Saarland (Germany) LPIS converter

CHANGELOG.md

@@ -19,6 +19,8 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - Support Esri JSON and server-side filters in EsriRESTConverterMixin (rest_format, rest_params["where"])
 - Converter for Bavaria, Germany LPIS field blocks (de_by_block)
 - Converter for Hesse, Germany LPIS reference parcels
+- Converter for Saarland, Germany LPIS field blocks (de_sl_block)
+- Fix parcel sizes written in scientific notation being read 10,000x too large (de_sl parser)
 - Update vecorel-cli to v0.2.16:
   - Converter output is sorted by Hilbert distance
   - Commands exit with a non-zero exit code when they report a failure

fiboa_cli/datasets/de_sl.py

@@ -11,7 +11,9 @@ def parse_flik(x):
 
 
 def parse_size(x):
-    match = re.search(r"Size in ha: (\d+(\.\d+)?)+", x, re.I)
+    # Small parcels are written in scientific notation, e.g. "Size in ha: 1.0999999999999999E-4".
+    # Without the exponent the value is read as 1.1 ha instead of 1.1 m², a factor of 10,000.
+    match = re.search(r"Size in ha: (\d+(?:\.\d+)?(?:[eE][+-]?\d+)?)", x, re.I)
     return float(match.group(1)) if match else None
 
 

fiboa_cli/datasets/de_sl_block.py

@@ -0,0 +1,86 @@
+import re
+from urllib.parse import urlencode
+
+import requests
+from vecorel_cli.conversion.admin import AdminConverterMixin
+
+from ..conversion.convert_gml import gml_assure_columns
+from ..conversion.fiboa_converter import FiboaBaseConverter
+from .commons.de_iacs import DEIACSMixin
+from .de_sl import parse_flik, parse_size
+
+BASE_URL = "https://geoportal.saarland.de/gdi-sl/inspirewfs_Bodenbedeckung_LPIS"
+PARAMS = {
+    "service": "WFS",
+    "version": "2.0.0",
+    "request": "GetFeature",
+    "typeNames": "lcv:LandCoverUnit",
+}
+# The WFS accepts larger pages, but 2500 keeps each response around 7 MB. It is also the ceiling
+# of the limit allowlist on the OGC API - Features endpoint for the same data.
+PAGE_SIZE = 2500
+
+
+class DESLBlockConverter(AdminConverterMixin, DEIACSMixin, FiboaBaseConverter):
+    id = "de_sl_block"
+    admin_subdivision_code = "SL"
+    short_name = "Germany, Saarland (LPIS)"
+    title = "Field blocks for Saarland, Germany"
+    description = """
+The reference parcels ("Referenzschläge") of the Saarland Land Parcel Identification System (LPIS),
+the reference system for area-based agricultural payments. The data is transformed into the INSPIRE
+"Land Cover" data model, so each parcel carries a land cover class from the national IACS code list
+alongside its FLIK. The area given is the eligible parcel size recorded in InVeKoS, not the
+geometric area of the polygon. The complementary application parcels ("Antragsschläge") are
+published separately.
+"""
+
+    provider = "Ministerium für Umwelt, Klima, Mobilität, Agrar und Verbraucherschutz <https://geoportal.saarland.de/spatial-objects/384>"
+    attribution = "© GDI-SL 2026, CC BY 4.0"
+    license = "CC-BY-4.0"
+
+    extensions = {"https://fiboa.org/flik-extension/v0.2.0/schema.yaml"}
+
+    columns = {
+        "geometry": "geometry",
+        "flik": ("flik", "id"),  # derived in migrate(); unique
+        "area_type": "crop:code",  # added in file_migration(), trimmed in migrate()
+        "area": "metrics:area",  # derived in migrate(); in hectares
+    }
+
+    def get_urls(self):
+        # numberReturned is always reported as 0 by this server, so the page count has to come
+        # from a hits request rather than from the responses themselves.
+        hits = requests.get(BASE_URL, params={**PARAMS, "resultType": "hits"})
+        hits.raise_for_status()
+        total = int(re.search(r'numberMatched="(\d+)"', hits.text).group(1))
+
+        query = urlencode({**PARAMS, "count": PAGE_SIZE})
+        return {
+            f"{BASE_URL}?{query}&startIndex={start}": f"de_sl_block_{start}.gml"
+            for start in range(0, total, PAGE_SIZE)
+        }
+
+    def file_migration(self, gdf, path, uri, layer=None):
+        # The land cover class is a nested xlink attribute, which the GML driver does not guess
+        # into its generated schema. GDAL separates the steps of an element path with "|".
+        return gml_assure_columns(
+            gdf,
+            path,
+            uri,
+            layer,
+            area_type={
+                "ElementPath": "landCoverObservation|LandCoverObservation|class@href",
+                "Type": "String",
+                "Width": 255,
+            },
+        )
+
+    def migrate(self, gdf):
+        # The FLIK and the size are both encoded in the INSPIRE description, e.g.
+        # "Size in ha: 0.11206, flik: DESLLI0000248744"
+        gdf["flik"] = gdf["description"].apply(parse_flik)
+        gdf["area"] = gdf["description"].apply(parse_size)
+        # …/codelist/de.iacs/AgriculturalAreaTypeValue/GL -> GL
+        gdf["area_type"] = gdf["area_type"].str.rsplit("/", n=1).str[-1]
+        return super().migrate(gdf)