Releases: scrapinghub/dateparser
Releases · scrapinghub/dateparser
Release 1.0.0
Breaking changes:
- Drop support for Python 2.7 and pypy (see #727, #744, #748, #749, #754, #755, #758, #761, #763, #764, #777 and #783)
- Now
DateDataParser.get_date_data()
returns aDateData
object instead of adict
(see #778). - From now wrong
settings
are not silenced and raiseSettingValidationError
(see #797) - Now
dateparser.parse()
is deterministic and doesn't try previous locales. Also,DateDataParser.get_date_data()
doesn't try the previous locales by default (see #781) - Remove the
'base-formats'
parser (see #721) - Extract the
'no-spaces-time'
parser from the'absolute-time'
parser and make it an optional parser (see #786) - Remove
numeral_translation_data
(see #782) - Remove the undocumented
SKIP_TOKENS_PARSER
andFUZZY
settings (see #728, #794) - Remove support for using strings in
date_formats
(see #726) - The undocumented
ExactLanguageSearch
class has been moved to the private scope and some internal methods have changed (see #778) - Changes in
dateparser.utils
:normalize_unicode()
doesn't acceptbytes
as input andconvert_to_unicode
has been deprecated (see #749)
New features:
- Add Python 3.9 support (see #732, #823)
- Detect hours separated with a period/dot (see #741)
- Add support for "decade" (see #762)
- Add support for the hijri calendar in Python ≥ 3.6 (see #718)
Improvements:
- New logo! (see #719)
- Improve the README and docs (see #779, #722)
- Fix the "calendars" extra (see #740)
- Fix leap years when
PREFER_DATES_FROM
is set (see #738) - Fix
STRICT_PARSING
setting inno-spaces-time
parser (see #715) - Consider
RETURN_AS_TIME_PERIOD
setting forrelative-time
parser (see #807) - Parse the 24hr time format with meridian info (see #634)
- Other small improvements (see #698, #709, #710, #712, #730, #731, #735, #739, #784, #788, #795 and #801)
Release 0.7.6
Improvements:
- Rename
scripts
todateparser_scripts
to avoid name collisions with modules from other packages or projects (see #707)
Release 0.7.5
New features:
- Add Python 3.8 support (see #664)
- Implement a
REQUIRE_PARTS
setting (see #703) - Add support for subscript and superscript numbers (see #684)
- Extended French support (see #672)
- Extended German support (see #673)
Improvements:
- Migrate test suite to Pytest (see #662)
- Add test to check the
yaml
andjson
files content (see #663 and #692) - Add flake8 pipeline with pytest-flake8 (see #665)
- Add partial support for 8-digit dates without separators (see #639)
- Fix possible
OverflowError
errors and explicitly avoid to raiseValueError
when parsing relative dates (see #686) - Fix double-digit GMT and UTC parsing (see #632)
- Fix bug when using
DATE_ORDER
(see #628) - Fix bug when parsing relative time with timezone (see #503)
- Fix milliseconds parsing (see #572 and #661)
- Fix wrong values to be interpreted as
'future'
inPREFER_DATES_FROM
(see #629) - Other small improvements (see #667, #675, #511, #626, #512, #509, #696, #702 and #699)
Release 0.7.4
New features
Improvements
- Add support for PREFER_DATES_FROM in relative/freshness parser (see #414)
- Add support for PREFER_DAY_OF_MONTH in base-formats parser (see #611)
- Added UTC -00:00 as a valid offset (see #574)
- Fix support for “one” (see #593)
- Fix TypeError when parsing some invalid dates (see #536)
- Fix tokenizer for non recognized characters (see #622)
- Prevent installing regex 2019.02.19 (see #600)
- Resolve DeprecationWarning related to raw string escape sequences (see #596)
- Implement a tox environment to build the documentation (see #604)
- Improve tests stability (see #591, #605)
- Documentation improvements (see #510, #578, #619, #614, #620)
- Performance improvements (see #570, #569, #625)
Release 0.7.3
Broken version
Release 0.7.2
Features:
- Extended Czech support
- Added
time
to valid periods - Added timezone information to dates found with
search_dates()
- Support strings as date formats
Improvements:
- Fixed Collections ABCs depreciation warning
- Fixed dates with trailing colons not being parsed
- Fixed date format override on any settings change
- Fixed parsing current weekday as past date, regardless of settings
- Added UTC -2:30 as a valid offset
- Added Python 3.7 to supported versions, dropped support for Python 3.3 and 3.4
- Moved to
importlib
fromimp
where possible - Improved support for Catalan
- Documentation improvements
Release 0.7.1
Features/news:
- Added detected language to return value of
search_dates()
- Performance improvements
- Refreshed versions of dependencies
Improvements:
- Fixed unpickleable
DateTime
objects with timezones - Fixed regex pattern to avoid new behaviour of
re.split()
in Python 3.7 - Fixed an exception thrown when parsing colons
- Fixed tests failing on days with number greater than 30
- Fixed
ZeroDivisionError
exceptions
Release 0.7.0
Features added during Google Summer of Code 2017:
- Harvesting language data from Unicode CLDR database (https://github.com/unicode-cldr/cldr-json), which includes over 200 locales (#321) - authored by Sarthak Maddan.
See full currently supported locale list in README. - Extracting dates from longer strings of text (#324) - authored by Elena Zakharova.
Special thanks for their awesome contributions!
New features:
Improvements:
- Improved support of Chinese (#359), Thai (#345), French (#301, #304), Russian (#302)
- Removed ruamel.yaml from dependencies (#374). This should reduce the number of installation issues and improve performance as the result of moving away from YAML as basic data storage format.
Note that YAML is still used as format for support language files. - Improved performance through using pre-compiling frequent regexes and lazy loading of data (#293, #294, #295, #315)
- Extended tests (#316, #317, #318, #323)
- Updated nose_parameterized to its current package, parameterized (#381)
Planned for next release:
- Full language and locale names
- Performance and stability improvements
- Documentation improvements
Release 0.6.0
New features:
- Consistent parsing in terms of true python representation of date string. See #281
- Added support for Bangla, Bulgarian and Hindi languages.
Improvements:
- Major bug fixes related to parser and system's locale. See #277, #282
- Type check for timezone arguments in settings. see #267
- Pinned dependencies' versions in requirements. See #265
- Improved support for cn, es, dutch languages. See #274, #272, #285
Packaging:
- Make calendars extras to be used at the time of installation if need to use calendars feature.