diff -pruN 1.2.1-1/CONTRIBUTING.rst 1.2.2-1/CONTRIBUTING.rst --- 1.2.1-1/CONTRIBUTING.rst 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/CONTRIBUTING.rst 2025-06-26 09:26:59.000000000 +0000 @@ -179,3 +179,16 @@ Whenever the content of the corresponding documentation table:: dateparser_scripts/update_supported_languages_and_locales.py + + +Updating the Timezone Cache +---------------------------------------------------- + +Whenever the content of +``dateparser/timezones.py`` is modified you need to rebuild the timezone cache. + +Run this command: +``BUILD_TZ_CACHE=1 python -c "import dateparser"`` + +which should update +``dateparser/data/dateparser_tz_cache.pkl`` diff -pruN 1.2.1-1/HISTORY.rst 1.2.2-1/HISTORY.rst --- 1.2.1-1/HISTORY.rst 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/HISTORY.rst 2025-06-26 09:26:59.000000000 +0000 @@ -3,6 +3,19 @@ History ======= +1.2.2 (2025-06-26) +------------------ + +Fixes: + +- Handle the Russian preposition “с” (#1261) +- Fix weekday search (#1274) + +Improvements: + +- Add Python 3.14 support (#1273) +- Cache timezone offsets to improve import time (#1250) + 1.2.1 (2025-02-05) ------------------ diff -pruN 1.2.1-1/MANIFEST.in 1.2.2-1/MANIFEST.in --- 1.2.1-1/MANIFEST.in 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/MANIFEST.in 2025-06-26 09:26:59.000000000 +0000 @@ -3,6 +3,7 @@ include CONTRIBUTING.rst include HISTORY.rst include LICENSE include README.rst +include dateparser/data/dateparser_tz_cache.pkl include dateparser_data/settings.py include requirements.txt diff -pruN 1.2.1-1/dateparser/__init__.py 1.2.2-1/dateparser/__init__.py --- 1.2.1-1/dateparser/__init__.py 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/dateparser/__init__.py 2025-06-26 09:26:59.000000000 +0000 @@ -1,4 +1,4 @@ -__version__ = "1.2.1" +__version__ = "1.2.2" from .conf import apply_settings from .date import DateDataParser Binary files 1.2.1-1/dateparser/data/dateparser_tz_cache.pkl and 1.2.2-1/dateparser/data/dateparser_tz_cache.pkl differ diff -pruN 1.2.1-1/dateparser/parser.py 1.2.2-1/dateparser/parser.py --- 1.2.1-1/dateparser/parser.py 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/dateparser/parser.py 2025-06-26 09:26:59.000000000 +0000 @@ -399,8 +399,8 @@ class _parser: return datetime(**params) except ValueError as e: error_text = e.__str__() - error_msgs = ["day is out of range", "day must be in"] - if error_msgs[0] in error_text or error_msgs[1] in error_text: + error_msgs = ["day is out of range", "day must be in", "must be in range"] + if any(msg in error_text for msg in error_msgs): if not (self._token_day or hasattr(self, "_token_weekday")): # if day is not available put last day of the month params["day"] = get_last_day_of_month( @@ -512,6 +512,10 @@ class _parser: dateobj = dateobj + delta + # set the token_month here so that it is not subsequently + # altered by _correct_for_month + self._token_month = dateobj.month + # NOTE: If this assert fires, self.now needs to be made offset-aware in a similar # way that dateobj is temporarily made offset-aware. assert not (self.now.tzinfo is None and dateobj.tzinfo is not None), ( diff -pruN 1.2.1-1/dateparser/search/__init__.py 1.2.2-1/dateparser/search/__init__.py --- 1.2.1-1/dateparser/search/__init__.py 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/dateparser/search/__init__.py 2025-06-26 09:26:59.000000000 +0000 @@ -57,6 +57,8 @@ def search_dates( ('on May 6th 2004', datetime.datetime(2004, 5, 6, 0, 0))] """ + text = _search_with_detection.preprocess_text(text, languages) + result = _search_with_detection.search_dates( text=text, languages=languages, diff -pruN 1.2.1-1/dateparser/search/search.py 1.2.2-1/dateparser/search/search.py --- 1.2.1-1/dateparser/search/search.py 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/dateparser/search/search.py 2025-06-26 09:26:59.000000000 +0000 @@ -295,3 +295,10 @@ class DateSearchWithDetection: language_shortname, text, settings=settings ), } + + def preprocess_text(self, text, languages): + """Preprocess text to handle language-specific quirks.""" + if languages and "ru" in languages: + # Replace "с" (from) before numbers with a placeholder + text = re.sub(r"\bс\s+(?=\d)", "[FROM] ", text) + return text diff -pruN 1.2.1-1/dateparser/timezone_parser.py 1.2.2-1/dateparser/timezone_parser.py --- 1.2.1-1/dateparser/timezone_parser.py 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/dateparser/timezone_parser.py 2025-06-26 09:26:59.000000000 +0000 @@ -1,4 +1,8 @@ +import os +import pickle +import zlib from datetime import datetime, timedelta, timezone, tzinfo +from pathlib import Path import regex as re @@ -84,8 +88,47 @@ def get_local_tz_offset(): return offset -_search_regex_parts = [] -_tz_offsets = list(build_tz_offsets(_search_regex_parts)) -_search_regex = re.compile("|".join(_search_regex_parts)) -_search_regex_ignorecase = re.compile("|".join(_search_regex_parts), re.IGNORECASE) local_tz_offset = get_local_tz_offset() + +_tz_offsets = None +_search_regex = None +_search_regex_ignorecase = None + + +def _load_offsets(cache_path, current_hash): + global _tz_offsets, _search_regex, _search_regex_ignorecase + + try: + with open(cache_path, mode="rb") as file: + ( + serialized_hash, + _tz_offsets, + _search_regex, + _search_regex_ignorecase, + ) = pickle.load(file) + if current_hash is None or current_hash == serialized_hash: + return + except (FileNotFoundError, ValueError, TypeError): + pass + + _search_regex_parts = [] + _tz_offsets = list(build_tz_offsets(_search_regex_parts)) + _search_regex = re.compile("|".join(_search_regex_parts)) + _search_regex_ignorecase = re.compile("|".join(_search_regex_parts), re.IGNORECASE) + + with open(cache_path, mode="wb") as file: + pickle.dump( + (current_hash, _tz_offsets, _search_regex, _search_regex_ignorecase), + file, + protocol=5, + ) + + +CACHE_PATH = Path(__file__).parent.joinpath("data", "dateparser_tz_cache.pkl") + +if "BUILD_TZ_CACHE" in os.environ: + current_hash = zlib.crc32(str(timezone_info_list).encode("utf-8")) +else: + current_hash = None + +_load_offsets(CACHE_PATH, current_hash) diff -pruN 1.2.1-1/debian/changelog 1.2.2-1/debian/changelog --- 1.2.1-1/debian/changelog 2025-04-14 16:30:31.000000000 +0000 +++ 1.2.2-1/debian/changelog 2025-12-05 06:30:36.000000000 +0000 @@ -1,3 +1,18 @@ +dateparser (1.2.2-1) unstable; urgency=medium + + * Team Upload + * New upstream version 1.2.2 + * Bump Standards-Version to 4.7.2 + * Drop "Rules-Requires-Root: no": it is the default now + * Replace transitional python3-tz dependency with python3-pytz + * Add debian/salsa-ci.yml + * Rewrite d/watch in v5 format + + [ Chris Lamb ] + * make the build reproducible (Closes: #1120352) + + -- Alexandre Detiste Fri, 05 Dec 2025 07:30:36 +0100 + dateparser (1.2.1-1) unstable; urgency=medium * new upstream release diff -pruN 1.2.1-1/debian/control 1.2.2-1/debian/control --- 1.2.1-1/debian/control 2025-04-14 16:30:31.000000000 +0000 +++ 1.2.2-1/debian/control 2025-12-05 06:30:36.000000000 +0000 @@ -16,18 +16,17 @@ Build-Depends: debhelper-compat (= 13), python3-parameterized , python3-parsel , python3-pytest , + python3-pytz , python3-requests , python3-ruamel.yaml , python3-regex , python3-setuptools, - python3-tz , python3-tzlocal , tzdata-legacy , -Standards-Version: 4.6.2 +Standards-Version: 4.7.2 Homepage: https://github.com/scrapinghub/dateparser Vcs-Git: https://salsa.debian.org/python-team/packages/dateparser.git Vcs-Browser: https://salsa.debian.org/python-team/packages/dateparser -Rules-Requires-Root: no Package: python3-dateparser Architecture: all diff -pruN 1.2.1-1/debian/rules 1.2.2-1/debian/rules --- 1.2.1-1/debian/rules 2025-04-14 16:30:31.000000000 +0000 +++ 1.2.2-1/debian/rules 2025-12-05 06:30:36.000000000 +0000 @@ -5,6 +5,7 @@ export PYBUILD_NAME=dateparser export PYBUILD_BEFORE_TEST=cp -r {dir}/dateparser_data/ {build_dir} export PYBUILD_TEST_ARGS=tests \ -k 'not test_dateparser_data_integrity and not test_custom_language_detect_fast_text' +export PYBUILD_AFTER_TEST=rm -rf {build_dir}/dateparser_data/ %: dh $@ --buildsystem=pybuild diff -pruN 1.2.1-1/debian/salsa-ci.yml 1.2.2-1/debian/salsa-ci.yml --- 1.2.1-1/debian/salsa-ci.yml 1970-01-01 00:00:00.000000000 +0000 +++ 1.2.2-1/debian/salsa-ci.yml 2025-12-05 06:30:36.000000000 +0000 @@ -0,0 +1,2 @@ +include: + - https://salsa.debian.org/salsa-ci-team/pipeline/raw/master/recipes/debian.yml diff -pruN 1.2.1-1/debian/tests/run-unittests 1.2.2-1/debian/tests/run-unittests --- 1.2.1-1/debian/tests/run-unittests 2025-04-14 16:30:31.000000000 +0000 +++ 1.2.2-1/debian/tests/run-unittests 2025-12-05 06:30:36.000000000 +0000 @@ -9,6 +9,7 @@ cp -a dateparser_scripts tests $AUTOPKGT ln -s /usr/lib/python3/dist-packages/dateparser_data $AUTOPKGTEST_TMP cd $AUTOPKGTEST_TMP +rm -vf ./tests/test_dateparser_data_integrity.py for py3vers in $(py3versions -s); do echo "Testing with $py3vers:" diff -pruN 1.2.1-1/debian/watch 1.2.2-1/debian/watch --- 1.2.1-1/debian/watch 2025-04-14 16:30:31.000000000 +0000 +++ 1.2.2-1/debian/watch 2025-12-05 06:30:36.000000000 +0000 @@ -1,4 +1,5 @@ -# Compulsory line, this is a version 4 file -version=4 +Version: 5 -https://github.com/scrapinghub/dateparser/tags .*/archive/refs/tags/v?((?:\d+\.?)*)\.tar\.gz +Template: GitHub +Owner: scrapinghub +Project: dateparser diff -pruN 1.2.1-1/pyproject.toml 1.2.2-1/pyproject.toml --- 1.2.1-1/pyproject.toml 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/pyproject.toml 2025-06-26 09:26:59.000000000 +0000 @@ -1,5 +1,5 @@ [tool.bumpversion] -current_version = "1.2.1" +current_version = "1.2.2" commit = true tag = true diff -pruN 1.2.1-1/setup.py 1.2.2-1/setup.py --- 1.2.1-1/setup.py 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/setup.py 2025-06-26 09:26:59.000000000 +0000 @@ -31,7 +31,7 @@ setup( install_requires=[ "python-dateutil>=2.7.0", "pytz>=2024.2", - "regex>=2015.06.24,!=2019.02.19,!=2021.8.27", + "regex>=2024.9.11", "tzlocal>=0.2", ], entry_points={ diff -pruN 1.2.1-1/tests/test_date_parser.py 1.2.2-1/tests/test_date_parser.py --- 1.2.1-1/tests/test_date_parser.py 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/tests/test_date_parser.py 2025-06-26 09:26:59.000000000 +0000 @@ -630,6 +630,25 @@ class TestDateParser(BaseTestCase): @parameterized.expand( [ + param("Monday", datetime(2015, 3, 2)), + ] + ) + def test_preferably_future_dates_relative_last_week_of_month( + self, date_string, expected + ): + self.given_local_tz_offset(0) + self.given_parser( + settings={ + "PREFER_DATES_FROM": "future", + "RELATIVE_BASE": datetime(2015, 2, 24, 15, 30), + } + ) + self.when_date_is_parsed(date_string) + self.then_date_was_parsed_by_date_parser() + self.then_date_obj_exactly_is(expected) + + @parameterized.expand( + [ param("10 December", datetime(2015, 12, 10)), param("March", datetime(2015, 3, 15)), param("Friday", datetime(2015, 2, 13)), @@ -875,7 +894,7 @@ class TestDateParser(BaseTestCase): ): self.when_date_is_parsed_by_date_parser(date_string) self.then_error_was_raised( - ValueError, ["day is out of range for month", message] + ValueError, ["day is out of range for month", "must be in range", message] ) @parameterized.expand( diff -pruN 1.2.1-1/tests/test_search.py 1.2.2-1/tests/test_search.py --- 1.2.1-1/tests/test_search.py 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/tests/test_search.py 2025-06-26 09:26:59.000000000 +0000 @@ -410,7 +410,7 @@ class TestTranslateSearch(BaseTestCase): "Die UdSSR blieb gemäß dem Neutralitätspakt " "vom 13. April 1941 gegenüber Japan vorerst neutral.", [ - ("Die", datetime.datetime(1999, 1, 28, 0, 0)), + ("Die", datetime.datetime(1999, 12, 28, 0, 0)), ("13. April 1941", datetime.datetime(1941, 4, 13, 0, 0)), ], settings={"RELATIVE_BASE": datetime.datetime(2000, 1, 1)}, @@ -1090,3 +1090,16 @@ class TestTranslateSearch(BaseTestCase): text=text, languages=languages, error_type=ValueError ) self.check_error_message("Unknown language(s): 'unknown language code'") + + def test_search_dates_with_prepositions(self): + """Test `search_dates` for parsing Russian date ranges with prepositions and language detection.""" + result = search_dates( + "Сервис будет недоступен с 12 января по 30 апреля.", + add_detected_language=True, + languages=["ru"], + ) + expected = [ + ("12 января", datetime.datetime(2025, 1, 12, 0, 0), "ru"), + ("30 апреля", datetime.datetime(2025, 4, 30, 0, 0), "ru"), + ] + assert result == expected diff -pruN 1.2.1-1/tox.ini 1.2.2-1/tox.ini --- 1.2.1-1/tox.ini 2025-02-05 11:57:37.000000000 +0000 +++ 1.2.2-1/tox.ini 2025-06-26 09:26:59.000000000 +0000 @@ -27,7 +27,7 @@ deps = {[testenv]deps} python-dateutil==2.7.0 pytz==2024.2 - regex==2015.06.24 + regex==2024.9.11 tzlocal==0.2 [testenv:min-all] @@ -36,10 +36,11 @@ extras = {[testenv:all]extras} deps = {[testenv:min]deps} convertdate==2.2.1 - fasttext==0.9.1 hijridate==2.3.0 langdetect==1.0.0 numpy==1.19.3 + # fasttext excluded due to + # https://github.com/facebookresearch/fastText/issues/512 [testenv:scripts] deps =