Add process engineering and chemistry value sets

[cmungall]

Summary

This PR adds comprehensive value sets for process engineering domains and chemistry identifiers, along with extended units for energy, power, and flow rates. These new value sets support process flowsheet modeling (particularly the PISCES Standard Flowsheet Format) and chemical substance identification.

Key Changes

New Process Engineering Value Sets

• Unit Operations (process_engineering/unit_operations.yaml): 100+ unit operation types organized by class (momentum transfer, heat transfer, mass transfer separations, membrane separations, mechanical separations, solids processing, reaction, and storage), plus 50+ process equipment types
• Process Streams (process_engineering/process_streams.yaml): Stream roles, phase states, and utility types for flowsheet modeling
• Process Industries (process_engineering/process_industries.yaml): Industry categories and operation modes (batch/continuous)
• Process Modeling (process_engineering/process_modeling.yaml): Design/simulation methods, flowsheet solution approaches, and process simulator software
• Thermodynamics (process_engineering/thermodynamics.yaml): Equations of state, activity coefficient models, property packages, mixing rules, and Poynting corrections

New Chemistry Value Sets

• Chemical Identifiers (chemistry/identifiers.yaml): Identifier schemes (CAS RN, SMILES, InChI, InChIKey, etc.) for referencing chemical substances

Extended Units

• Energy units (units/measurements.yaml): Joule, kilojoule, megajoule, watt-hour, kilowatt-hour, calorie, kilocalorie, BTU with conversion factors
• Power units: Watt, kilowatt, megawatt, horsepower, BTU/hour with conversions
• Flow rate units: Mass flow rate (kg/s, kg/h, t/h, t/d, lb/h), molar flow rate (mol/s, mol/h, kmol/h), and volumetric flow rate units

New Business Value Sets

• Currencies (business/currencies.yaml): ISO 4217 currency codes for 50+ actively circulating currencies with numeric codes, symbols, and minor units

Bioprocessing Updates

• Extended bioprocessing/scale_up.yaml with clarification, flocculation, ultrafiltration, diafiltration, and tangential flow filtration operations

Implementation Details

• All new enums follow project conventions: CamelCase names, UPPER_SNAKE_CASE permissible values, with ontology mappings via meaning: field where applicable
• Extensive use of OBO ontology terms (CHMO, OBI, PROCO, UO) for semantic grounding
• Annotations include symbols, conversion factors, abbreviations, and classification metadata
• All new value sets marked as DRAFT status with contributor attribution
• Updated main schema (valuesets.yaml) to import all new modules
• Updated ontology term caches for CHMO, UO, and OBI

https://claude.ai/code/session_01JkQ7CEGzWMmM8ReU2KTwiP

[claude]

Claude finished @cmungall's task in 3m 34s —— View job

---

PR Review: Process engineering and chemistry value sets

I read through all 13 changed files, cross-checked every ontology CURIE against the term caches, and spot-checked the unit conversion factors. Overall this is a high-quality, well-structured PR — consistent conventions, good descriptions, and the new modules are wired into valuesets.yaml correctly. A few things worth addressing before merge.

• Gather context (read changed files)
• Review schema conventions (naming, structure)
• Verify ontology mappings (CURIEs/IDs)
• Check unit conversion factors & data correctness
• Post review feedback

✅ What's solid

• Ontology IDs check out. Every CHMO/OBI/UO CURIE I checked resolves in the caches with a matching label, and the title: on each mapped value matches the ontology label (e.g. DISTILLATION→CHMO:0001532 distillation, GASIFICATION→CHMO:0001501 gasification, BIOREACTOR→OBI:0001046 bioreactor). The cache files were updated for the genuinely new terms. This is exactly the discipline CLAUDE.md asks for.
• Conversion factors are correct. I verified the energy, power, and flow-rate factors — e.g. BTU=1055.06 J, cal=4.184 J, kWh=3 600 000 J, hp=745.7 W, t/d=0.0115741 kg/s, gpm=6.30902e-5 m³/s, cfm=0.000471947 m³/s. All accurate.
• Currency numeric_code stored as strings ('036', '032', '048') correctly preserves ISO 4217 leading zeros — good call. Using alpha-3 codes as permissible-value names is the right "existing standard" exception per CLAUDE.md.
• Clear domain grouping via class annotations, and the UnitOperationType (function) vs ProcessEquipmentType (asset) split is a clean, well-documented modeling choice.

🔶 Suggestions

1. CRYSTALLIZATION → PROCO:0000052 is the one unverifiable mapping (unit_operations.yaml:131-134)
Unlike CHMO/OBI/UO, there is no PROCO term cache in the repo, so this CURIE isn't covered by the cache-based validator and I couldn't confirm it. Meanwhile CHMO:0001477 "crystallisation" is already cached and used elsewhere. Recommend switching to CHMO:0001477 for consistency and verifiability (and dropping the otherwise-unused PROCO prefix). Please double-check PROCO:0000052 against OLS before merging if you want to keep it.
Fix this →

2. Inconsistent numeric typing in conversion annotations (units/measurements.yaml)
Some factors are quoted strings, others are bare YAML floats:

KILOJOULE:  conversion_to_joule: '1000'   # string
MEGAJOULE:  conversion_to_joule: 1e6      # float (1000000.0)
MEGAWATT_HOUR: conversion_to_joule: 3.6e9 # float

Within one annotation key the values should have a uniform type or a consumer parsing them gets a mix of str and float. Recommend quoting them all (e.g. '1000000' / '3.6e9') to match the majority style. Same pattern in conversion_to_watt, conversion_to_kg_per_s, etc.
Fix this →

3. Cooling utilities categorized as HEAT (process_streams.yaml:140-159)
COOLING_WATER, CHILLED_WATER, REFRIGERANT, HOT_OIL, and BRINE all carry utility_category: HEAT. That's confusing — HEAT reads as heating. Consider either a neutral label like THERMAL/HEAT_TRANSFER, or splitting into HEATING vs COOLING.

4. PURIFICATION uses an undocumented class value (unit_operations.yaml:299-304)
It's tagged unit_operation_class: SEPARATION, but the enum description enumerates the classes as momentum transfer, heat transfer, mass-transfer separations, mechanical separations, solids processing, reaction, and storage — SEPARATION isn't one of them, and it's the only value using it. Use an existing class (e.g. MASS_TRANSFER_SEPARATION) or add the new category to the enum description.

💭 Minor / optional

• The classification annotations (unit_operation_class, equipment_class, utility_category, method_class, pisces_category, …) are free-text strings. They're effectively controlled vocabularies — a candidate for promoting to their own enums later so they're validated, but not blocking for DRAFT.
• ChemicalIdentifierScheme and CurrencyCode have no meaning: mappings. Reasonable (most are named methods/codes without OBO equivalents), though identifier schemes could optionally map to IAO/EDAM data-format terms.

⚠️ Validation note

I was unable to run just validate / just site in this sandbox (command approval / uv + network restrictions), so I couldn't execute linkml-term-validator or the generators end-to-end. The YAML structure and CURIE checks above were done by inspection and cache lookup. Please make sure CI's just site + just validate pass — particularly to confirm the PROCO:0000052 mapping (item 1) doesn't trip the term validator.
• branch claude/pisces-terms-coverage-nae73q

[claude]

Claude finished @cmungall's task in 3m 32s —— View job

---

Addressed review feedback

I applied the still-applicable items from the earlier review. Changes are committed locally as 51fd916, but I could not push — see the note at the bottom.

• Fix CRYSTALLIZATION mapping — PROCO:0000052 → CHMO:0001477 (crystallisation, already in cache/chmo/terms.csv); removed now-unused PROCO prefix from unit_operations.yaml
• Fix PURIFICATION undocumented class — SEPARATION → MASS_TRANSFER_SEPARATION
• Re-label cooling utilities — split misleading HEAT into HEATING (steam variants, hot oil) vs COOLING (cooling water, chilled water, refrigerant, brine) in process_streams.yaml
• Conversion-factor typing (review item 2) — already resolved: the EnergyUnitEnum/PowerUnitEnum flagged in the review were dropped in commit 2476be1 (slot-name collision fix). The remaining new flow-rate enums (MassFlowRateUnitEnum, MolarFlowRateUnitEnum, VolumetricFlowRateUnitEnum) already use consistently-quoted string factors. No change needed.
• Run just site / just validate — just/uv aren't installed in this sandbox, so I couldn't run them locally. CI will validate; the CHMO:0001477 term is cached so the term validator should pass.

Diff summary

process_engineering/unit_operations.yaml

-  PROCO: http://purl.obolibrary.org/obo/PROCO_
...
   CRYSTALLIZATION:
-    title: crystallization
+    title: crystallisation
-    meaning: PROCO:0000052
+    meaning: CHMO:0001477
...
   PURIFICATION:
-    unit_operation_class: SEPARATION
+    unit_operation_class: MASS_TRANSFER_SEPARATION

process_engineering/process_streams.yaml (UtilityType)

   STEAM / *_PRESSURE_STEAM / HOT_OIL:   utility_category: HEAT → HEATING
   COOLING_WATER / CHILLED_WATER /
   REFRIGERANT / BRINE:                  utility_category: HEAT → COOLING

⚠️ Could not push

The push was rejected: Permission to linkml/valuesets.git denied to github-actions[bot] (HTTP 403). The commit exists locally but isn't on the branch yet. This looks like a token-permissions issue on this run rather than anything in the change itself.

To get these in, either:

1. Re-trigger me once the workflow has contents: write for the claude[bot] token, or
2. Apply the diff above directly — it's three small, mechanical edits across two files.

I left items 3 (free-text classification annotations → future enums) and the meaning:-on-identifier-schemes note as optional/non-blocking, consistent with the original review.
• branch claude/pisces-terms-coverage-nae73q