1.3.3

What's Changed

• (docs) refine docs
• (chore) flag methods and modules as internal
• (refactor) replace fixtures with dynamic mock data
• (refactor) use default device for embeddings
• (fix) fix key mismatch
• (fix) define module within mostlyai namespace

1.3.2

What's Changed

• refactored Makefile
• simplified docs layout
• fixed issue with reading large Int64 columns
• fixed issue with binning of bool columns
• fixed issue with binning of constant datetime columns

Full Changelog: 1.3.1...1.3.2

1.3.1

What's Changed

• updated pypi package info
• move code folder one level up
• fix compatibility with pandas 2.0

Full Changelog: 1.3.0...1.3.1

1.3.0

What's Changed

• provide docs
• use rich.Progress as default progress
• new interface for progress reporting
• relax pandas dependency to 2.0

Full Changelog: 1.2.1...1.3.0

1.2.1

What's Changed

• trim metrics precision after all other validations are done
• filter out empty buckets while calculating embeddings
• makefile support for release to PyPI

Full Changelog: 1.2.0...1.2.1

1.2.0

What's Changed

• changed return type for qa.report from dict to Metrics
• added quick start example
• (internal) simplify calculate_embeddings + add fine-grained progress
• (internal) trigger CI workflow for PRs from forked repos
• (internal) change INFIX for stored statistics files

Full Changelog: 1.1.4...1.2.0

1.1.4

What's Changed

• Update README.md
• Added quick start example

Full Changelog: 1.1.3...1.1.4

1.1.3

What's Changed

• fix issue with special character in column names

Full Changelog: 1.1.2...1.1.3

mostlyai-qa 1.1.2

What's Changed

• fix issue with max_sample_size_embeddings argument
• improved README

Full Changelog: 1.1.0...1.1.2

mostlyai-qa 1.1.0

What's Changed

• add package version (#1)
• fix typo in README (#2)
• harmonize max_sample_size_embeddings (#3)
• changed defaults for max_sample_size to unlimited (#3)
• drop zero-sequence-length records in pull_data_for_accuracy (#4)

Full Changelog: 1.0.0...1.1.0