Skip to content

Releases: OpenAdaptAI/openadapt-types

v0.3.1

Choose a tag to compare

@abrichr abrichr released this 16 Jul 00:21

v0.3.1 (2026-07-16)

Bug Fixes

  • Keep release lock metadata consistent (e3c78f4)

Pin release tooling, verify editable lock metadata, and build the reviewed lock state during semantic release.


Detailed Changes: v0.3.0...v0.3.1

v0.3.0

Choose a tag to compare

@abrichr abrichr released this 13 Jul 16:34

v0.3.0 (2026-07-13)

Features

  • Add canonical Benchmark* types (Task/Observation/Action/Agent) (#5, 82ad485)

Move the Benchmark* vocabulary into the canonical schema package so both openadapt-ml and openadapt-evals can import it without depending on each other, breaking the historical ml<->evals import cycle.

Definitions are dependency-free (dataclasses + abc) and faithfully match the previous openadapt-evals definitions.

Claude-Session: https://claude.ai/code/session_01CKrVJJy5jWVCkXAqgUqtqZ

Co-authored-by: Claude Opus 4.8 noreply@anthropic.com


Detailed Changes: v0.2.0...v0.3.0

v0.2.0

Choose a tag to compare

@abrichr abrichr released this 29 Mar 18:21

v0.2.0 (2026-03-29)

Continuous Integration

  • Switch to python-semantic-release for automated versioning (#3, d5b3f3f)

Conventional commit PR titles (feat:, fix:, etc.) auto-bump version, tag, publish to PyPI, and create GitHub Releases on merge to main.

Co-authored-by: Claude Opus 4.6 noreply@anthropic.com

Features

  • Add universal action parser for DSL, JSON, and BenchmarkAction formats (#4, 0c2ac58)

Adds openadapt_types.parsing module with five public functions: - parse_action(): auto-detect format (DSL or JSON) and parse - parse_action_dsl(): parse DSL strings like CLICK(x=0.5, y=0.3) - parse_action_json(): parse JSON with canonical, flat, and coordinate formats - from_benchmark_action(): convert BenchmarkAction-style dicts to Action - to_benchmark_action_dict(): convert Action back to BenchmarkAction dict

Handles Thought:/Action: prefixes, markdown fences, coordinate normalization detection, clamping, and legacy field mapping. All edge cases return Action(type=DONE) with a logged warning instead of raising.

Includes 66 tests covering DSL, JSON, auto-detect, and BenchmarkAction round-trip conversion.

Co-authored-by: Claude Opus 4.6 (1M context) noreply@anthropic.com


Detailed Changes: v0.1.0...v0.2.0