Skip to content

Release 0.13.0
Compare
Choose a tag to compare
@paulgc paulgc released this 02 Mar 01:32
· 885 commits to master since this release
v0.13.0
b1bf593

Major Features and Improvements

  • Use joblib to exploit multiprocessing when computing statistics over a pandas
    dataframe.
  • Add support for semantic domain related statistics (natural language, image),
    enabled by StatsOptions.enable_semantic_domain_stats.
  • Python 3.5 is supported.

Bug Fixes and Other Changes

  • Expand unit test coverage.
  • Modify validation logic to raise SCHEMA_MISSING_COLUMN anomaly when
    observing a feature with no stats.
  • Add utility functions write_stats_text and load_stats_text to write and
    load DatasetFeatureStatisticsList protos.
  • Avoid using multiprocessing by default when generating statistics over a
    dataframe.
  • Depends on joblib>=0.12,<1.
  • Depends on tensorflow-transform>=0.13,<0.14.
  • Depends on tensorflow-metadata>=0.12.1,<0.14.
  • Requires pre-installed tensorflow>=1.13.1,<2.
  • Depends on apache-beam[gcp]>=2.11,<3.
  • Depends on absl>=0.1.6,<1.

Breaking Changes

Deprecations

Assets 10
Loading