marcel/familienarchiv
1
0
Fork 0
Code Issues 106 Pull Requests 2 Actions 2 Packages Projects Releases Wiki Activity

feat(normalizer): unresolved-names report + fix ambiguous-pair over-flagging

Browse Source 
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
Marcel
2026-05-25 15:54:37 +02:00
parent f10b80a03f
commit 97ab9e38df
4 changed files with 44 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
tools/import-normalizer/normalize.py
View File

@@ -21,7 +21,8 @@ def run(*, document_workbook, document_sheet, person_workbook, person_sheet,
person_dicts = [{f: (row[i] if i < len(row) else "") for f, i in p_fields.items()} for row in person_rows[1:]]
register = persons.parse_register(person_dicts)
alias_index = persons.AliasIndex(register)
ctx = persons.ResolutionContext(alias_index, name_overrides)
given_names = persons.build_given_names(register, config.EXTRA_GIVEN_NAMES)
ctx = persons.ResolutionContext(alias_index, name_overrides, given_names=given_names)
# --- documents ---
doc_rows = ingest.read_sheet(document_workbook, document_sheet)
@@ -93,7 +94,15 @@ def run(*, document_workbook, document_sheet, person_workbook, person_sheet,
writers.write_review_csv(review_dir / "duplicate-index.csv", ["source_row", "index"], duplicates)
writers.write_review_csv(review_dir / "blank-index-rows.csv", ["source_row", "kind", "content"], blank_index)
writers.write_review_csv(review_dir / "skipped-x-suffix.csv", ["source_row", "index", "base_index"], skipped_x)
writers.write_review_csv(review_dir / "ambiguous-receivers.csv", ["raw", "part", "source_row"], ctx.ambiguous)
unresolved_agg: dict[tuple, list] = {}
for name, category, row in ctx.unresolved:
unresolved_agg.setdefault((category, name), []).append(row)
unresolved_rows = sorted(
([cat, name, len(rows), " ".join(map(str, sorted(rows)[:5]))]
for (cat, name), rows in unresolved_agg.items()),
key=lambda r: (r[0], -r[2], r[1]))
writers.write_review_csv(review_dir / "unresolved-names.csv",
["category", "raw", "count", "example_rows"], unresolved_rows)
writers.write_review_csv(review_dir / "index-file-mismatch.csv", ["source_row", "index", "file"], mismatches)
dated = sum(1 for d in canon_docs if d.date_raw.strip())
@@ -115,7 +124,13 @@ def run(*, document_workbook, document_sheet, person_workbook, person_sheet,
"distinct_unparsed_formats": len(unparsed_by_raw),
"# NAMES": "",
"unmatched_name_strings": len(ctx.unmatched),
"ambiguous_receivers": len(ctx.ambiguous),
"unresolved_name_occurrences": len(ctx.unresolved),
"unresolved_unknown": sum(1 for _, c, _ in ctx.unresolved if c == "unknown"),
"unresolved_single_token": sum(1 for _, c, _ in ctx.unresolved if c == "single_token"),
"unresolved_relational": sum(1 for _, c, _ in ctx.unresolved if c == "relational"),
"unresolved_collective": sum(1 for _, c, _ in ctx.unresolved if c == "collective"),
"unresolved_prose": sum(1 for _, c, _ in ctx.unresolved if c == "prose"),
"unresolved_ambiguous_pair": sum(1 for _, c, _ in ctx.unresolved if c == "ambiguous_pair"),
"# ANOMALIES": "",
"empty_rows": empty_count,
"blank_index_rows": len(blank_index),

17
tools/import-normalizer/persons.py
View File

@@ -264,12 +264,14 @@ class AliasIndex:
class ResolutionContext:
"""Resolves raw name strings to person ids; accumulates provisional persons and review data."""
def __init__(self, alias_index: AliasIndex, name_overrides: dict[str, str]):
def __init__(self, alias_index: AliasIndex, name_overrides: dict[str, str],
given_names: set[str] | None = None):
self.index = alias_index
self.name_overrides = name_overrides
self.given_names = given_names or set()
self.provisional: dict[str, Person] = {}
self.unmatched: dict[str, list] = {}
self.ambiguous: list[tuple] = []
self.unresolved: list[tuple] = [] # (raw_name, category, source_row) for non-RESOLVABLE names
self._raw_to_pid: dict[str, str] = {}
self.override_hits = 0
@@ -296,6 +298,9 @@ class ResolutionContext:
return pid, self.index.display(pid) or name, True
# provisional person (unmatched) — never reuse a register id
self.unmatched.setdefault(name, []).append(source_row)
category = classify_name(name, self.given_names)
if category is not NameClass.RESOLVABLE:
self.unresolved.append((name, str(category), source_row))
if name in self._raw_to_pid:
return self._raw_to_pid[name], name, False
last, first = _last_first(name)
@@ -315,13 +320,7 @@ class ResolutionContext:
return pid, name, matched, len(parts) > 1
def resolve_receivers(self, raw: str, source_row: int):
results = []
for part in split_receivers(raw):
pid, name, matched = self.resolve_one(part, source_row)
if not matched and " " in part and find_known_last_name(part) is None and len(part.split()) == 2:
self.ambiguous.append((raw, part, source_row))
results.append((pid, name, matched))
return results
return [self.resolve_one(part, source_row) for part in split_receivers(raw)]
def _last_first(name: str):

17
tools/import-normalizer/tests/test_documents.py
View File

@@ -93,10 +93,17 @@ def test_resolve_one_override_increments_hits():
assert name == "Eugenie de Gruyter" # display comes from the alias index
assert ctx.override_hits == 1
def test_ambiguous_space_pair_flagged_not_split():
# US-PERS-02 AC4: "Ella Anita" is kept as one provisional + flagged, never guessed into two.
ctx = _ctx()
def test_ambiguous_pair_recorded_in_unresolved():
people = persons.parse_register([{"last_name": "de Gruyter", "first_name": "Walter"}])
ctx = persons.ResolutionContext(persons.AliasIndex(people), name_overrides={},
given_names={"ella", "anita"})
raw = documents.RawRow(source_row=7, index="C-0200", sender="", receivers="Ella Anita")
doc = documents.to_canonical(raw, ctx, date_overrides={})
assert len(doc.receiver_person_ids) == 1 # not split
assert any(part == "Ella Anita" for _, part, _ in ctx.ambiguous)
assert len(doc.receiver_person_ids) == 1 # not split — one provisional
assert any(name == "Ella Anita" and cat == "ambiguous_pair" for name, cat, _ in ctx.unresolved)
def test_resolvable_first_surname_pair_not_unresolved():
ctx = persons.ResolutionContext(persons.AliasIndex([]), name_overrides={},
given_names={"ella", "anita"})
ctx.resolve_one("Mieze Schefold", source_row=1) # surname is not a given name
assert ctx.unresolved == [] # RESOLVABLE -> not recorded

7
tools/import-normalizer/tests/test_normalize.py
View File

@@ -10,7 +10,7 @@ def _doc_wb(tmp_path):
"Eugenie Müller", "15.2.1888", "Rotterdam", "Brautbriefe", "Geschäftsreise"])
ws.append(["W-0001x", r"..\__scan\W-0001x.pdf", "", "", "Walter de Gruyter", "Eugenie Müller", "", "", "", ""])
ws.append(["", "", "", "", "Section banner row", "", "", "", "", ""])
ws.append(["C-0001", "", "", "", "Hans Wittkopf", "", "Freitag 1919", "", "", ""])
ws.append(["C-0001", "", "", "", "Hans Wittkopf", "?", "Freitag 1919", "", "", ""])
ws.append(["W-0001", r"..\__scan\W-0001.pdf", "V", "1", "Walter de Gruyter",
"Eugenie Müller", "15.2.1888", "Rotterdam", "Brautbriefe", "dup"])
p = tmp_path / "docs.xlsx"; wb.save(p); return p
@@ -42,6 +42,11 @@ def test_run_end_to_end(tmp_path):
assert (review_dir / "unparsed-dates.csv").exists()
# C-0001's "Freitag 1919" is unparseable -> must appear in the review file (NFR-DATA-01)
assert "Freitag 1919" in (review_dir / "unparsed-dates.csv").read_text(encoding="utf-8")
assert (out_dir / "canonical-documents.xlsx").exists() # (keep existing asserts above)
assert (review_dir / "unresolved-names.csv").exists()
unresolved_text = (review_dir / "unresolved-names.csv").read_text(encoding="utf-8")
assert "unknown" in unresolved_text and "?" in unresolved_text # the "?" receiver
assert not (review_dir / "ambiguous-receivers.csv").exists() # replaced
# determinism (NFR-IDEM-01): a second run yields identical canonical content + review files
def _matrix(p):