52yn4+dUiXM+3S?#Xnk!SoUArgLFA?|2GD_eEm~VHwl5Ap*TD-LX z@GjF_Q#4>iw5QCcXW!$t2ikTH@!jI(8OzpD3`@Rb83qqL+c33(ipWr#>)WBtzi940 z94ngAbb(dIDOalPdnD%k(DtN%8hHf=rb)ig31#w?5uWKG9a8JKv`9gZ z+h_7<>24gfy|iSqF1`|l#0$}6#C09U1S_K0)%kv}Q?AS;&)U=i(iM|ZtDa@Gmoh|f z^^yf<>BJ7Ah|P&nC{hF6&&;#Ct*zrAtwq$w1IHUO2MASf^;;Tq#`<1D0E#$J~ht{{g zGLCKz^N@|2*-s4^UDhh{F;3Z@ZefQ$`%<6Cane|`|7AqdOAWf?9h}yfSYpuSFgMaG z9ey|%W)CuCmFz)Fz1fkGjPJvRk~_ZhhYIVK;jzEf(;=@ zlX?0X9~Sf~Oe-1y^3+P73}4uI07)j(+4sN7alLY=dB5+!Xke7tJv_Ego+6sk#PUem|OG2_J;Pkj~jK`Y!84VP;u<$ zeB3{wvXWtSMd#k|h=p;BWTj})gf|s$W-T?w0P4`dxH#N(s)rQN>iHeCpP!yH*C$SkGob zSSrGBx|5%b0piiXV%ta->R-Tlr5!6+t2v1Aj7KzEVSmlz9&Q;E6 zhC_WCXzFTR)4lwz-abZw)86kkZ?T@zN!I|wfj;(s^~@s* zI-L{rErsel)}l>K&`Ke}<@iHod&hkb);<%?Fxcy?vugNpPgG7E5_*t8M~Dy zH4HLi&oZl@Zny5}t$DX2qC05HOg8fdWW{fVkYR?84IKP38)L2HDZ^TN6Syvu*(a<*JZq zWMClc_%o`vJf9!UA=V12nj^y{Nw`kBH_8($r?Np-#onpnN2t0*-@WDB9>tbK<>x2! zQ8Xp-$|WE6omh4d0X4VbE4GeP<<;KVfPaLNn+Nt9D`opkY#W1u?~dRrV4&1gnbST6 zO@>p=yuntFjn(0VDn1L!zOu9u>6(^}jEvOQ=Usq=b)VJh-E@ttn^(%?F~_6@lhO(y)aD5M(nGkpvn+?a~m7hrUsLA9sP)irvEDC}Ho+lg$@1&ALWj$cip9zVxKroJ6%u>dIYfoe%!C z55}z~`v7Hqc}z1G}qC~mD*^Z z`Ws`#n8z%NOFfY-@#f5wDodjB7IHgIM|^@+fn1#TuU+SA>&R3}y)Lc<9h)kO&JG$j zrPh?KOVziHP@9pA++Z%~wpLWw515sVHD*vLUbzbo*&I?qzCG#Q?ZtnhjrhL$ypw$n(shyC0Jl&Z8Ipj<^Rq+G>o=m>mCe$i-5C3syxC`d8i zp^K-5#gd7os8mZ<6{&A~P(AhO-AbM7Q(1ix!HP>tbq``}nrR!cqA$}NEnqF7OMLl}IHY}|8Tsni zA&W0gVQuODoY4+}(iVH79=D9Crpk z4d;Z2{B2c^>~*fs5^w5&k3)5gxnaFfx6}K~MbyCsIC-=LynK8`g@uJhMc_hU9>AQc z9`QVv_}aT9ijTkCzqS;?>nxMfxn$RZaXO;ALNG zez_(Zl^5#9x296cD(4$%RA_Sd6k1ce?ex|YuVs}b=^{sUtYuX2Mk+8;wl}F1tSc3D z)7N5jaS89GtdGWtO(CgAEpy}M``w{Moj4qd_fp#A?L7Ob%Mf3ouL;6&*4fT~f z#j;(P-i7vKN2*Kw-o#vYI<-L2cdiNn^^13NYh1?8JBtJdEQ+$`N(g^Wk|Y4~Rg1~m za;(#yxgOK(xOF_b4Z>|L0&lP->|w_mFwxX8%}@ro+o@85xE-M=9_ zP;odsI&}e5TvtqXH>p(Qta2KBoBGt3G<5;#Gj=GqIFgk1ljJazrfO@V@|C0=@tCKY z!07SRRci>9HztyWExPC%kqP~Fl_%O&^EW`zt@eY-t0Z?T@7y- z`+DRudURY3OvXpza4ku{kQP-W#pf0E+o}CbZT`UeIe_bZA^;T=%ytjLuy*S*ww4i~ zf$$sPPn19E?1Vd%WX>#zrpMT5xO)P_3v7-5Jnn@Fh~o(`*_$9vQ=cTmnf~k!_+g`tB=3SH2b0FeY+sHy1FGh1qHUWkMVI?;G=Fhd_Ri($h;&huU+*2!~RmW+}pZjxO$`sOVi1|E!<48&~ln zA|p}nC&!!qH($uhn;zcY_kMf0)MwQG_Ve_N_h0X;H=<^qN+kt)IBhqb)ipGDUcUxc z?Q#-=5O}9-e($ zHK|iKv1N8v8`QRUTsvJ z^0firRZ&q=m0R1z-{iTU0q$^=6X;9b^0op%vZfa`Gq`y zHvxW7)jv0aOlN1b1^9$1xRMkiLUoSjUZ<7KlJ`zS_S0@Yzc(~A1PChp{QL$6=_m^Q z-t{=;nA0cF=u**ten*|Bm%M#8CDTn`w%=#jJkEWX!O6|_-~iwny?pz-wweFipJlu` z{5us&LWe30J#CMQQ+xDF6h{Fj2>^}*Kha8``)mdKYPn=bB8ngNfyn&V(Cxg~E5{FABFn)JyPoJ>$?vACvwjMRl9!L&^W=EK%%2P;yt|K2L4$cbXLqwd+)a_sn=X~HnmhVQ!)k%ztz=c zPVz?eZkB~LOP@3t!2y`Pn$tB1#5f$a6u;N#rc@%5>pJAqP2Kq(b6yCf1G^`*^BM-~ zx#C-gk^r%g0`S^GUP(zwdO8J1lu6YE$SdNf&d8USK8fz~lwOsgl>1rK~bQ>cE&P7En{KO54TR@hH$G zT?@_f0;HHMF7EN-^CZ+vA>&rUVqzz=ktP810>%BGFM`K)BxVt(CptIe* zeH#vda{!J5c<<-G@A)@r0&t}eQbXoQK=|96SMea2t}p+HrfM2W3BP>%jpq_A2$^#)FC33fPA+CJ?lPucH zz;H-QOQY87hx^$Wt^v4`vcwhud6f*%OTc?%opBd%@y=#>H}UOr?bZkqQ7Ktjz_g?! z7yqU#Df0Ur73g*_n4knu&3{c$|DYFENuo4pSoB-4NGZUfg7Zrj0eVi7N#`UK=jC+) zJ#45oV*It-GV<~f7J9kwQLKOs6cW#wQVz=#Aq}+H?~zrpw6X$HMEU@9{@33Dni>i^ z$Ly9UD4!0*L9zEaP8t9x=a&<}9z7|^KQA_W@o500833Ep^;_lKl9I&WHPrEnzvwUtaEVSkbpEHqMZhy{ z(d>OaJt)Q$0mC>?z5@%kX!*=Ld$H(W!1xWQ%Q2&0O>RqozM7+dO0)_Z-Jb3|2U;sfm&q_*RBBTo&o-~y=0gqlh zd3otCQM$i5+j<=f%LoA67#ad=FLDFu*(*F}oT&8X&6{75EaDN*si=SlGmHYBKQC^0 za^*I;Yv4AVQ>iM>1d~3$tbCJ`lK_4yTd@S--#h_FOEdqargC0hp0i1}I{+>MvqSKK z*_~VXE@AMVqGD9b%)!vu2`INOh60!8Jj)kMnZVcf=fLfdW?=fJk^f8|0R~S)PoI~W zi2#s!D1iz*vw778cW&RFG=ZCLObddK8@>LAwC?~zzmC%dKEtV|9KP`{`S>R$CIGzm zT}C}X8ZJP(^OP5*k}r79$XJk^41mc1pyKJ+&eK+u_stuJ!=bSdQh>M!bW-umKMZvM zaNh!`+W||#;V9x2{zKi$ZvAaoD(|rfi1ApR$VY_0)MNlJUTEy23~df zr)zm(0f8`p0tWDL0RSYBSePGo2jnl(j|7kiC$@B@h?k)pnd;2fCVw-Yl9|~CuoeRY z19FOrz}Oc{|NRY!`aqr*deG^fMm&+TIc~Q z#mmbJg`z;7s*QO6$K4;(f!3jCU`TRL;P*voC0qi$k8wa3|JTMNmX{4n6i)y^l&p*l z>H)wz=Y_AB^vM(T1$Ywtjhfm7ASIyV9bge4w7CB>5fH^^3@A*-a4IS)FbFV2jDL8% z*hlKUe6G3+<2xxSi}guf)HGH9R;|VFUcat^RUIQ%R#pJGC`z?o`CGxhgrH7gzK;8z znF&@k1E}YV!1F&(iHqPc+t~4n_330wM~qRQ0ZQhaje~qjN=gD)f9Vw!BZkc2BlY-y z(?j(wg-nLuq6k36{H-p$DE>AuwBL)1ivcL6-egA{4~ijSH$VMhNQi5m#ezhz&^ACT zjN-7B0J_Q%#V}NgUdr|cJi7DmwNXYJG}6J~7>bLK`d~$Bm0b0@Sme}V1O(cqjX}5mA0yGys@CtyF2B0lbP>D`QQ$%hqF5q!vJk%t-sXK0_`c%P3 zNPfTYv5Ct0MgpXMC{L?p*{ZB-;b^N%d~;(13}Ff!DWnbSJg%bFPDVN6r;Ya5+06|F z3FtHzZ7l}Cw=zwj6m)bztZ4~REQ^5(?@pU48nBavV5)lR00J^PIvPcXJ7-oLLR_fX z6ab=xLKuvQ%|ZYuhC3h;dvxb-&HK?mKwk_6rCVE516&e-AqS3gN~?UkP^}8S_g|yE zMn^|yYikSt*+AqpJUB=WPTJi8OOkW$5jH@SrluxBUmbu{1|$t|U?(tCxF#2Ifmisa zCqAR?#UAqDMHFnR#k?QjpA66k&zTjf{tmPi3V9oVP|Bol1CTh-(L0c`qoq9Kb+5V% zxqNU+_^~JoMboMbfb~FB70(1p)ys=-3Lx-Oi7tE&(F`p!vmn0fpCGZnn)KND@px2rUwiZ{r_$(!YN%;3PmCjd+o?0@yJ$SXK=;9 zQAN~e-OnDIQk|#)D7^0?`T+*({^lIeX*{O{7$yAElb7#33YlTsxoC+`i|tM)=8}?u zfC8qV1`QQr)1sdvm2XZ?j?LX>iMh^cxTbm&MUT3nKaS_e55%D@ec38R zckXm`cIKMTwY7xO0NK0bLm!?I7pK-9!+8<_xc7K}_Xa8ktA2DwLfR96sJoAiT*C$_3MM8d z0741e2u0t=&;IK0^)CR;3lsJZD z9&E-Uhet-Jz!rEIcgIZ}98w78??cO{-NL?h>MG`Ii-Jx@jK?bMS;DVC!rD;o2+r9Q zjH5w>L3A{JRboG_sA%VScgz}fn3t9c1j*+mWmp2(1X}>V3`&_77Z*pNyCcBk1e-FB zryZUNrg0&F#d9NaeDUPTBtR8kE#DO5;rRme9Rtb}^E~5D*e9_R`)PlEkOrbrsXQ9B zUmZYzGge_@|90pb78b@&Xsg+qA*ZCIRA#*jU}C*c3iPzYmwy-!0`rzgJ!)_|Y5>@p zF)=aN+(&DB>r-GZoFJD+OL@lh!(zYdy_EXa#i=8l3M!vB8-F7-tSl zliLN)Syvq@=^X&Kv>$*h4E+N>)#k5`>k(hLDs@W~-qb(ib+(IBde z;2Z-EiucbtR?Q0oNTRja+?1g$G|J_BZS1T4r_+xIfo8{B@}lZ*=eWat-lEj$AjJvf zp)AV)7~m>^C=3E5;QdDzHIa~#&hm2~be7w00v*k@KOK1K>Ww{5WiYd8pl<+o&^|qr zX%RL+h;-R(fT{?b>O%QUr-QCoYL!ah97&=BdtEz;xbPo`RzF4YG11AI8w#TmKzE~9 z+G)M+KA^AkDSgj%tIQT{JXP-vpGgu9y+Oov1Tckrrqm3Z2m;)F{t5bGCI~9;~l05)YQL{5MnaUONq=8DniU0O(9C%V7_C?D_1%N=F z&gs&d4`*8=Y1_VO{mBKuK1|gE8)j*3?R2!>;I!YOg6nt@@;S7Mhe!AMJq6Xb<(z+H z5gkvz(5}`MYiz&GI#Cb6ze5kLTw{Uo&tNkGqk0|k1*n1l+DlrHDFseS)#3OT>``z= zJ~K0OweoOfeX8LY+5)mCiQqZ}6nV~39W53}h&Mh9DJiLPiP=nkJC{AM(NSV4U@F)k zTC7D~O!B+Ed@GSEg<|B2-BF5rwRrgWV~&T2c5Y-jNalis!3E&ny*~>((bYY1s1DYB zoj447DN|{87E)p>ftv+9Ij~9#|41|%>;Z!|$3`PS&t?VM@`fAmSqOkb1s*Lyh3O!q z{WFZu$^rrc0F;#bxQn~4zTS=_hM3={9C>nRWMl*$m;2}Ow z$D&G#IyfBFd*g>vE8*eb^mcTt0Rei*Y1fTnv+Bejes})zDJ~1%%d=l8(E&^eawb63 zf$y~sb2?->wVi=4eSe$IGscwc1=z}td$mMaiX|HW5nb27;IKvIWaGzGBE?b*LExoP zIQ2aY7XcdgEFB7wjB3(6XcBkK1DB5I9iTMy6cO^hlvMYrl7_iEmAkL4nVueF`a{~V zgC|d(01srnmgvM3YNb*}O3I9+ zq!}8(&A?JQP1IZlX|+P|%HJU;B-tOBZq)RRRHUbah_u;!K)@Oa1?HY>Q>;e22bmJE_1EXKn&W56%~ zF$kqmIhjxJ9d$f3a;pWQ;vJCptrj@TiK!ldq;<*RN*S^Nxkm)J_8JK44Gk%uKS%j0 zb)NSxLsV-}T!Qh0`wyu^tgpFD0y9woUIRlv0Lhu-KBr?@5T(e=55kj#rOAVV&&dKC zLhXUno+a-fD46s0>Py(&y}g8-b_IajGctau=Td@rmx4l=I`lfEdJpi}c-K8#JfQae zYab23oRAB#Cwjou1b!oj`0q*Kno2Ka*<+n;xeQJS}MJJTdMI2eTm zoRBaDVBP`PHSlP|xt$zQ1|2et%D?k{KFx%mzWC@P;D0YKFMcAZ1&Gchnh;<%SFs5l z4iPX43JO5-S(FaB4Yd8A_9dW6b9dOh9hfHQxjq06=eGqUgYV zB!9hC`<6TyxZm9SP2ygB30nqAN+T;HW!JA=JIz3`bE zlbM~31bXs!ya8u5p9QjtzULMNBE5^%;~=>D)9UAe>ktdw3GdP0MhazIfbh+aA&2x9vw8a1R7C=mb5-PB+!0>&rwFT)TCr4Bm z4~L@`9{AX;ZEZfDo_8R$-ut!6w{G3~E3o*8@i?#vqz&#8L7ghb}?&mc{;t)8#@2H>18rNlZO0ImraN#^y$X+_mcRBw%AGolzd!w+?u^Uq?N zUKgw@ui6LT#QWKTVq(g`8Hi#vM}vr1!c)dToEQ}obF%A%+yyLhrlLL$O#y!fxvhe< zJlI;ut~wcXY6U8ro}L~MMH7HpejMKg8Mt1Am7B9oOCCIF4eMuD)f2%K_rbAf{}0Q^a(py1%`qz4C^Dho7yJEG36t^p>H zgEM6Y39PC9@Vr3Y2y7;I>SAJJ--q0E2HS{0lM&K(m|j8}#n`W~UifzZ%%akkM6p|E z0WAc=MyDjTOCP@#0r>#wGyh>Qg$?&-F*hNg_=?oG5KEd0u&4b(LVNTz)4R{0#& zA0WjIxE`#NIw~1v3$)NjQM@BrS2s74!2-cdnd^|_b2NITp<3$-)Lhw0=@M9&A`3Fn z>mbK1ub?pIu-W?238eQ!sFm#6Id{<^eP{L#@>yI#Vd%b!gIex;u*oxCH4Ch97`M~O z=gZe8r=~7LMChGPHe=J%(u{TC<4EaD8I$4Sccg(Be|ORkAk&R<-)G3A4fa19|PZ_;_`YX<$)Yz-ByJbK{QUv@ZfN=()*nm6nwysZW4hHwZuw*-!zd*A>bp zjCfT*&o|AOU4X!PKrZcMM&)E?W2S{gksDmky9#nI7hgU)C;VU^rxQ#9MaAS~QX~@) zbTAA$Ja2dsBNb$vnbm6f{f7^^PJ?5Sk2yGwo^kIBN=XHQ`&rILLVKBskrBwoEJzIl zU9h17LTyP&NpC(4Gcz+J-SUO;UR^ysl%YQAI(bP@w_XERg3w?W94*aRv(wWJN_=!Hj_G1~#vORTrqdc0Tt)yDd0Vo0dkN-lLyC zz(MGXZ2k`JvEyRSAALghBML<`?mzi3`@Z*`YhBk`*Gg4*)^&O9bKRU2O!=3bEnoxe zP!R$-vYHMxI#|TC-e!GF*Xd!;yM9~OBT^a~zyaMp^kB zdMtNfdefnNhQwzhJ)5_;8?l4p!TP8ux#{1Vz)Fr^esZzNCO-Jv>1vk>^cy(zq0p*pnK&8KJ>cwn0EAs6 zJ~l^UVf`f^_D6HK(5?brf>dYG_^*?;itBb{5j4&c9)x-zJ?Lu&6NrVgLXT*|Nat-Sip+0s$5c7uV6vwJnqLuzLAyinFax1`js@ zU=4*;GNk_zLHp17^DzpT85zLV_XqZ(_WTjB|AFSY$nos$4{@rUy%9(ASBpLg!$8Z} z#9Q(=7VH0>&>IvR7^az-8K|j~)8XUv?Ck9C-`PLm+z7srtAFRQ_}7`jvQq($1Asoz zP*U0lFzgdufZ08CZ)i411U^+c{}DMvc>WAReYvab|C%xS7!QlR`gj~S_1$##Z(!Sx z)3vRwqQm@x0*AU;FBJWY7pK$QeK0Or1A&{dk$Ckj`H!&y1F*{*pR0P|8Ni6}Z`=f| z$R@C_UI_tkj*}OHy)Qdk)6;SVZ)1gu?Q#F!TlP=Dfn-oV$;iZnOc$QlNWj6xm6emL z?*-q)#ALm@`M|iwm-yFb!!bT5sRowA|C0EC%?$*$0@S9OscDWAKuJtO8?@jG($Y7S zd31EN$?G0;Yt~Jp@lDdAXBSJr1(2K@LmLZ>S(T*3moErhF;q2_JaT7pZ&5&u^ zkm-ye^E=f_b}`}oq*?ND&+`(CN%^Otnv}!HyusRu=>5D&+b+06aymzZ0dD@ zNrUgY=X53|b^Z?vP)vVo00ry~_6Ly@R3;$U9&j90hi+~64 z99*28k95-`BYy(pJF*U(jct$q86>jivEP^9Tf)L)VC`92T5i+)Bmn3nZ6tDso)WC^*m<9wo5j&{z?WK zSl%-J-Vjn`;ZP&V9kN&bm~47lOUu%FRrX2oyQeJ74Cplcq#=hPJuO!w{T!w}FBi~b z0UhsL)5vqe9!&3RjSZ43{m0{FUgBXkzO(A2sg>4-;jEr_s#MLGwNvuy7TRZ)cg&$6 z7v5w_8F-k2;`~uC=1qtOWwxoE+M~EFK#}iV*hRcAOY2)!*@bRVE+D2bU$=D1MENY2 z%If5)AQ<=X*+Ejb)d9Zy~hA$(tNNC^v!+rnRYks~)d-y$K7oli6+yo!SZ` z2_NTEgC>0WRc4#!>Y5reI$=hqONVY@CI#ALRab1zfpl#X6p+ECkYpom%lu>UL5>2Ca~zqC(=e}oh#-26KxsoS}sSqO~=Y8Bar(l$%yuMki|XN zlmFvNfYYOyo}VAkxZDJ+AD8r-$cE-sfXoCgbLgB~o?eNff75=zu>&@vymk7uGN02jnT6hk9V)M9Am0TX}sYTd#0c*YG3aWKcDhG<>vh z(?G_OxbBiBq=LNewmRDXqqtKIx9NyzNV%GZkB$2HQ%1{2zF*{E=cl4c`HjC2Qn=-s z!i-Nr^qH{qz)f=Y8dR)v||uUn6RX| zkB9l|P9c$!epEv!pz5El0H9WYZn z!i}?@a}=*&c9*yCe%dl{rO&XET-}C0uW=bzOy1Fk)Z3c*%j9U}tQ|4*qIVetFS>jA zkz;tSIyJ9t-3D&$`flj@OP=4QY^k(Z-$ajU_;7xKB?z%RsU9-D@8vQzb$D;V)-L6H zEW%HGL!MxTjB7a}J9Utd3TBc5+VHCHJfx372g$eiTFUuYvTDfk{9>U{Vsc zv$b_dPX!pe+Q16;>@zDD$&`|jIe^+G0`O9ziJs4SE8v3Eybf-DdCReCI@#T7NhXK8 zgQq|$)ONYm3QZbk?KY+(z@9;9M@m$z@3hNgKcua9AJqFr%*2HL=)eD&%l+}bioL=2_oV~rUE?Tpo2e=?&Tcf7n?1@iq$2Ejs&Q)Ub| zie;V7r7A?kM5Z893l=V}+m;hD@(RV=GB9%pG~e^cVTfzZE5Fh$iE`Jk?lM}3giclS zO%M)UykYR*PVHw6Xaj0Db9w4bGWRlsUB=%=ZAbKt*=D2YgY}Xp?dQUiU{8YvA)RE8 z-Q4I74H?^ShfL+_t3@b}yejp??DUkhL4;ac}x=Gyg{BgJ`;E2nBs7zTx0Ba zdg|RM>=*8*E`X61hfUW8+TG8R0{D7Ew1rt3j>TFZ4wKq?kCJJ=r?b_vd3Dv0!`unb^9;o$&P z2=KH`*P|+mmjL2=2D#pDqNAq|3=Gt-by^2F4qm&TKW-eDDgbScQ53o{9Rn1Or(9%dN@46)xW}oK;kC7n#3qp zVSoWlT4ko9Pz=cX@YohRR5A8J&+TZBXx=hHzUc(~#@>O0F_lC;?<9wu2O=8I zt!!`$;3Fi&+ygPL+uOH|%6>EK+2H`SXtn>Voj~aHCxHH44vG>0qP&11Duv?V=1${t z-T;{534{%BZ|h%g0vr|K1pvUfQPG73_WA)3i9lOJeU1wA5N0H;0F!eiz%Nqzp|2Z_ zaurG*OH)jnGL@$6k2U&h#biA7a?i*)`=SODo{8m#zI`eZRonLSiO8>CZtL8wPR}C6 z!AmbX+;>ZNx`H+Ug`WgJp78IlRXBTUt0DQY0kZgWu)YjARPwq!DQ?heL55HxA9{xvpJ?B`Z{27{&Jlo* z34#$JH!4zDiqZ1Jck8Fe`$y35Jrmf@Sz!zWWh!=a>gYG`JP~P8x-zwpMGZn})00xbs?x;8 zJ&x^M#qOJvMv#-wcCFEF0i}4KhMn_Umv;qdG>&(178qx*yKVJz=K1%;U`=kmKHqV6 z`x(Q%mEQ}+M?LJw9?^3+h~yC@wC&MM$$El%=-rFaJxB6+#+=EibabKrN^|j$&4b7g zxjBz9$iKm!|M0j)Pua``kE2Wi>gA@HDNK;bS;QXh&|5jWGo`y)=|n{=ogSi9X&gBxlup2;jq0#6ti7e%Or@C1HmG<}dpRR&j-_Q+vKi`G44-au{ z43~gVK-3BtHKyf=#pN7!ezlkl$_!1TVb^*s9xL4WbKN1r?gwEGqxzd3&EnOHRCMY{ zNq5iq@$Q`rj=uhb2Q&-JtPU0Ujz~jki(EWN8X(>!rWPhyooso@Li)(P)uBNo6!u&l zzK2}dsTPtQvTrn4wUBN4L=UnEp+3RqUMNejp7xQ&PF@>Pt=rrai`#3MZWrE>2;~;# z`Q`iyHzF3}h&OBVbiP2Pp#0Eso{J@iE*y~w-Zcj$oNHuGnT&=WP&khrz7$YS2!poV zzAbm~$EdK!0tm$o+*+s(cLyRpZrBf>le&3%X=xyg7VvCILzaIw`!-!)?gF5l-iHq# zG&R@v@wCcz=A+-iii?ZA0S*|*Kp@=f@Nw;&Zh4oPE!ax#N0&o^$2e~iQr?*5ai{zh zp6m*xr_Ftx@TuVVsJ_>4L$U@nTE}zy>&9_@@cW4ctHZu)kWBFfP(VZ|ngn>}v>hPO zM_SrmJUibXshv~enL9X`MR>N|5ytej2`r9L*0*oeXJXHr31aF5oi*;JO+=y(1DtlN z$#()~|E8CPbQ=&~0#$la1x!v>un*saprbn~{pJ^#E6*i~mDVeRi4gN&Mv;K#A}**} zn~}1CNxFFSMgQUDwvn-gF3Jx2(_tdzij|cTF2UKuQ(D(v3T^7=uVXY-r8PCcd+8Pv z@P8Q)0IedIf&|#@RDf!|$p&zt2Ar}jM*w;SN+-b8lT%WPP@B$tsj7@<cIhq%>%=80lc&IK~WnrF9 zo?oKnW6D!bT&fEdsa&X>5){Vq;h4WX?LFtc}s;%`$o5P$&^%}eiDfd8rew%$m zidAe{*M2lQb+1N}YYBBpsJ)a8o@^QrJIoTZU0nn2O2)D+%E*(p6Q`zASFOG&<=Gyd~dVp#6y=+~i4Lq@2OpAx+Y)!@K0| zCjY9XV_{{HQc}X!&`_TfC_oOTTLa9vA}&`0 zDdY^lIssyLbEcfGBP5d{8jsR^N`~n=f(oJ4VN%W+H1uXKg{lqR=EjbQ0kw)4!_oFL z&U?1{KiW@%BMV2Pu9yUM6C@tGSwJyWu_F25YWJtjll7cHTY{e4yZL`>sr@>;FhPg+ zoN=sJvkL$L>j3yxe|-GcZ;(Qw^8W1<>gH){>-lkW7T;r@uYPtap>WMZZ{d>>(d}LC zElXo;#RLy##caS@JY+3g0C(K8ITp0SMK*DG;}GD1EdD^mepD;vtG=YRsI94)1eRM- z5niGhVIu+`Fj0VTJRqbBNL1?S?#3keV+9-g=CqYntz8%_@5zqG zOg})mofCFFe|R_*3KA{WqceJ#eu`n-A8t&x!W-0u2x_=<(DA$k3`1AP8bdeb>MBVQ zr6apa%=bHs_002ogvSSt@tepMNMXK^vSUOy;Zr~#BOI~S@FR;r2bZ{O%rd4qGA8v~ zT;Qf7+K09_ZI;7?_s57v!R|wz)>(_HsXiVK-p}P1Chw9KY7W0l^tD^zk3l|k=J`>N zO($l;X@5R(;oIOlJ8D?7w6cJ|;`bP|<{xp`uBo1$p04-(axYU=jM1N*6+@F?yOM0@ zHpLdklC)sKJmX^v=g?(~&@Y|(esU7;{we5})i9wKX1Kk3{C6A(>!nzcqM6xBFI$s9 z^fVa>NorvcX0sEpLj3>h?l^+VtEo1FW!=FHWq=4`8bEP*qorhu80|-L+rfJ5^KuEd zpekyam4eWqo>+?>75ZG6quq`v(Q%iP!Se2}oDN_~x4hc6zM2RKNauQ$J*=o#dop%2 zo-JrzgV^A47*)$1e@VrrQt-e?(}zlRvSMVziG@|`cymWVqN)H{-(BX-5zd^4S1>h4 zf~g+KO4Hjh;`HmDw3e>ExSzPiQkUD|=?|KQ__dxqX<5PtiP7h`wz2mMw}7<5ga-%J zQG6>(0x9Fsx~7f|I1%C@D5?U)LGiCcN*j-sT6Qk*DkFbC2zCUwo%!xFzB(HLok&$KFb z`m*(C9d~>sNB_F~na1vQp;SL8r=VS`I+=0a6Ou;Ax}E6fX8zCIjkWloffvm8A=6LA`wNEzUW zNL{FT@0wBHQa1=ADa6Cpw6kt|whS45CSQ+O{Fn&(4cq?7y}7;)1ccW()$|@$p6J*D zUu6j-yI;kK&%Ji9X7I5OKhEvxP$+EISxwy=D^Fc{-f3rp{V10XE!TLzX45F7VSr+< zb7-dy&g$m>jzJ<9K3q1t<*^3{XdFOTzuEV?0!8z*FX%4n$Iyp7R9jKwHfNstzStUJ z@Cu8qRzoSahrRg{^N}P(6tc?E0XpoMzNI*9eeCvaUNh%8IY=nCw&uGa;e}FHFRiwI zdI5I7FI|a|k?(zCusPlUJna8w1?%OY!J_=(;h|Il5K!rbx`4N!UVJ>ptMaW|nekZZ z+wu(i!eQMt0>d4Ma7)A*8gk9fHf&(kgE>vwNNyDy+Lna(_J*FYh;37`d3~>YX*Z-B z<93~lWFEZltnOkYf~h$pN3(lC)KZ+qO!)jR2>w21s_&dK{S}}Kpe)D zI{XQyo)~SlM3*^o&r|Qp#ygP$r3h_}*okOrtz>$FHRxA5?)lp<1jtQ?h{Hd{+%H=z z8Vlj{WC&hy2*JEW)#O#4+6B>ZqF!jME++r9cFP8)% zLYz88y2k9L>-EG5Mc|OTY~K7ft1?s##HALwp+phX3oa=h$ocGdab+}Y98xfl6y3#0 z&YTTUA1*%S8K4}W*03_<BwYj$aQJsIM7pT6vPURH?HUK|Fo`y8^03y1Gyl9hbxNy{4w0R*_B@mD&C z)YnRIEtllY-;0(XM5!^|y)Mt%^9|!KkmQb+lw4nk5ue0N78pX_jQZ-8O&@Oy>%AZ4 zJ^o9~_9svdX#Z2)qKEc@vyy}SPOFt8b-cPg*fH=+=z2nfIb`B!ypcyn3sOceMh)$+ z*r;>Cm?vB#t?!&uOb*uGkm?KajrNMnP5Gq^rl&! zvgJl-z7ghCzHMzW!j1{zuIsl(j&Az5u(N6Cdw1})lwaR`2FdtN+G$2dzazi(?h~7? z4^K+S`*n(sUBI#)Ub>?G+n)9{@QMyDy{OMt0`Qg3x>T~m)brn6Yv!u@y-^VU_bd+mH_@D+vM8qWuWMTw<`39VDgJqm!>8WMG z(}8dQFER_umTHO~}DPj84V>B17} zR>bysyUp*$dDUMx)J$~U_8)8UC-*;Za{vO(g1e?C|1b+|xqxoGeP2y`;PMBX_)&ZN zs7g5drWf9Ti|`x$iZ!DQ_~duPQy$**!@!Q}Piu|kh5@_xysGri-|*f1bB1HxA0*Qc z_zmrPkFV+vk_&6*`OQB+bJqUzpIL7Bt$WHsQn4x z&bL2U@LJKIs?S^X* z3z@Ox;J3BOeVZ%QU$Yt{&NyUS$W~IkUfphgKQtAfV^?!Rn$8#X6PP^}{(SM<3UK19 zjmZdMcZfoYomNEC`}dkY+UKNV1De&#QBDnIBW2fnOe?sy9Gcx<9iR5h$L2l^L)bu) zWVj2(JmAU_~8wkATpL*DGE}wV- z^3Tk}nC0X=Y;IusyxBSPCMn-Lp`QUClTizHI!k@c`10XD1%CeXR(2&IxIoTm_DEdP zTkFL~4uG2K7!z8Mr~EDW?qkr+yeCH6%Kv$}h{&Oj@eY!vQb4}l89-^C<+KNqsUt@4 z*9qw*yX3CDOI#T2hN50Q2X}q9LZFgrZ!ASG>FRpRtg~8Oxcaz9J<5bdJH>b~F?Oy= z(n3C&yQ9j@9Wwu+ZgZ`Ma$PCbo=H7Vw#uSBjwF$4Iu?jrAuH+{-7s!p<>H>QuIvo9 z25Z$qTntC2xG*ncxH+27SNHrX+LPn9y&Nking2GEOeW8l+TKc$1E@efYFbt8zApP$+v?8 zl-CeoM3QHP9okAj-{G)sECerCI1F9x1@bWKaYrmvyQr%Lx7Nw zG}#k6M=lS2f`6(7h0~L85Xg{lb|E$(Z1o9Uq(}sUsw@sceKhT?+Y&zI>-=cWx z*TdvVzMt)$vz)M~xTI*qFS4-wp^h(zkSG1}!`Nj>ISJ|DpKucCj-{fCHdUJnA&G_K zMgdD%#mT<7C$=07GoSwi)M#=c9@t3^S7jdQKIUn9-rIaJO3X;@i}L?eqd7mz_$a_? zZ{uC7_a!*oF%aAAo}Z<+?U|NW^*3U(ayHRtsuVRzR!0p!5koAh;}4}6oiL}W zp3@Z5^%_Qd>Dl*N9DPmi-EgEK)tl%yfG$x?|3nDpwj-^^#4AMoX-weXE*-jz*&st`ncOki7X`$#{3n zkg~P@u9sssTw~S2!%t63X&9GLMafW5XJS-v4`UvROReAVO#Pn6!sK>ZA+7+-hT|4c z(tzrKXLbDy@_%J~9wL%U2yFbULUWJY_UE=g>+J+ zsxW}28df=HtJ~#bZZM>-@;;c5&9k}G`vH*h$xnOd4Yqu&)-49K>-h41P2{mX`e1Dw zR)2>UKv@P5|DK3gzhN!pJ|oqFzJRjtKjk-E^Z2*NH-4#9hLhWj(J9o44;cq*|b}_!{ba4J(s+EuR#e z7J+=BhN`tk!ulf@EomnbY8H)Yby@pelBN%AjOLz)9mcXv+L&nrKOcz7q40aTJvRpu zO=x`yuVAeEg8X38#g#3vpWQP|^u36JH2(m;;*!^{4lds-Ot&%L&MmY>8y8xop~}a&T5$sJMS5cDIC8*w>J9uO}q73GgZoNie+-VDDLA}D`|Wx zG#}t_5`uAd4Yug|4Z>RUak<$j}YnhA5R%s!E=hW z3l%{Wzg9NK?k_tN;bZ!0l6^`obJ&DNt@*WGW!zz`HOj_XMp_o{b5>TGNb_b6Hdb>+ z7)AzB*_msHHCs65zlZ8zEkdS`Z;zAXP$J8C|wNYwbMBwC>uTn$Uf*jezMHvyNdmYFIA71ps}&4#In&N z53|wI<(&Lfn#p?|U7D^w_iLhRM9KW3*dj`Tt#+dqP!u2G*eD8cKMn{S_Ob|wr-gX$ zeXW7MAxOmUJ-Bv@Y0kdSs}1eXA8F$EJbr&I5C}?Or(a&WMjN9m{$ae@`3q97s1iJ{ zWH6yz!ioc5f(Xf;1iC!M8THO6?Ds|()L{&Uj(p#FzIeXMJag99Fv$TU*&HU&>IX|8 zIzSzCNH{u|Lg%;)5SMna$x2q-Ws;jd=<}+v--x4fkljWFUq(3&KW&;%FsuG>RHq?4 zX^yX;P-zdZ13~Vrbgkv{xIa( zi!C=O37)bLjeHmMW#(JGa?77<@dPV1FKkLW+QoT0xlRYU?3&Ef-_-7`Hmzkm_VK)Y zi!xEl7({&{9dDL{4~`gG28_aSlL_BXx{lMM1= z6f33{t;N^IA3hH)^(Hu_z8QkBgv}!&tr~{z>fgy;iu1MzX!MviJ2u06)NvRW@K(E7 z@4%=D*j4qZ`(+SXX-u9vo(Jp1}Ke4rWpdh2wtB~KCIV|k@oJAwYogi?p=!5)ksfw zu)5!~MaSu}^l(mN_;qyjPDlXg3r6DKbQ!FbNkJ;)vkIAZRfTM_T@dMUw9Fz*zC<#3$WLom6W-E0>qi`mWijOA|LW3lD0q zI(f9=;opH#Hpt*lN7^7^dg$g<{K~opCYH^zZ-KN)SPo*j=S99Kn5_8vD z@-4rUNn@EN_j-yq8Qvl#SdZ@I42zN35?ViO51)h(zLr0LX-In=yans&gBsBiraBr= zYcxNNJI18%O@Nb!8l(nF;lW!t!}?If90@Hp?OK5g4}-6hhu&Fs9s383We0idj}!+8 zCyBJeN9D`vI5nG!+;jE>MDPX>iusyn#^u1e9P02)=KJDIdzT=$rVKOssEaND z$a7*-%M*Ace7;TH?rGc$hy(oh>C5xrNqff{MHxCS#ZjwxR@O~sEQ!;USChZ`mt%kw zXA-Ac@$?okM|$lGXwf#yg-yTwr?DhH!wT}S0<-I1a%QjN2A;BTLS06k+h2{JPirN+ z%fKmA8gz}B${@^KtK?2ed|`V0I#_B}BXI>~T3IN%5DHoD;tHErNVZ61Qf3VDjSjDd z1?{_s4DKog7d=mUF%q)2;RgfFE{c$?WqfRb#&?D?(wyB@Zz~99-2G|=H3T_&FrwQU z*)FAnbqcRiNu33mdN`G>3E&9bwa2siqUi}uLW-ITI~{D6FNIanc5KRh+{W2;IjC9E z(3+Norx57dM{CNVGBYBpr^dIQGO%NG7v6TUH5Di*6-G-!X1ZpHSi&^4cd9>q88?WS zBTq+dhtz2HMOHEWBom>nc!P~uWObdE)`_}C>=GP`i&FMH)uCGF{yl#>poSo-d(YI*XVH4iV?iL%4%V_o&X(J*#J& zlD~q3y#pdL`Ivl=Dsgr9ovZjIxKdvRnKZnk3Tl*MaJvY?jhH|F5GO&Ll&6&-baBra zV|4tsYuOX>;RBg#YqXulW)il(TMbQFQ%7eyk3=sT_L-|NYua#BtqKv5K6f#L8`r^` z@+-1ov$CPi)}u%(CrXjmJ1(H*Hivg!%U7RatM0R+wb}Iqfxnun7uy2Tl?~m0<~4c# zL^krS1%Px;&b^PPg|8Nx$=WY6Z`Aqz`=D3!!oh?x{?1E<7e?&IM7Mnx z7S1+}#@7Q{pecPK-09ycr&2;kgBUItZ7S-$tcI|2VR>s*4lGt*vd5dWcdx|P-X9juS;Aksv}+bpu!cN1RMahfsu&l|UG+6hD{rV%2)$+^ zl4;Sw*=P1zbgo0u=jiT9+usyx@DQ<+z`NL4`ei6{=N?aNqdra*e#}=&4{?j^&q8bN z*R!{Gb)QfNMOpVUl5&EXnWsqC&DIL9*E*Xn4$=ETL}wzsPdU7N7t-EKP5NduZNTRn zbY}TePr4%BSG*`eDGXs{3}^T6K&=9H4{F)?r-Pv%{3^}kW(T(pak6YGTfbg(jqLgN z@*GdN)su-{z0G!`n>mZom{YqYK-#l5D?QHC1aF>(ii8T~w4nE6?xv>C+lK^uAaD7D z2LDaX`-riR=;Zv;<$^W)mr@3?d(6SbY1Nk6@GB+?&;eS`y5ZAoEw9dp0?{k_i@Ul zNuOOJoJk!6U? zpQ|FF7lGneF2uj_lNE`X4TRZclUle7~x));;!9&7dt(rm=OXc@8-{l;`9oZg}>2_cSW; zD-Hg^PjJ&^Rw6y<-y9Kxr?2+$sAqve3pVJi39*qW)v%dVN)q0mz6hAHiDPq$x~QDbNx2 zKP(_@JXTexg2sicD?^x^p`*}Ay>*HGoW|!c`DEkp;iliy;H8wWl^)}3ZE}8fT1b!> zA0_j#U$@2*%$9&PGueR-o;2?<7>ZD6-hwXzsD58?7>e=!Y7shQ7^*%MO zsW(axU4MT8Vj%<-*IeooIqz`rkTy}lIce|R!O_JnuOn00t8KDFFx+;WY^}SW=UFzU z1^2eye$X?UAGQ>qXzjt6*T~;0B2Ol4rL4|dY27>Jtit{F#U?(?qh|Kq)wwXkdZW3N zZz91T<4uSp4Ea8GS{p?6mTb23DscMTSg}mDDPDZL>bj|H$qUp;c@F#`MbF!!UsJAac>ss%USr6 zrWBE~qK+wY6RS4Ch<8QyThomj-snEdT5srO(3!Ve27~8mTxy~1*kWZaejOupQrV~; zoPio6U&k4OPn5432D(-2l?-S#I-iVdRDH<(HXkUkGbglMy#k^eF zZzG@(3;3yJ{?+JYe&~HhrjrN@-}lKHJ02TR(G2d47sF#P^9oTz%%-AuYZQ+s?xEZ9 zrLdNM8jYFp4{gU%HBDCConn?Ky_Anm?sDh7Yn>@$4T+yij@r%>;_5O6=Js0#M3d2u z9kW>`i!$d|8dv*$3;iU}P_=4bgW1UtwVqjNhHo=xX9IDIlFOy=wbFjdw+inH=je9V2;qTFFp?Ym%EBdyB)KD_S(ka3ql7{NLwe z4NcktsXltBs<4PHHet=lbbcDsx&2o@)1&yp@1&w2vu8{B!VWXDLhQ$a&9x_N>Llv% z$$lMYKhSJ~d!Emp{eAc7ET+5kH_0K9aKfmxIWO3kPvTM}-k;j-)Ov^h%_=lb#B{ zr@UbAi^q({Aq9;OUeS~o4D<6JKXqPH{K2t&k!^Sq&>zY5U^0_h{ ztC4^}PZ7i$7FS~nF`XOBJr{}HuiDqAZU-X!M@#}{$b?)m&~!DATQnS0J>v6`Y?N;> zYXW1wX|ihP1lRo}p-1t>OpbB-5-6UZH&n4+tHI_S!%R9&i4`W z*fP>ylIVic)S3gWHSZwC9e2mZF#&Bl_-<5JXNx=oX-<;Di(UR-XYq`_caJPkdz}8L zL_^YDH{b;$VU1k_uZQlc2?-7N6YeC;>=1+5s*slXyOdcp_<`4*3>;wTgX@1PPW@-E zvum^1q{h|h#S|(8H1+0W8a}phmo~h*dagrs9RppiE zW>(Vio=3y^dapK`dmK{N>YQ9o*y0lJIFCC>MtL5L?jX3Uv>l|I_l)GtrlK7W`dp^> zUg(DPr$c8kUpJT?r88^VzHaVp+3|4KWe+k>J$RX>l5ES=$5u))>|W0bwVI97Yf9vI zlXgT+%Bjm*>T3s1k1SZ$Y|~~AuT1tG3xsfq7RSnRg=tLt_SgtIeJ_)vuVwSz8-+A} zQR3IIII!;BW@1s-{V_73H8QcT+je$#F^{{vxRg-ThuAn1w3&S-$9$DXk>2e}<*8qhW!EF{miQyD=eRc6z zl+ygSEs(Nx?K5{Z;x9`H*K1=* zi%mP{zguoWaWDu(#jC1H=|=eNTVJ%%4~HXPTnkJ5?`2q6oTJ0znl+nTH4`SMbvie1 zB%UFstR0_w-q!DPxr;1e!oU*WJp8SFs+;z07aNzBMr-~w_$jO`HU-c_c?Zw9Y?83B zp2e4}ar#LWT-Vp+?g+Q_4gKaDP|yMywsw+^corGIIf|U30=I$8la+_Ve~e9lD^TJT z)KB-PmxZ;DhxL(y0^j2FbRJCl=P!PC4H0|+1G-o9fK7AZ6G_p)A^zJp$G-q%Y;&_G zvsog4@6$vT*bz`=eKo-Y;(Z3le#TJo*!o=v5n+V8Lg%D@pRH#|qp4G=e1YmY6z$DF zUNu?)Yv&bXpZ@t#N&i~d??#l_@nvKr@BtMw)%ID!gY*R67fb3KqJR9$JwOjmP6d>< zK-G>&>fpTh`^@6{0T%Ginz}%%%5`93&2OO-eeXfp3yNf{y1Sc0Y@9fKRPuW3jDh{n z_@-0FG_*d`vK0=a-l)d`+j?~)$+BCxKgL3y3h`*VvSqhecx5YPr|#i^Y6xeYc!-B3 zx@T9th`kUyxjO6wZ=k=!1JgCLi2s6mh&syJnV&ktOO0eP5;6C&mAvIz6{mSs69_dS15#gSn_|we=I?K3s%5EbG^xVCrzMKE#N{G?v<}Lq znebrgPh`RByL;JlSKqsSqF@`^;Bzd$TItcC|1NH4LbE1Mik@#)(tEK8jq@V%O$TS! z&c9ppN72^#!U~XKzUO>B_*3v&-o&8s+X$RHYb;vVK$?WGTcSMwhY|DlTq?EhTS3M< z`*@#!ckp2khxg3$1@0Z|0j~@t4+5V~6)jTS`5hq&@$$Ht)KJcsX@-S~zQ50!`pvAZ zvi6|_Y3jG;p?dB7 zwPa+|rD}Y%cC$91&1jgFiNVW2G=CBnHG`UDKxP(9UDiOh9)l2NUv2!cIet0=GCDzm z4S@Uc>q(A8<-M`g>r+ozDLL(1EHJ*_tyg7}HXXM@)CCD+r|OG4*i!L<12^Wu7>wJR z0b%7ouj|t{xvpWc%T@_t?q^}QSQ0O`x+u#%3+Cr-D5UJuX4-af1c1E2SirqGulzI= zw_m5*nG)@Dpo*XCYU@o`GT#wY{IGyhe*z~HlW#(2P4ZUbG zLwe@OW)$BhWn+;=J@?fJ-D;drbrxvaGR(Rm%5d(xqnEUuNpx1kH_N2O(qGQGkD5B2 ze|=s`^?ZYQA&!0Ywi9!r>zt){-iGP5Dl7O3LRk*4i+2X;r}NW_hpSrH_%n@@z~C!R zjsX&E0kv_<(Mtp$P!I8NNNRI4=xFJJY{^l_As zyt>wEfFDHVoi%R#JFBSQEq6B(gkAb5TfZa)^FHf=#p&$y&SN8M8XCM6A;Q zKADORje`35t1*=zBiV~QF>y9jEQ<;ngu1LW*=Gf!fUM6zRkffrj#19IM zsSVzl|%1khqgxf8zr&S|@+?mb%#mVoomx}D` zEz3=M&`z>x1CfKd+BT$aFv^f^%&gE!bJKWOlNWC3cg~B^Zg*W1i`#k$b^T+nxG+mN z9cCnGa1SLf+! z-dweU$|~R=RF6lglxu03GSqX{Dtj_Y%KMjhoO(kZkCMZ4qeU=K`Y|0HZ}Z(j$3Wp! z+l!8d7nI2F#U{o7&v`j3NIaZTDtHF>ifZxvwp|Xm3aJSaJ%waKL{VP%W$>yP2RDl1 z6nDc%BW}Yxy^ouRI)$E$$MGv05vs8kc_XKRMKO1xOr=Y%dU~d33!5ri-Tw;X5uzyWJIa;Rh(6Xo2SVhTN0!-;A%P2PEmqcUe;N3=GfIluksR*7XG zndUc%DcaF^8 zJr%XO3Z%C2$z@mjpzc$(rjk$&%_IXfBHh0&59bLBdtl^5S0`3Np?;i5CCo^kdNQYV z9iyYwob9m$LGk5Me}JjBeivJ}6p7$)(&Fj>@+5#Z#)nJ7|L!fWfdX`b&H$6KOM+eS zbx!TtMCH*~@K!}jWnJ)0ppkRcOs~Wuh4-PNV;DGcmj3^YpN};FQs3oku*+qCq6}W% zoqV29;lF0z^ajB${CtLPOosw)Yq>>208w=~y{*3-LG3|7iAEcwZcBGOVO`yhrwMXGhGgyQQ{h@4k?DI#=qJ|Z$QB3aGa+n>+4>e zf>EUMaXYTo&Y&L|z&XXiZ~vvsAyv>psRjU7=aei}J6(Nzj`lg6y(Zs&U%6~b-Bk@R z(=JwHTG6cDUMuNz;_Ui7oPVP8)HIvv6=)rDX|O=*@-p2WfNzikP)q(xVr|0!t%6iZ zTE(mzEc+Sh(2{-_(f?Q7cfZ56y>Ck+L9~eIL^%;9iD-!sj4n~4M2ixV=)I0EN}|^g z1Q9Jv7(^W{QKB10XAr#)MrWAsmUGVM%(&iv;GLgb_BDIh=z(oMM0>6zy_-7KnN6py)P$XNM&#(M3! zTyX3c>*IouDpe7#YaP6$P=Q-fU};B{81a(x0qNQXX@(NgAim^5O_PD!0cj93`kJ_o zVWEHJOs34`$)!mC1%T!nV0l>8{Z*M}D<7IM<<-$-!`qM=n=4PZs4RNC$Z^B=<-&>v z`KgrI{*M3E*mpsb-iujrT5&!?D6V1Mg=eF85#QddfUxMl;twLsD6Aj;PnGcl$C&Zk zA76bZ4P%!a6BdB^w8|jd*ooThU0kS@ip4AF_YT@1b;7QfCskcE>(lFpT52TwF_Z z{VhlC4|06-y4AC()Gn+I7cv%~vo7;%zo-mK?aF;_VJ*8Gkv-2Ca7QDz*Xo_sprs8A z(fWCC(PayH;J|>!q}XC#mURq%5LbItRYhvT19e6Cft$V$vf4 zkaYi^IXGTPSV?HQ3aBjHLpC7;E!*0ZfXtMrCyjudqyKf#FYO4>)NrT@(s&3B^ z!=;v9XpIf7>k2;C4_cdn5mtpm%M!a>i#H4C^ll->5z1-d>T!s+sAhR)24v+1WpT^4nM^i#BW!1!*;G8vl!pFlT>& zqV3`9NiQ3?<@fu{tGlIxRkdrWwP}TZMixT~e0iAp$OcV+{&{iyP6j(`zY`=4y3h+P zP+j3!o${ZugFn$)N1Z5dF}dk~Z1$>&NIzL7M}-6}`%l0WDGx1EpqW3y{JleD{8MYPYl)bBvXQHlP1E zr^3Q*D*L)L<)v8D`JPtznF&YY}6ncPv{Y-bxCV9Fc}~ zvUw^#cS>FJr<(I2^D%S*Ka~h5X>3lCFiJ%?Y*v_8)O(!D?12-n1oI1@PPl*ygya%= zsJE5E9#uTl8x3O0QL3LzC?pj%kcS(lmewRcwL1JZO((}jquIvqaYIf-Jn1H|=O;w> zc8;6T;aG9-L)>I)r0 zaVynJ7lVJSe6!S2^0MrRT!Wx<^5-Cea?&L__Q9bo*4Gw{TgRWrK&UGknY4;WscUS9 zz;jzU@z&s1su}9nQMUF2E?w4;AQ`>VKcZR2;sDQ9#5Eu8yV`T1g0E*?M4=wdpesZk z$1y3V;@5)JSc1DdJrPJ<4NN7I(3U1N#ho16{!X|l+?kGcn@BpQYUx|B&Jj83e)8z~ z_8;;3FH<)BEUd3v4ZVIpayj^r&6`vX^3V>H9ynO~((~C*FQF9VbsDZS+3GeFEO_ye zN}ZeT6+VjK2)OO|!o8&yjGCR{0lBKJuhRlL=EgfcONqyGs6_cBtGb;q?Lj-8JSOc6 zT6|8{6WhD`s#p`xLPOrl(rJ0akt+0axwk(tE+7)WMqX%i7Dd&Y&VnO6$Kt4U2q>;y+SGme>P=FfLaB5eu`}Iuu&>-EZ5eV?MqgSLsPwHgvMn!{Q9}v6Iaa z^&HbswH{-NhJjvI4d*M1m~mx`EMamh$o34x4+qoxJw?`vvfqq`8i5PkeWRe_v=&Ti z7LHG0=6PCFB!zJAk=Hr~&#TrZnIlvljF^Y)a$WA-9f_Gi-5#Z$<#zY(Nw) zXp22wVLgSfvRgtd-qr1gAaIC^_HJ#eFY>WN6)5d6FVU)QzQ+UYWz{@WWznEk_uAzyipNR zp9GN3rt>(`$QHxJ;#pPG%9ZvQkw|O@r`@)b1t5gHvN;k<2j1h68(Aeil{ev{Q^#(j zZXCKrnC<4XH364b;gQMr8hyI+S<}e#_wR?fs-2Pbn|4HGKJKG^EXStFH7}Gz8b2?n z4@6CV&954B?N7=ExIb?RBpZz+qANw@znxU~CJjZmp+G{0$fdZt1qp^L2`6{CLipj9 zf>s(0KgqNMG#60YD&zGT-S?q9f(yEcK93sZ*jY!tFFJi+z`rXl-sr!FF#5L# z%hYO!Vp-?r$684~naY{Ypf@l&g7AS!x3&2$2CRkh)OisW4HeJ{^#)0P=|Ts^ctR&!sKli)okupq>}r6L8wZ$GnG zeEWBK=rq%HVaRg=ol*Q7C0qo;M6qyq6n`oV7$u*%N_={?;!hr`O`nnvPp;udLH#BW zh3UcA7W@DQo<`vTH93Cqn!KS>8PMz8Myg`uYO~RD!bn&UWfi7cSFMDiUq+{R51_)op zYR0s|quPbYba6*1G}*f-(nHzuOcR$JirsTYIEW3DY^`k{*s)!k&o#vx6L zVgaW!#=jrl0zjzxliNTS2t$#32sgHI(qor(VE>@0QE9k!vS^1{tCJ8$%)xOY5>UQ1$I9&9n~p9=C(vyCIb7E+Wz%(zpY49W%LBB{iOyydsZ+! z=e<1vO{0RC^iKr7##A(mGZG)L`ni8G(%oKu0#s^^z>U;L9@9TIA?zU}pwu$-(s=eI z%bVv5TFB~-`{CWk%kZmapRTSmVz*ATRWVPz>nUzu`mmCI2SHZ<^HwSprAC;5Ax?2c zWS4T1rD)N)Tjk3;G5IwMpUNUSmbmE`LW%IJjJM`>tT{ncv@3%GAE_`tc31>0eDr1$ z+zovv(k?;epKr>-Y*SNqC8*(6trJrBMzlO(OD2zFurxiWFsx8??m?qy%VJpLAMPn0 z864f>J<3G=gKBB7vAGVnwRY}=m|RvE9kawNj~bP9?PX44Tm*HqmD)}|o-*sDdZ;5cO8L(i9I-x%{#yN$qa?+uVH@5OSwgrsy0EgO% zdkEgZVr2#fmV0cncGXg@CreSjEWm&U5UJMoJt%IL#Nk*11ptmEw5G%Oo~69U4MGpA zPqvE8i=7N%B?Fm6knl(}QZQrH&;j<7f#wI|{VxHp%wd`{L#D~!ENytBgQXmoXQbB|@nFOV8Ixmq5J_ zUugO=X)H{!wDH|qWw#e9M!BtfYq?5y9rgVkMeQOq&?QU62?~N~8T%)gG9zsAdvcEd ztbVM9K+Gz8zz1D>pfT4Eqb|Wasy12>6((5E5{s{byYW33f1u_L3NeE@_-z(Is@|) z6O3O!=HEE)mAGW~ zLp`Jse+ry#7f`hZULZ<*=_r^`aiWEa9xW_NUbFeE4563FUa9(*_+>o|>!$nIbQ-o!rV@Oj5MK!n$ilOi3z$u#M36(+ z#IL@Nvh(%ca-3O&6V`O+4?YHeE?*t4Cj9WrTp&UK@37?fEBWw78>SMd2nHL|ZV~IqilhJ0-zLC-Ht;lSO z5VMW8003lqHRGWh3RS@6nR}w{R11}_@8s~~oU?Oa@-V%J$?Rg_16)jXTd?JR|Cz3d zb2q)Tqw02U?Mx^?vd&8?pVl_ynV6ihB9~LA(M)EHy91*PqCWa2%~!S~r0>`v5HArZ zhU!CoCP-*77i)yY?uVLD;HDPY=|n@7Q0KlC5qFC{MP`_%;>3BO{gZ-H1p#MOEMp2`eh^d*=#q7dTofOUehnS{&f$Vy)lbC!eTYu(%7?e*lZBY5d zQ8asr1zJ%sQK=3tsy^*tX943YOI({yV^TUj5wC7R*oKg{D-9S&%!{GR@L@9kTX zNV?WmzrD@wS>FTTTqZCeT{w-dIfNJr|6893n)3dcSt0`kd7r2}yTbZ2@Bf8G9|28p zbK(FX3yY6_6AzA&fg<;tk&yS1z7>ED{~8{OD+l`gq8YE({y=ScX|IGEEQ-;si~{bL zZST#Ima+O%i_H+%F|sejh@d#-V_CD#h1_A>hKUy$tMn6pPQd-khZ?0Mt03zz=dHu0Zl&*Ql?lAz(E3joYCR-xj-Cr|>7kMk{NGPB<_&IxsMU)hGavDIh`*->_e@ zjKRgMk_~_%h6I3g_cTIi@o=0Z;IJFeJm{hVQ*Z!H%i*KMcRwibV&_64(DjV~8vU`& zVUAyVLz)xp3NiBme;z&UB0dUbof11bo{A1;kPZOO|0)9(p$o2dG$Fw__6rVL{knr< zY^@B}&8tJesb7{u^3!xDeaWC0;5giWJDossdIYFb=z8gMr}0w~u|qDZ4&zVOR55&$ zm~_B8S9^+@t*Zn}Ov>cG|W z1nAaq7oM~!q7ZugTThqFrPR*>nN08(>^giSLgq{mGSO-;IhLzoDTZ`>+%AY oh=2f#<#g3s2O2_nx^}-n@V&E@c~~rn;WPw_vMMr#Amg|H2MLTyMF0Q* literal 0 HcmV?d00001 diff --git a/docs/manuscript/figures/figure1_workflow.svg b/docs/manuscript/figures/figure1_workflow.svg new file mode 100644 index 0000000..fdbc9ff --- /dev/null +++ b/docs/manuscript/figures/figure1_workflow.svg @@ -0,0 +1 @@ +

    exemplars

    exemplars

    execute

    scored

    scored

    Evaluation modes (error attribution)

    Mode 1 · Skill-1 isolated — bundle vs expert labels

    Mode 2 · Skill-2 oracle — gold bundle into harmonizer

    Error-propagation gap = Mode 2 − Mode 3

    Mode 3 · end-to-end — full pipeline, errors propagate

    Agent pair under evaluation

    Curator — Skill 1
    retrieve · classify files · resolve location
    infer time series · detect manipulation · pick exemplar

    Curator bundle
    decision · payload/location files · coords · interval · qc_flag

    Harmonizer — Skill 2
    map variables · convert units · generate code + change-mapping

    Python transform code
    + change-mapping JSON

    ESS-DIVE dataset identifier
    (DOI / package ID)

    19 expert exemplars
    (in-context references)

    Raw package files

    Expert transform code

    Agent harmonized table

    Expert harmonized table
    (gold standard)

    PRIMARY ENDPOINT
    output-data equivalence
    cell-by-cell on the natural key

     \ No newline at end of file diff --git a/docs/manuscript/figures_and_tables.md b/docs/manuscript/figures_and_tables.md index 8d2fd94..fa3d92b 100644 --- a/docs/manuscript/figures_and_tables.md +++ b/docs/manuscript/figures_and_tables.md @@ -34,7 +34,7 @@ | # | Title | Notes | |---|---|---| -| 1 | Workflow + evaluation design schematic | **Drafted** — `figures/figure1_workflow.mmd` (Mermaid source) + `figures/figure1_workflow.html` (in-browser SVG render). Curator → bundle → Harmonizer → code+mapping → executed output vs. expert; three modes; gap. This is the main **Figure 1**. | +| 1 | Workflow + evaluation design schematic | **Drafted** — `figures/figure1_workflow.{mmd,svg,png}` (source + static vector/raster, rendered offline via Chromium) and `figure1_workflow.html` (live in-browser render). Curator → bundle → Harmonizer → code+mapping → executed output vs. expert; three modes; gap. This is the main **Figure 1**. Regenerate the static assets after editing the `.mmd`. | ## Presentation From 4d723eda8c67b2a9eac62b90970c6d88367a3e16 Mon Sep 17 00:00:00 2001 From: Claude Date: Tue, 30 Jun 2026 23:38:55 +0000 Subject: [PATCH 4/4] Vendor deck libs, add landing page and GitHub Pages workflow - vendor reveal.js 5.2.1 + Mermaid 11.16.0 under docs/manuscript/vendor/; point the slide deck and the Figure 1 renderer at them so both work with no external CDN (verified rendering with all network blocked) - add docs/manuscript/index.html landing page linking deck, figure, sections - add .github/workflows/pages.yml to publish docs/manuscript/ to GitHub Pages (auto-enables Pages; serves the landing page as the site root) - note hosting + offline behavior in the manuscript README Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_01FD9C8gFuJGBBYqemVFtwTx --- .github/workflows/pages.yml | 49 + docs/manuscript/README.md | 11 + docs/manuscript/figures/figure1_workflow.html | 11 +- docs/manuscript/index.html | 106 + docs/manuscript/slides/index.html | 12 +- docs/manuscript/vendor/VERSIONS.txt | 2 + docs/manuscript/vendor/mermaid/mermaid.min.js | 3587 +++++++++++++++++ docs/manuscript/vendor/reveal/reset.css | 30 + docs/manuscript/vendor/reveal/reveal.css | 9 + docs/manuscript/vendor/reveal/reveal.js | 9 + docs/manuscript/vendor/reveal/theme/white.css | 359 ++ 11 files changed, 4174 insertions(+), 11 deletions(-) create mode 100644 .github/workflows/pages.yml create mode 100644 docs/manuscript/index.html create mode 100644 docs/manuscript/vendor/VERSIONS.txt create mode 100644 docs/manuscript/vendor/mermaid/mermaid.min.js create mode 100644 docs/manuscript/vendor/reveal/reset.css create mode 100644 docs/manuscript/vendor/reveal/reveal.css create mode 100644 docs/manuscript/vendor/reveal/reveal.js create mode 100644 docs/manuscript/vendor/reveal/theme/white.css diff --git a/.github/workflows/pages.yml b/.github/workflows/pages.yml new file mode 100644 index 0000000..a228d71 --- /dev/null +++ b/.github/workflows/pages.yml @@ -0,0 +1,49 @@ +name: Deploy manuscript site to GitHub Pages + +# Serves docs/manuscript/ (landing page, slide deck, Figure 1) as a static site. +# The deck and figure vendor reveal.js + Mermaid under docs/manuscript/vendor/, +# so the published site needs no external CDN. + +on: + push: + branches: + - main + - claude/agentic-curation-research-docs-6dgt7l + paths: + - 'docs/manuscript/**' + - '.github/workflows/pages.yml' + workflow_dispatch: + +permissions: + contents: read + pages: write + id-token: write + +# Allow one concurrent deployment; newer pushes supersede in-progress ones. +concurrency: + group: pages + cancel-in-progress: true + +jobs: + deploy: + runs-on: ubuntu-latest + environment: + name: github-pages + url: ${{ steps.deployment.outputs.page_url }} + steps: + - name: Checkout + uses: actions/checkout@v4 + + - name: Configure Pages (auto-enable if needed) + uses: actions/configure-pages@v5 + with: + enablement: true + + - name: Upload manuscript directory as the site + uses: actions/upload-pages-artifact@v3 + with: + path: docs/manuscript + + - name: Deploy to GitHub Pages + id: deployment + uses: actions/deploy-pages@v4 diff --git a/docs/manuscript/README.md b/docs/manuscript/README.md index 37e2083..d89a69a 100644 --- a/docs/manuscript/README.md +++ b/docs/manuscript/README.md @@ -26,6 +26,9 @@ Working drafts for the paper describing this study: | `figures/` | Figure 1 schematic (Mermaid source + HTML render) | Drafted | | `slides/index.html` | HTML slide deck (reveal.js) | Drafted | +A landing page (`index.html`) links the deck, figure, and sections; it is the +homepage of the published site (see **Hosting** below). + Read in numeric order for a continuous draft. The intended journal/venue is TBD (candidates: *Scientific Data*, *Environmental Modelling & Software*, *PVLDB / SIGMOD data-systems venues*, or an agentic-AI venue — decide once results frame the @@ -51,6 +54,14 @@ strongest contribution). 6. Reconcile the reporting checklist at the end of `03_methods.md` (model version, seeds, counts). +## Hosting + +`.github/workflows/pages.yml` publishes this directory to GitHub Pages (it +auto-enables Pages and serves `index.html` as the homepage). reveal.js and +Mermaid are vendored under `vendor/`, so the deck and figure render with **no +external network** — both as a hosted site and when opened from disk. To preview +locally without hosting, open `index.html` (or `slides/index.html`) in a browser. + ## Citation policy `references.md` separates sources the research pass **verified by fetching** from diff --git a/docs/manuscript/figures/figure1_workflow.html b/docs/manuscript/figures/figure1_workflow.html index 02dc8e8..c497adc 100644 --- a/docs/manuscript/figures/figure1_workflow.html +++ b/docs/manuscript/figures/figure1_workflow.html @@ -81,12 +81,13 @@

    Figure 1. Agent-pair workflow and evaluation design.

    OUT -. scored .-> M3 -
    Source: figure1_workflow.mmd. Renders to SVG in-browser via Mermaid - (loaded from CDN). To export: open in a browser, right-click the diagram → save, or paste the - .mmd source into mermaid.live.
    +
    Source: figure1_workflow.mmd. Renders to SVG in-browser via a + vendored Mermaid (no network needed). Static .svg / .png renders sit + alongside this file. To re-export, paste the .mmd source into + mermaid.live.
    - + diff --git a/docs/manuscript/index.html b/docs/manuscript/index.html new file mode 100644 index 0000000..1db5c9b --- /dev/null +++ b/docs/manuscript/index.html @@ -0,0 +1,106 @@ + + + + + +Evaluating an LLM Agent Pair for Environmental Data Harmonization + + + +
    +
    +

    Watershed Function SFA · Berkeley Lab

    +

    Evaluating an LLM Agent Pair for
    Environmental Data Harmonization

    +
    +

    A curator + harmonizer + agent pair that turns heterogeneous DOE ESS-DIVE soil-moisture datasets into an analysis-ready + schema, benchmarked against a domain expert with leave-one-out cross-validation and blind prospective + evaluation.

    + Pre-results scaffold — manuscript & framework +
    +
    + +
    +
    +

    Explore

    +
    + + Deck +

    Slide deck →

    +

    15-slide overview: motivation, the agent pair, design, metrics, positioning, and results placeholders. Figure 1 rendered live.

    +     + + Figure 1 +

    Workflow schematic →

    +

    The agent-pair pipeline, three evaluation modes, and the cell-by-cell output-equivalence endpoint. Also as .svg/.png.

    +     + + Index +

    Manuscript README →

    +

    Section-by-section status and how the drafts map onto the codebase.

    +     +
    +
    + +
    +

    Manuscript sections

    + +

    Markdown files open as source on a plain static host. For rendered prose, read them on the + repository page; the deck and figure are full HTML and render anywhere.

    +
    + +
    + Static site served from docs/manuscript/. reveal.js and Mermaid are vendored under + vendor/ — the deck and figure need no network. Built with Claude Code. +
    +
    + + diff --git a/docs/manuscript/slides/index.html b/docs/manuscript/slides/index.html index 78c965a..aa4dfc9 100644 --- a/docs/manuscript/slides/index.html +++ b/docs/manuscript/slides/index.html @@ -4,8 +4,9 @@ Evaluating an LLM Agent Pair for Environmental Data Harmonization - - + + +