feat(normalizer): add persons_tree skeleton + year extraction

tools/import-normalizer/persons_tree.py

@@ -0,0 +1,46 @@
+"""Normalize Personendatei 2.xlsx into canonical-persons-tree.json."""
+import argparse
+import datetime
+import json
+import re
+import sys
+from pathlib import Path
+
+import config
+import dates
+from persons import _strip_accents
+
+
+def _parse_year(raw: str | None) -> int | None:
+    """Extract a birth/death year from an Excel cell string.
+
+    Handles four cases:
+    1. ISO string (openpyxl date-formatted cell) → parse_date()
+    2. Numeric string that is an Excel serial (1-80000) → timedelta conversion
+    3. Any other string → parse_date()
+    4. Unresolvable → None
+    """
+    if raw is None:
+        return None
+    s = str(raw).strip()
+    if not s:
+        return None
+
+    # Try parse_date first (handles ISO, DD.MM.YYYY, year-only, month+year, etc.)
+    result = dates.parse_date(s)
+    if result.iso:
+        year = int(result.iso[:4])
+        # Reject years outside 1700-2100 (same guard as expand_year())
+        if 1700 <= year <= 2100:
+            return year
+
+    # If it's a pure integer string, try Excel serial conversion.
+    # parse_date() may parse large serials like "7568" as year 7568 or other edge cases.
+    if re.fullmatch(r"\d+", s):
+        n = int(s)
+        if 1 <= n <= 80_000:
+            d = datetime.date(1899, 12, 30) + datetime.timedelta(days=n)
+            if 1700 <= d.year <= 2100:
+                return d.year
+
+    return None