feat(normalizer): add persons_tree skeleton + year extraction

This commit is contained in:
Marcel
2026-05-25 20:41:25 +02:00
parent b37fd1728b
commit 9ae1196d1c
2 changed files with 100 additions and 0 deletions

View File

@@ -0,0 +1,54 @@
import sys
from pathlib import Path
sys.path.insert(0, str(Path(__file__).parent.parent))
import persons_tree
def test_parse_year_iso_string():
assert persons_tree._parse_year("1920-09-20") == 1920
def test_parse_year_excel_serial_birth():
# 7568 days from 1899-12-30 = 1920-09-19 or -20 depending on leap counting
assert persons_tree._parse_year("7568") == 1920
def test_parse_year_excel_serial_death():
# 36222 days from 1899-12-30 ≈ 1999
assert persons_tree._parse_year("36222") == 1999
def test_parse_year_excel_serial_small():
# 177 days from 1899-12-30 = 1900-06-25
assert persons_tree._parse_year("177") == 1900
def test_parse_year_german_date_string():
assert persons_tree._parse_year("30.8.1862") == 1862
def test_parse_year_year_only():
assert persons_tree._parse_year("1930") == 1930
def test_parse_year_free_text():
assert persons_tree._parse_year("August 1941") == 1941
def test_parse_year_none():
assert persons_tree._parse_year(None) is None
def test_parse_year_empty():
assert persons_tree._parse_year("") is None
def test_parse_year_unresolvable_truncated():
# "2.9.196" has no valid 4-digit year — returns None
assert persons_tree._parse_year("2.9.196") is None
def test_parse_year_typo_year():
# "4.3.1023" — year 1023 outside 1500-2100 guard — returns None
assert persons_tree._parse_year("4.3.1023") is None