Skip to content

v1.2.0
Compare
Choose a tag to compare
@shouples shouples released this 21 Aug 21:43
v1.2.0
90ba268
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

What's Changed

1.2.0

2022-08-21

Added

  • pd.Series as a default renderable type (to go with the existing pd.DataFrame and np.ndarray types)
  • Support for the following data types inside pd.DataFrame columns:
    • type and np.dtype
    • shapely.geometry objects
    • pd.Timedelta and datetime.timedelta
    • pd.Period
    • pd.Interval
    • complex numbers
    • ipaddress.IPv4Address and .IPv6Address
  • Extra dataset generation functions for development/testing under dx.utils.datatypes
    • dx.quick_random_dataframe(n_rows, n_columns) to get a pd.DataFrame of 0.0-1.0 floats (convenience wrapper for pd.DataFrame(np.random.rand(n_rows, n_columns)))
    • dx.random_dataframe() with different boolean values to enable based on available datatypes (dx.DX_DATATYPES)
  • settings_context context manager to allow temporarily changing a setting (or multiple)
  • Logging via structlog (default level: logging.WARNING)

Changed

  • Default sampling method changed from outer to random

Fixed

  • Displaying a dataframe with an out-of-order index (like with .sample()) no longer resets the index before sending data to the frontend.
  • Index/column flattening and string-formatting is behind settings and is handled more explicitly
  • dx should no longer interfere with other media type / mime bundles (e.g. matplotlib) formatted by the existing IPython display formatted
  • 1.1.3 updates by @shouples in #19
  • fix github workflow for publishing to pypi by @shouples in #15
  • display ID registration prep, formatter registration fixes, index updates, dtype handlers, expanded data generators by @shouples in #16

Full Changelog: v1.1.2...v1.2.0

Contributors

shouples
Assets 2
Loading