- Add extract_preamble() to retain SEO metadata before the heading when rewriting the doc
- Add normalize_version() to map -preview suffix to -rc.1 (preview == RC1 in ABP versioning)
- Add unit tests (Test 17-19) covering preview, rc.N, and stable version normalization
Updated the AI response handling and improved raw notes cleaning process. Enhanced version mapping table extraction and file writing to preserve all content.
Updated the workflow to save release notes to an environment variable instead of a temporary file. Adjusted validation and fallback mechanisms accordingly.
Replaces git diff logic with GitHub API to reliably detect added/modified markdown files in docs/en/ from merged PRs. Improves branch creation and file processing steps, ensuring only relevant files are handled for SEO updates.
Improved modularity and readability of add_seo_descriptions.py by refactoring functions, enhancing block handling, and adding statistics saving. Updated auto-add-seo.yml to correctly detect changed markdown files using PR head and base SHAs for more accurate file selection.
Enhanced the detection and updating of SEO description blocks in add_seo_descriptions.py. The script now checks for the presence and non-emptiness of the Description field, and updates or inserts the SEO block as needed, handling invalid JSON gracefully.
Moved the SEO description generation logic from an inline script in the GitHub Actions workflow to a dedicated Python script at .github/scripts/add_seo_descriptions.py. Updated the workflow to call this script directly, improving maintainability and readability.
This workflow automatically adds SEO descriptions to modified markdown files in the 'docs/en/' directory when a pull request is merged. It checks for existing descriptions, generates new ones using OpenAI, and commits the updates in a new branch.