Changelog¶

All notable changes to the AI Documentation Agent will be documented in this file.

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

[Unreleased]¶

Planned¶

Response caching for faster subsequent runs
Support for more programming languages
Plugin system for custom extensions
Web UI for easier usage
API server mode
Integration with popular IDEs

[2.0.0] - 2025-01-26¶

Added¶

Iterative Refinement: AI agent now critiques and refines its own output
Self-Critique System: Automatic quality assessment and improvement
Smart File Prioritization: Intelligent file ordering based on importance
Project Type Detection: Auto-detect frontend, backend, or mixed projects
Multi-Format Output: Support for Markdown, HTML, and PDF
Comprehensive Logging: Detailed logging with file and console output
Retry Logic: Exponential backoff for API failures
Configuration System: Environment-based configuration via .env
Quality Thresholds: Configurable acceptance criteria
Docker Support: Containerized deployment option
PyInstaller Support: Standalone executable builds
Organized Structure: Clean project organization with src/, config/, docs/
Complete Documentation: Comprehensive guides and API reference

Changed¶

Breaking: Restructured project layout (moved to src/ directory)
Breaking: Updated configuration to use .env file
Improved error handling and user feedback
Enhanced documentation generation quality
Better handling of large codebases
Optimized file discovery algorithm

Fixed¶

Timeout issues with large projects
Memory leaks in long-running sessions
Unicode handling in documentation
PDF generation on Windows
API connection error messages

Removed¶

Direct hardcoded configuration (moved to .env)
Deprecated single-pass generator as primary mode

[1.5.0] - 2024-12-15¶

Added¶

HTML output format support
PDF generation capability
Basic project type detection
Priority file identification
Configurable file limits

Changed¶

Improved prompt engineering
Better code extraction
Enhanced error messages

Fixed¶

File encoding issues
Empty documentation output
API timeout handling

[1.0.0] - 2024-11-01¶

Added¶

Initial release
Basic documentation generation
Ollama API integration
Markdown output support
Simple file discovery
Command-line interface
Basic configuration options

Features¶

Generate documentation from Python, JavaScript, TypeScript files
Support for common web frameworks
Automatic file discovery
Simple command-line usage
Configurable output location

Version Numbering¶

This project uses Semantic Versioning:

MAJOR version: Incompatible API changes
MINOR version: New features (backward compatible)
PATCH version: Bug fixes (backward compatible)

Example¶

2.0.0 → Major release with breaking changes
2.1.0 → New features added
2.1.1 → Bug fixes only

Release Process¶

1. Pre-Release¶

2. Release¶

# Tag version
git tag -a v2.0.0 -m "Release version 2.0.0"

# Push tag
git push origin v2.0.0

# Create GitHub release
# - Go to GitHub Releases
# - Create new release from tag
# - Add release notes from CHANGELOG

3. Post-Release¶

Announce on project channels
Update documentation site
Monitor for issues
Plan next version

Detailed Changes by Version¶

[2.0.0] - Detailed Changes¶

New Files¶

src/ai_agent.py - AI agent with iterative refinement
config/.env.example - Configuration template
build/Dockerfile - Docker containerization
build/docker-compose.yml - Docker Compose setup
build/ai_agent.spec - PyInstaller specification
docs/ - Complete documentation structure

Modified Files¶

src/doc_generator.py - Enhanced with prioritization
run.py - Updated launcher
setup.py - New package configuration
README.md - Comprehensive updates

Architecture Changes¶

Old Structure:          New Structure:
├── ai_agent.py        ├── src/
├── doc_generator.py   │   ├── ai_agent.py
├── requirements.txt   │   └── doc_generator.py
└── README.md          ├── config/
                       │   ├── requirements.txt
                       │   └── .env.example
                       ├── build/
                       ├── docs/
                       ├── tests/
                       └── run.py

API Changes¶

# Old (v1.x)
generate_documentation(directory, output_format)

# New (v2.x)
agent = AIAgent(
    directory=directory,
    model=model,
    output_format=output_format,
    config=config
)
agent.run(max_iterations=3)

Configuration Changes¶

# Old (v1.x)
# Hardcoded in script

# New (v2.x)
# .env file
OLLAMA_API_URL=http://localhost:11434/api/generate
MODEL_NAME=llama2:7b
API_TIMEOUT=300

Migration Guide¶

Upgrading from 1.x to 2.0¶

1. Update Project Structure¶

# Backup old version
cp -r old-version old-version-backup

# Clone new version
git clone https://github.com/your-repo/ai-doc-agent.git
cd ai-doc-agent

2. Update Configuration¶

# Create .env file
cp config/.env.example .env

# Transfer your old settings
# Old: MODEL = "llama2:7b"
# New: MODEL_NAME=llama2:7b in .env

3. Update Scripts¶

# Old usage
python ai_agent.py /path/to/project

# New usage
python run.py --directory /path/to/project

4. Update Imports (if using as library)¶

# Old
from ai_agent import generate_documentation

# New
from src.ai_agent import AIAgent
from src.doc_generator import generate_documentation

Breaking Changes Log¶

v2.0.0¶

Configuration: - Environment variables now required in .env file - MODEL renamed to MODEL_NAME - OLLAMA_URL renamed to OLLAMA_API_URL

Project Structure: - Source files moved to src/ directory - Configuration moved to config/ directory - Import paths changed

CLI: - New command-line argument names - --model flag added for model override - --iterations flag added for refinement control

API: - AIAgent class signature changed - Configuration now uses AgentConfig class - Return values updated (exit codes)

Deprecation Notices¶

Deprecated in v2.0¶

Direct script execution (use run.py instead)
Hardcoded configuration (use .env)
Single-pass mode as primary (use AI Agent)

Removed in v2.0¶

Old directory structure
Direct configuration in scripts
Legacy command-line arguments

Security Updates¶

v2.0.0¶

Added .env to .gitignore to prevent secret leaks
Improved input validation
Sanitized file path handling
Updated dependencies to latest versions

Known Issues¶

Current Issues¶

v2.0.0: - PDF generation may fail on some Windows systems without wkhtmltopdf - Very large projects (500+ files) may timeout - Response caching not yet implemented - Some antivirus software flags PyInstaller executables

Workarounds: - Install wkhtmltopdf separately for PDF support - Use --max-files to limit file count - Disable caching (already disabled by default) - Add executable to antivirus exceptions

Contributing to Changelog¶

When contributing, please update CHANGELOG.md:

Add changes to [Unreleased] section
Use these categories:
Added - New features
Changed - Changes to existing features
Deprecated - Soon-to-be removed features
Removed - Removed features
Fixed - Bug fixes
Security - Security updates

Format:

### Added
- Feature description with context
- Another feature

Next Steps¶

Contributing Guide - How to contribute
Testing Guide - Testing instructions
API Reference - API documentation
User Guide - Usage guide

Note: This changelog is maintained manually. For detailed commit history, see GitHub commits.