Repository navigation
search_dates() silently drops a French date when STRICT_PARSING=True and multiple languages are passed #1326
Description
"""
Bug: search_dates() silently drops a French date when STRICT_PARSING=True and multiple languages are passed,
if the string ends with a bare uppercase token that is a valid ISO 639-1 language code (e.g. "FR").
dateparser version: 1.4.0
Python version: 3.13
"""
import dateparser
import dateparser.search
print("dateparser version:", dateparser.__version__)
TEXT = "Date de facture 23 juillet 2020 Condition Redevable livraison FR"
# Works — single language, no ambiguity for the language detector
r1 = dateparser.search.search_dates(TEXT, languages=["fr"], settings={"STRICT_PARSING": True})
# Fails — same string, same settings, but 7 languages passed
r2 = dateparser.search.search_dates(TEXT, languages=["en", "fr", "es", "pt", "de", "it", "ar"], settings={"STRICT_PARSING": True})
print(f"languages=['fr'], STRICT_PARSING=True → {r1}") # Expected: [('23 juillet 2020', datetime(...))]
print(f"languages=[7 langs], STRICT_PARSING=True → {r2}") # Bug: None
assert r1 is not None, "Baseline broken"
assert r2 is not None, f"BUG: date not found when multiple languages passed — got {r2}"Note that it works well when we remove FR in this input TEXT
It works also if we replace FR with EN or US or Frn
Activity