From fa4b6b5fc29e873d691906d57374b2543c8902db Mon Sep 17 00:00:00 2001
From: Marcel <marcel@familienarchiv>
Date: Mon, 25 May 2026 21:03:46 +0200
Subject: [PATCH] feat(normalizer): add SPOUSE_OF resolution to persons_tree

---
 tools/import-normalizer/persons_tree.py       | 35 +++++++++++++++
 .../tests/test_persons_tree.py                | 45 +++++++++++++++++++
 2 files changed, 80 insertions(+)

diff --git a/tools/import-normalizer/persons_tree.py b/tools/import-normalizer/persons_tree.py
index af3ceb09..9b7f6095 100644
--- a/tools/import-normalizer/persons_tree.py
+++ b/tools/import-normalizer/persons_tree.py
@@ -207,3 +207,38 @@ def _deduplicate(persons: list[dict]) -> tuple[list[dict], list[str]]:
         result.append(p)
 
     return result, skipped
+
+
+def _resolve_spouses(
+    persons: list[dict], index: dict[str, list[str]]
+) -> tuple[list[dict], list[dict]]:
+    """Emit SPOUSE_OF edges from each person's _spouse_raw field."""
+    relationships: list[dict] = []
+    unresolved: list[dict] = []
+    emitted: set[frozenset] = set()
+
+    for p in persons:
+        raw = (p.get("_spouse_raw") or "").strip()
+        if not raw:
+            continue
+        row_id = p["rowId"]
+        matched_id, reason = _resolve_one(raw, index)
+        if matched_id:
+            edge = frozenset([row_id, matched_id])
+            if edge not in emitted:
+                emitted.add(edge)
+                relationships.append({
+                    "personId": row_id,
+                    "relatedPersonId": matched_id,
+                    "type": "SPOUSE_OF",
+                    "source": "verheiratet_mit",
+                })
+        else:
+            unresolved.append({
+                "rowId": row_id,
+                "field": "verheiratet_mit",
+                "raw": raw,
+                "reason": reason,
+            })
+
+    return relationships, unresolved
diff --git a/tools/import-normalizer/tests/test_persons_tree.py b/tools/import-normalizer/tests/test_persons_tree.py
index ea4a1b61..97bbc8bd 100644
--- a/tools/import-normalizer/tests/test_persons_tree.py
+++ b/tools/import-normalizer/tests/test_persons_tree.py
@@ -273,3 +273,48 @@ def test_deduplicate_both_none_birth_year_kept():
     result, skipped = persons_tree._deduplicate(persons)
     assert [p["rowId"] for p in result] == ["row_A"]
     assert len(skipped) == 1
+
+
+def _make_persons(*args):
+    """Helper: args are (rowId, firstName, lastName, maidenName, spouse_raw) tuples."""
+    return [
+        {"rowId": a[0], "firstName": a[1], "lastName": a[2], "maidenName": a[3],
+         "_spouse_raw": a[4], "_bemerkung_raw": None,
+         "birthYear": None, "deathYear": None, "birthPlace": None, "deathPlace": None,
+         "generation": None, "familyMember": True, "alias": None, "notes": None}
+        for a in args
+    ]
+
+
+def test_resolve_spouses_success():
+    persons = _make_persons(
+        ("row_002", "Elsgard", "Allemeyer", "Wöhler", "Allemeyer Werner"),
+        ("row_003", "Werner", "Allemeyer", None, "Elsgard Wöhler"),
+    )
+    idx = persons_tree._build_index(persons)
+    rels, unres = persons_tree._resolve_spouses(persons, idx)
+    assert len(rels) == 1
+    assert rels[0]["type"] == "SPOUSE_OF"
+    assert set([rels[0]["personId"], rels[0]["relatedPersonId"]]) == {"row_002", "row_003"}
+    assert unres == []
+
+
+def test_resolve_spouses_not_found():
+    persons = _make_persons(
+        ("row_007", "Charlotte", "Blomquist", "Ruge", '"Tante Lolly"'),
+    )
+    idx = persons_tree._build_index(persons)
+    rels, unres = persons_tree._resolve_spouses(persons, idx)
+    assert rels == []
+    assert len(unres) == 1
+    assert unres[0]["rowId"] == "row_007"
+    assert unres[0]["reason"] == "not_found"
+
+
+def test_resolve_spouses_empty_spouse_field():
+    persons = _make_persons(
+        ("row_004", "Jürgen", "Allemeyer", None, None),
+    )
+    idx = persons_tree._build_index(persons)
+    rels, unres = persons_tree._resolve_spouses(persons, idx)
+    assert rels == [] and unres == []