Releases: CaptorAB/openseries
Releases · CaptorAB/openseries
1.8.0
- Adjustments to meet demands of new Poetry version 2.0.1, among them format changes to pyproject.toml
- Set strict=True in all uses of zip()
- Set arguments in console example and changed position of code sample into badge at top in README.md
- Some widening of test in test_package.py
- Refactored folder name types -> owntypes to adhere to A005 error in ruff updated to ver 0.9.3. This is a POTENTIALLY BREAKING CHANGE.
1.7.8
- adjustments to make typing work with new version of numpy
- Removed strict type DatabaseIdStringType on database identifiers and replaced with regular string
- checks that project still work with updated dependencies
- some tooling updates in github scripts
1.7.7
- Checked and project works with Python 3.13. Project meta data changed to this effect.
- Development updates only otherwise. E.g. adapting to changes in ruff, pandas-stubs etc.
1.7.6
- Changed worst_month() to be consistent with resample_to_business_period_ends().
- Updated Poetry to version 1.8.4.
- Miscellaneous dependency and lockfile updates.
1.7.5
- Added pandas .ffill() as precursor to all remaining pct_change() to suppress pandas FutureWarnings in dependent projects.
1.7.4
- Tightened existing checks to not allow mixed series types as methods input. The coverage level of the checks has not been widened.
- Added pandas .ffill() as precursor to pct_change() to suppress pandas FutureWarnings in dependent projects.
- Fixed method .resample_to_business_period_ends() so it considers renamed labels.
- Corrected warning in this changelog for release of version 1.7.0. Added 'NO'.
- Miscellaneous dependency and lockfile updates.
1.7.3
- Consolidated all_properties() method and its string validations.
- Simplified pandas.pct_change() after new pandas type stubs and fill_method no longer required by mypy.
- Simplified dependencies after bug in statsmodels resolved in version 0.14.3.
- Miscellaneous dependency and lockfile updates
1.7.2
- Replaced the dot operator from numpy and pandas with the Python @ operator.
- Made OpenTimeSeries Pydantic validator methods private.
- Adjustments to adhere to mypy unreachable code warnings.
- Removed table in security.md to limit maintenance.
- Removed setup_class from readme.md as it was deleted in version 1.7.1.
- Cosmetic improvements on raise exception statements.
- Limited dependency scipy version <1.14.1 due to bug in statsmodels
- Miscellaneous dependency updates
1.7.1
Commit with full changes is this one 10292b9
- Changed so that resample_to_business_period_ends method on OpenFrame now retains original stubs on all constituent OpenTimeSeries.
- Made do_resample_to_business_period_ends private.
- Cleaned up calc_range and efficient_frontier functions.
- Improved calc_range test and exception message.
- Removed classmethod setup_class on OpenTimeSeries. Replaced with Pydantic field_validator decorator for domestic (currency) field and countries field.
- Made input type more strict on OpenTimeSeries.from_df().
- Introduced branch coverage and added tests to bring coverage back to 100% from 99%.
- Miscellaneous cleanup and depedencies updated.
1.7.0
- Changed code to enforce PEP 604 on typing. This means that the PACKAGE WILL NO LONGER WORK FOR PYTHON 3.9.
- Limited GitHub workflow build.yaml to no longer run on Python 3.9
- Adjustments to adhere to ruff TCH type checking imports.
- Introduced strict requirement that generate_calendar_date_range argument trading_days must be positive.
- Further simplified common simulation for CommonTestCase class.
- Removed unnecessary .resolve() from pathlib.Path and simplified Path.open()
- Made all Path.open() statements follow the same syntax.
- Cleaned up test setups and added tearDownClass method on common testcase class.
- Removed year from copyright notice in license.md
- Suppressed pycharm type checker on OpenFrame class.
- Miscellaneous dependency updates