Add data change transforms

[palewire]

Summary

• Add DataChange and DataChangeList models for individual value corrections stored in metadata.data.changes from Datawrapper's Check & Describe step.
• Wire the new public API through Transform.changes and package exports, while omitting empty changes from default serialization.
• Document usage and add regression coverage for both documented list-shaped changes and API object-map changes.

Verified configuration shape

• Datawrapper's developer documentation represents corrections under metadata.data.changes as a list of objects with zero-based row and column, replacement value, edit time in Unix milliseconds, and optional previous.
• Existing repository sample API payloads also show metadata.data.changes as an object keyed by Datawrapper internal IDs. Those entries include the same core fields plus API-only fields such as id, ignored, and _index; some responses use an empty object {} for no corrections.
• This implementation accepts both shapes and round-trips object-shaped API responses without converting them to list-shaped payloads.

Compatibility behavior

• Existing transformations are unchanged: Transform() still does not emit an empty changes key.
• Existing column-format, column-order, headers, transpose, external data, CDN, and upload-method serialization behavior is preserved.
• User-authored corrections serialize as the documented list shape; API object-map responses preserve their original keys and fields.

Tests

• uv run ruff check ./datawrapper ./tests
• uv run ruff format --check ./datawrapper ./tests
• uv run mypy ./datawrapper --ignore-missing-imports
• uv run pytest (1114 passed, 3 skipped)
• uv build --sdist --wheel

Closes #498.

[Copilot]

Pull request overview

Adds first-class support for Datawrapper “Check & Describe” value corrections stored under metadata.data.changes, exposing them through the public chart transformation API while preserving existing serialization behavior (not emitting empty changes by default).

Changes:

• Introduces DataChange and DataChangeList models and wires them into Transform.changes, including round-tripping both list-shaped and object-map API payloads.
• Exports the new models from datawrapper, datawrapper.charts, and datawrapper.charts.models.
• Adds documentation plus unit-fixture regression tests for both documented (list) and observed (object-map / {}) API shapes.

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
datawrapper/charts/models/transforms.py	Adds DataChange / DataChangeList and integrates them into Transform.changes with shape-preserving serialization.
datawrapper/charts/models/__init__.py	Re-exports DataChange and DataChangeList from the models package.
datawrapper/charts/__init__.py	Surfaces DataChange and DataChangeList in the charts-level public API exports.
datawrapper/__init__.py	Exposes DataChange and DataChangeList at the top-level datawrapper import.
docs/user-guide/api/models.rst	Documents DataChange / DataChangeList usage and the accepted/round-tripped API shapes.
tests/unit/models/test_transform.py	Adds unit coverage for serialization, optional fields, shape handling, empty {} handling, and object-map round-tripping.
tests/fixtures/metadata_data_changes.json	Adds a documented-shape fixture for round-trip validation through Transform.
tests/integration/test_scatter_chart.py	Minor formatting adjustment to a list comprehension in an existing integration test.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.