Skip to content

Changelog

All notable changes to the Documentation Automation Examples project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Added

  • Changelog page added to the doc site (docs/changelog.md); uses the snippets extension to render CHANGELOG.md directly so content stays in one place
  • Project 03: README Generator (03-readme-generator/) — new automation project that analyzes a local directory or GitHub repository and generates a complete README section by section using AI
  • readme_generator.py: main script with repo analysis (language detection, dependency parsing, entry point detection, license detection), section-by-section AI generation, and Jinja2 assembly
  • prompts/: four prompt files for description, installation, usage, and contributing sections; each accepts {REPO_CONTEXT} placeholder
  • examples/sample_project/: built-in TaskFlow CLI app for --sample mode (no API key needed for testing)
  • templates/readme_template.md.j2: default Jinja2 output template
  • CLI flags: --input, --output, --config, --prompts, --template, --sections, --analyze, --sample
  • tests/test_readme_generator.py: 39 tests covering language detection, project name detection, license detection, dependency parsing, context summary building, sample mode, and README assembly
  • Full tutorial (5 steps), reference docs (configuration, prompt engineering), sample outputs, troubleshooting, and FAQ added to the doc site under docs/readme-generator/
  • config.example.yaml: added readme_max_file_lines and readme_sections fields
  • mkdocs.yml: README Generator nav section added after Doc Site Portfolio

Changed

  • Tutorial steps 1, 3, 4, and 5 updated to add Path C: Use Claude Code CLI as a third AI workflow option alongside the existing script-based paths
  • Step 1: Added Path C to "Choose your path", added a "which path is right for you?" guidance paragraph, added full Path C setup section (install, authenticate, verify), added skip note in "Configure AI access" for Path C users, and added jump-ahead links on all three path names
  • Step 3: Added "Option C" blocks for generating nav structure, homepage, and about page via Claude Code CLI
  • Step 4: Added "Option C" block for generating project pages, plus a callout explaining how Path C users can direct Claude to write output to a file
  • Step 5: Added "Option C" block for content refinement, plus a callout explaining how to reference file paths in prompts

[1.1.0] - 2026-04-02

Added

Release notes automation (01-release-notes-automation)

  • GitLab integration via python-gitlab — GitLabProvider class with support for self-hosted instances via gitlab_url config
  • Bitbucket integration via atlassian-python-api — BitbucketProvider class with app password authentication
  • Azure OpenAI provider — AzureOpenAIProvider class using the existing openai library's AzureOpenAI client; requires azure_endpoint and azure_api_version config fields
  • Hybrid categorization pipeline — three-stage strategy: conventional commit prefix parsing → keyword rules → AI fallback; AI only processes commits not matched by earlier stages
  • categorization_rules.example.yaml — example keyword rules file with pre-built rule sets for Security, Performance, Database, Infrastructure, and Dependencies categories
  • Batch processing — --repos CLI argument and batch_repos config field; --batch-output controls separate files per repo or a single combined report
  • Custom Jinja2 output templates — --template CLI argument accepts any .j2 file; templates/default.md.j2 ships as the built-in default and preserves existing output format
  • vcs_provider config field and VCS_PROVIDER environment variable for selecting GitHub, GitLab, or Bitbucket

Changed

  • generate_release_notes() signature updated: now accepts categorized commit data (List[Dict]) instead of a raw AI response string, enabling template rendering
  • load_config() no longer requires github_token as a universal required field; VCS token validation moved to each provider class to support multiple VCS sources
  • config.example.yaml updated with sections for all new providers and batch processing options; model recommendation updated to claude-sonnet-4-6
  • conftest.py updated with stubs for gitlab and atlassian so tests run without those packages installed
  • Test suite expanded from 13 to 47 tests, adding TestParseConventionalCommit, TestMatchKeywordRules, TestParseAiCategories, TestCategorizeCommitsHybrid, TestTemplateRendering, and TestBatchProcessing

[1.0.0] - 2026-04-02

Added

Doc site portfolio project (02-doc-site-portfolio) - Implemented generate_content.py — generates About, Home, project, and site structure pages from a writer profile using Anthropic or OpenAI - Implemented improve_content.py — refines existing Markdown content for clarity, tone, or consistency using AI - Added profile.example.yaml — template writers copy and fill in to drive both content generation scripts - Wrote all three prompt templates: about_page_prompt.txt, project_page_prompt.txt, site_structure_prompt.txt (previously placeholders) - Both scripts support --sample mode for full testing without API calls

Tests - Added tests/ directory with pytest test suite (39 tests) - tests/test_release_notes.py — covers sample data, release notes formatting, and fetch with sample flag - tests/test_generate_content.py — covers profile loading, profile/project formatting, and output wrapping - tests/conftest.py — stubs optional dependencies so tests run without installing anthropic, openai, or PyGithub

Release notes automation project (01-release-notes-automation) - Initial project setup with working release notes generation script - GitHub API integration using PyGithub - Support for Anthropic Claude and OpenAI GPT providers - Sample data mode for testing without API calls - Prompt engineering examples showing evolution from 65% to 91% accuracy - Workflow worksheet for documenting manual processes - Configuration management with YAML and environment variable overrides - Comprehensive MkDocs documentation site - Troubleshooting guide with 40+ common issues - Comprehensive FAQ with 50+ questions - Contributing guidelines and Code of Conduct

Changed

  • Doc site portfolio tutorial steps 1–5 updated to reference the new scripts alongside the existing manual prompting approach
  • Tutorial step 1 corrected: AI setup now uses config.yaml (matching the actual scripts) instead of a .env file with wrong variable names
  • Tutorial step 2 now includes a profile setup section for profile.yaml

Documentation

  • Complete tutorial covering setup through iteration (release notes)
  • Doc site tutorial steps 1–6 covering setup through deployment
  • Reference documentation for API integration, prompt engineering, site structure, theme customization, content types, and AI collaboration
  • Sample outputs demonstrating categorization quality
  • Getting started guide for first-time users
  • Portfolio examples with five profile patterns

Release Note Categories

This changelog uses the following categories:

  • Added - New features or functionality
  • Changed - Changes to existing functionality
  • Deprecated - Features that will be removed in future versions
  • Removed - Features that have been removed
  • Fixed - Bug fixes
  • Security - Security vulnerability fixes
  • Documentation - Documentation improvements

Future Releases (Planned)

Version 1.2.0 (Planned)

  • Web-based configuration UI
  • Scheduled automation support
  • Integration with CI/CD pipelines
  • Multiple output format support (JSON, HTML)
  • Internationalization (i18n) support

Versioning Strategy

This project follows Semantic Versioning:

  • MAJOR version (X.0.0): Incompatible API changes or major rewrites
  • MINOR version (1.X.0): New features, backward-compatible
  • PATCH version (1.0.X): Bug fixes, backward-compatible

Contributing

See CONTRIBUTING.md for how to contribute changes to this project.

Questions?