Add hfb_pwn dataset and align drains_pwn with the Bergen model

src/nhflodata/data/mockup/drains_pwn/v1.0.0/drains_pwn.geojson

@@ -5,10 +5,10 @@
 "xy_coordinate_resolution": 0.01,
 "bbox": [ 107080.87, 521548.54, 107453.90, 521643.63 ],                                                                            
 "features": [
-{ "type": "Feature", "properties": { "name": "drain_dwk", "elevation": 0.73, "conductance_per_meter": 1.0, "conductance_per_squared_meter": null, "bgt-identificatie": "", "mover_lake_name": "", "comment": "wordt met pomp bemalen" }, "bbox": [ 107080.87, 521548.54, 107253.51, 521590.54 ], "geometry": { "type": "LineString", "coordinates": [ [ 107080.87, 521548.54 ], [ 107253.51, 521590.54 ] ] } },
-{ "type": "Feature", "properties": { "name": "drain_pomgebouw_ZUID", "elevation": -0.62, "conductance_per_meter": 1.0, "conductance_per_squared_meter": null, "bgt-identificatie": "", "mover_lake_name": "", "comment": "via put uitloop op -0.5mNAP" }, "bbox": [ 107341.82, 521617.28, 107406.43, 521639.18 ], "geometry": { "type": "LineString", "coordinates": [ [ 107342.3, 521639.18 ], [ 107341.82, 521617.28 ], [ 107394.21, 521617.91 ], [ 107406.43, 521627.28 ] ] } },
-{ "type": "Feature", "properties": { "name": "drain_pomgebouw_NRD", "elevation": -0.8, "conductance_per_meter": 1.0, "conductance_per_squared_meter": null, "bgt-identificatie": "", "mover_lake_name": "", "comment": "via put uitloop op -0.5mNAP" }, "bbox": [ 107347.06, 521635.69, 107407.07, 521643.63 ], "geometry": { "type": "LineString", "coordinates": [ [ 107347.06, 521643.15 ], [ 107398.49, 521643.63 ], [ 107407.07, 521635.69 ] ] } },
-{ "type": "Feature", "properties": { "name": "drain_dienstgebouw", "elevation": -0.844, "conductance_per_meter": 1.0, "conductance_per_squared_meter": null, "bgt-identificatie": "", "mover_lake_name": "", "comment": "uitloop onder water op -0.5mNAP" }, "bbox": [ 107422.94, 521619.66, 107453.9, 521636.01 ], "geometry": { "type": "LineString", "coordinates": [ [ 107453.9, 521636.01 ], [ 107453.74, 521620.29 ], [ 107436.12, 521619.66 ], [ 107422.94, 521629.18 ] ] } },
+{ "type": "Feature", "properties": { "name": "drain_dwk", "elevation": 0.73, "conductance_per_meter": 1.0, "conductance_per_squared_meter": null, "bgt-identificatie": "", "mover_lake_name": "vijver 1, 2 en 3", "comment": "wordt met pomp bemalen" }, "bbox": [ 107080.87, 521548.54, 107253.51, 521590.54 ], "geometry": { "type": "LineString", "coordinates": [ [ 107080.87, 521548.54 ], [ 107253.51, 521590.54 ] ] } },
+{ "type": "Feature", "properties": { "name": "drain_pomgebouw_ZUID", "elevation": -0.5, "conductance_per_meter": 1.0, "conductance_per_squared_meter": null, "bgt-identificatie": "", "mover_lake_name": "vijver 4", "comment": "via put uitloop op -0.5mNAP" }, "bbox": [ 107341.82, 521617.28, 107406.43, 521639.18 ], "geometry": { "type": "LineString", "coordinates": [ [ 107342.3, 521639.18 ], [ 107341.82, 521617.28 ], [ 107394.21, 521617.91 ], [ 107406.43, 521627.28 ] ] } },
+{ "type": "Feature", "properties": { "name": "drain_pomgebouw_NRD", "elevation": -0.5, "conductance_per_meter": 1.0, "conductance_per_squared_meter": null, "bgt-identificatie": "", "mover_lake_name": "vijver 4", "comment": "via put uitloop op -0.5mNAP" }, "bbox": [ 107347.06, 521635.69, 107407.07, 521643.63 ], "geometry": { "type": "LineString", "coordinates": [ [ 107347.06, 521643.15 ], [ 107398.49, 521643.63 ], [ 107407.07, 521635.69 ] ] } },
+{ "type": "Feature", "properties": { "name": "drain_dienstgebouw", "elevation": -0.5, "conductance_per_meter": 1.0, "conductance_per_squared_meter": null, "bgt-identificatie": "", "mover_lake_name": "vijver 4", "comment": "uitloop onder water op -0.5mNAP" }, "bbox": [ 107422.94, 521619.66, 107453.9, 521636.01 ], "geometry": { "type": "LineString", "coordinates": [ [ 107453.9, 521636.01 ], [ 107453.74, 521620.29 ], [ 107436.12, 521619.66 ], [ 107422.94, 521629.18 ] ] } },
 { "type": "Feature", "properties": { "name": "bassin", "elevation": -0.5, "conductance_per_meter": null, "conductance_per_squared_meter": 1.0, "bgt-identificatie": "", "mover_lake_name": "vijver 4", "comment": "" }, "bbox": [ 107403.66, 521624.72, 107425.88, 521641.05 ], "geometry": { "type": "Polygon", "coordinates": [ [ [ 107403.66, 521638.85 ], [ 107409.53, 521640.66 ], [ 107425.54, 521641.05 ], [ 107425.88, 521627.06 ], [ 107423.49, 521624.98 ], [ 107412.31, 521624.72 ], [ 107404.06, 521625.46 ], [ 107403.66, 521638.85 ] ] ] } }
 ]
 }

src/nhflodata/data/mockup/drains_pwn/v1.0.0/drains_pwn.py

@@ -22,15 +22,24 @@
 gdf_dir = get_abs_data_path("drains_en_damwand_bergen")
 gdf = gpd.read_file(gdf_dir / "drains_en_damwand_bergen.geojson")
 gdf = gdf.loc[gdf.naam.str.startswith("drain")]
-gdf["elevation"] = gdf[["diepte_van", "diepte_tot"]].min(axis=1)
 gdf = gdf.rename(columns={"naam": "name", "opmerking": "comment"})
 
+# Drain stage = the controlling drainage (outlet/weir) level, following the Bergen model
+# (04v2pwnbergenmodel). This is the DRN elevation, NOT the pipe invert min(diepte_van, diepte_tot).
+# The pump-building drains discharge via an outlet ("uitloop") at -0.5 m NAP (see the 'comment'
+# field), while drain_dwk is pumped and discharges at +0.73 m NAP.
+gdf["elevation"] = -0.5
+gdf.loc[gdf["name"] == "drain_dwk", "elevation"] = 0.73
+
 # Add conductances
 # m3/day per meter head difference per meter drain length
 gdf["conductance_per_meter"] = 1.0  # Default conductance
 gdf["conductance_per_squared_meter"] = np.nan  # Default conductance for open water
 gdf["bgt-identificatie"] = ""  # Placeholder for BGT identification
-gdf["mover_lake_name"] = ""  # Placeholder for mover lake boundary name, to move drained water to.
+# Route drained water to a lake via the MVR (mover) package, following the Bergen model:
+# drain_dwk -> "vijver 1, 2 en 3"; the pump-building drains -> "vijver 4".
+gdf["mover_lake_name"] = "vijver 4"
+gdf.loc[gdf["name"] == "drain_dwk", "mover_lake_name"] = "vijver 1, 2 en 3"
 gdf = gdf[columns_to_keep]
 
 # %% Add open water drainage

src/nhflodata/data/mockup/hfb_pwn/v1.0.0/hfb_pwn.geojson

@@ -0,0 +1,11 @@
+{
+"type": "FeatureCollection",
+"name": "hfb_pwn",
+"crs": { "type": "name", "properties": { "name": "urn:ogc:def:crs:EPSG::28992" } },
+"xy_coordinate_resolution": 0.01,
+"bbox": [ 107034.61, 521536.91, 107511.14, 521642.41 ],                                                                            
+"features": [
+{ "type": "Feature", "properties": { "name": "damwand_pompgebouw", "hydchr": 0.01, "depth": 5.0, "elevation": null, "comment": "" }, "bbox": [ 107280.34, 521589.56, 107511.14, 521619.63 ], "geometry": { "type": "LineString", "coordinates": [ [ 107280.34, 521589.56 ], [ 107511.14, 521619.63 ] ] } },
+{ "type": "Feature", "properties": { "name": "damwand_dwk", "hydchr": 0.01, "depth": 5.0, "elevation": null, "comment": "tot in veenlaag (0mNAP) OF SDL dieper (-5mNAP)" }, "bbox": [ 107034.61, 521536.91, 107280.34, 521642.41 ], "geometry": { "type": "LineString", "coordinates": [ [ 107034.61, 521642.41 ], [ 107058.75, 521592.47 ], [ 107070.99, 521536.91 ], [ 107280.34, 521589.56 ] ] } }
+]
+}

src/nhflodata/data/mockup/hfb_pwn/v1.0.0/hfb_pwn.py

@@ -0,0 +1,64 @@
+from pathlib import Path
+
+import geopandas as gpd
+import numpy as np
+from nhflotools.geoconverter.utils import round_bounds, validate_crs, validate_geometry
+
+from nhflodata import get_abs_data_path
+
+# Properties are those consumed by nlmod.gwf.hfb.get_hfb_spd(ds, linestrings, hydchr,
+# depth=None, elevation=None), which places a horizontal flow barrier on the cell faces
+# crossed by the line and spans the model layers vertically:
+#   - hydchr:    hydraulic characteristic of the barrier [1/day]; 1/100 == resistance of 100 days.
+#   - depth:     depth of the barrier below groundlevel [m, positive]; the top is at groundlevel.
+#   - elevation: alternative to depth: absolute elevation of the barrier BOTTOM [m NAP].
+# Use either depth or elevation (not both); here depth is populated and elevation left null.
+columns_to_keep = [
+    "name",
+    "hydchr",
+    "depth",
+    "elevation",
+    "comment",
+    "geometry",
+]
+
+# %% Create horizontal flow barrier (sheet pile wall / "damwand") input
+# The damwand features live in the same source dataset as the drains.
+gdf_dir = get_abs_data_path("drains_en_damwand_bergen")
+gdf = gpd.read_file(gdf_dir / "drains_en_damwand_bergen.geojson")
+gdf = gdf.loc[gdf.naam.str.startswith("damwand")]
+gdf = gdf.rename(columns={"naam": "name", "opmerking": "comment"})
+gdf["comment"] = gdf["comment"].fillna("")
+
+# diepte_van/diepte_tot are DEPTHS below groundlevel (negative = downward), so the barrier is
+# defined by its construction depth, not by an absolute elevation. `depth` is the deepest extent
+# below groundlevel as a POSITIVE number of metres -- exactly what nlmod.gwf.hfb.get_hfb_spd
+# expects: the barrier top sits at groundlevel and extends `depth` metres straight down.
+# `elevation` (absolute m NAP) is therefore left null; get_hfb_spd requires exactly one of the two.
+#
+# Deliberate modelling choice: representing the wall by construction depth (not an absolute tie-in)
+# means that where the ground surface is high (dunes; AHN up to ~7.5 m NAP along this alignment)
+# the wall bottom stays above the veen/SDL aquitard (0/-5 m NAP) and provides little cutoff there.
+# This is intended -- the damwand is its physical 5 m of sheet pile, not forced down to the aquitard.
+gdf["depth"] = -gdf[["diepte_van", "diepte_tot"]].min(axis=1)
+gdf["elevation"] = np.nan
+
+# Hydraulic characteristic [1/day]. Bergen model (04v2pwnbergenmodel) used 1/100,
+# i.e. a resistance of 100 days for a unit gradient.
+gdf["hydchr"] = 1.0 / 100.0
+
+gdf = gdf[columns_to_keep]
+
+# Checks
+validate_crs(gdf)
+for idx, geom in enumerate(gdf.geometry):
+    is_valid, error = validate_geometry(geom)
+    if not is_valid:
+        msg = f"Geometry at index {idx} is invalid: {error}"
+        raise ValueError(msg)
+
+out_geojson = Path(__file__).parent / "hfb_pwn.geojson"
+gdf[columns_to_keep].to_file(out_geojson, driver="GeoJSON", write_bbox=True, coordinate_precision=2)
+
+bounds = dict(zip(["minx", "miny", "maxx", "maxy"], gdf.total_bounds, strict=False))
+print("extent:", round_bounds(bounds, rounding_interval=1000.0).values())  # noqa: T201

src/nhflodata/data/repository.yaml

@@ -45,6 +45,27 @@ data:
       changelog:
         previous_version: 0.0.0
         log: Initial version of drains of PWN. Uses mostly data from drains_en_damwand_bergen v1.0.0.
+  hfb_pwn:
+    - version_nhflo: 1.0.0
+      owner: PWN
+      publication_date: "2026-07-01"
+      version_owner: 1.0.0
+      description_short: Horizontal flow barriers (sheet pile walls) of PWN
+      description_long: Horizontal flow barriers (sheet pile walls / damwanden) of PWN, near pompstation Bergen
+      contact: bas.des.tombe@pwn.nl
+      timezone: Europe/Amsterdam
+      extent:
+        - 107000.0
+        - 521000.0
+        - 108000.0
+        - 522000.0
+      paths:
+        local: hfb_pwn/v1.0.0
+        nhflo_server: /data/hfb_pwn/v1.0.0
+        mockup: mockup/hfb_pwn/v1.0.0
+      changelog:
+        previous_version: 0.0.0
+        log: Initial version of horizontal flow barriers of PWN. Uses mostly data from drains_en_damwand_bergen v1.0.0.
   rivers_pwn:
     - version_nhflo: 1.0.0
       owner: PWN

src/nhflodata/get_paths.py

@@ -7,10 +7,10 @@
 import re
 from datetime import datetime
 from pathlib import Path
-from zoneinfo import ZoneInfo
 
 import yaml
 from yaml.loader import SafeLoader
+from zoneinfo import ZoneInfo
 
 logger = logging.getLogger(__name__)
 

tests/schema_repository.yaml

@@ -44,7 +44,8 @@ dataset_version:
   timezone: str(required=True)
 
   # Spatial extent [minx, miny, maxx, maxy]
-  extent: list(num(), length=4, required=True)
+  # Note: yamale ignores `length=` on list(); use min/max to enforce exactly 4 elements.
+  extent: list(num(), min=4, max=4, required=True)
 
   # Path information
   paths: include('paths', required=True)

tests/test_geojson_content.py

@@ -0,0 +1,49 @@
+"""Content contract for the committed GeoJSON artifacts of the drain/HFB datasets.
+
+The registry test (test_repository_yaml) and the folder test (test_mockup_folder_structure)
+verify metadata and folder structure but never open the GeoJSON files. This lightweight,
+dependency-free (stdlib json) check verifies that the committed artifacts consumers rely on are
+structurally sound: the expected file exists, has non-empty features, declares a CRS, and every
+feature carries the property columns its model consumer reads. It deliberately does NOT assert
+specific data values, which are allowed to change for a data repository.
+"""
+
+import json
+
+import pytest
+
+from nhflodata.get_paths import get_abs_data_path
+
+# Required property columns per dataset, matching what the model consumers read.
+REQUIRED_COLUMNS = {
+    "drains_pwn": {
+        "name",
+        "elevation",
+        "conductance_per_meter",
+        "conductance_per_squared_meter",
+        "bgt-identificatie",
+        "mover_lake_name",
+        "comment",
+    },
+    "hfb_pwn": {"name", "hydchr", "elevation", "depth", "comment"},
+}
+
+
+@pytest.mark.parametrize("name", sorted(REQUIRED_COLUMNS))
+def test_geojson_content_contract(name):
+    """Committed <name>.geojson has non-empty features, a CRS, and all required columns."""
+    path = get_abs_data_path(name=name, version="latest", location="mockup")
+    geojson_file = path / f"{name}.geojson"
+    assert geojson_file.exists(), f"{geojson_file} is missing"
+
+    with open(geojson_file, encoding="utf-8") as f:
+        gj = json.load(f)
+
+    features = gj.get("features", [])
+    assert features, f"{name}.geojson has no features"
+    assert gj.get("crs"), f"{name}.geojson has no CRS"
+
+    required = REQUIRED_COLUMNS[name]
+    for i, feat in enumerate(features):
+        missing = required - set(feat.get("properties", {}))
+        assert not missing, f"{name}.geojson feature {i} is missing columns: {sorted(missing)}"

tests/test_mockup_folder_structure.py

@@ -38,7 +38,9 @@ def test_mockup_folder_structure():
                     location="mockup",
                 )
             )
-            if not path.endswith(version["paths"]["mockup"]):
+            # Normalize separators so the comparison also holds on Windows, where
+            # str(Path(...)) uses backslashes while the mockup path literal uses "/".
+            if not os.path.normpath(path).endswith(os.path.normpath(version["paths"]["mockup"])):
                 pytest.fail(f"Path in repository.yaml does not match the expected path: {version['paths']['mockup']}")
 
             assert os.path.exists(path), f"Path {path} does not exist"