Accept official Statbel geometry date format

scripts/statbel_population_preflight.py

@@ -59,6 +59,7 @@ GEOMETRY_MEMBER_PATTERN = re.compile(
     r"^sh_statbel_statistical_sectors_(?:31370_)?(20[0-9]{2})0101\.geojson$",
     re.IGNORECASE,
 )
+GEOMETRY_DATE_PATTERN = re.compile(r"^(20[0-9]{2})([-/])([0-9]{2})\2([0-9]{2})$")
 POPULATION_SOURCE_PATTERN = re.compile(
     r"^/sites/default/files/files/opendata/bevolking/sectoren/"
     r"OPENDATA_SECTOREN_(20[0-9]{2})(?:_(NEW|OLD))?\.zip$",
@@ -123,6 +124,18 @@ def _layout_from_variant(variant: str | None) -> str:
     return variant.lower()
 
 
+def _normalize_geometry_date(value: Any) -> str | None:
+    raw_value = str(value or "").strip()
+    match = GEOMETRY_DATE_PATTERN.fullmatch(raw_value)
+    if not match:
+        return None
+    normalized = f"{match.group(1)}-{match.group(3)}-{match.group(4)}"
+    try:
+        return datetime.strptime(normalized, "%Y-%m-%d").date().isoformat()
+    except ValueError:
+        return None
+
+
 def _validate_source_url(url: str, pattern: re.Pattern[str], *, code: str) -> re.Match[str]:
     parsed = urlsplit(url)
     match = pattern.fullmatch(parsed.path)
@@ -428,7 +441,7 @@ def parse_geometry_archive(content: bytes, year: int) -> GeometryArchiveData:
                 "Geometry artifact contains duplicate sector codes.",
                 sector_code=sector_code,
             )
-        if str(properties.get("dt_situation") or "") != expected_date:
+        if _normalize_geometry_date(properties.get("dt_situation")) != expected_date:
             _fail(
                 "STATBEL_GEOMETRY_DATE_MISMATCH",
                 "Geometry situation date does not match the population reference year.",