|
| 1 | +# Changelog |
| 2 | + |
| 3 | +All notable changes to Ouroboros will be documented in this file. |
| 4 | + |
| 5 | +The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), |
| 6 | +and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). |
| 7 | + |
| 8 | +## [2.0.0] - 2025-12-21 |
| 9 | + |
| 10 | +### Added - Phase 5: The Integration Loop 🎉 |
| 11 | + |
| 12 | +#### Safety Gate |
| 13 | +- Tree-Sitter syntax validator for Python, JavaScript, TypeScript |
| 14 | +- Automatic syntax validation before any code touches disk |
| 15 | +- Detailed error reporting with line numbers and context |
| 16 | +- Self-healing retry loop (up to 3 attempts with error feedback) |
| 17 | +- Configurable safety gate enable/disable |
| 18 | + |
| 19 | +#### CLI |
| 20 | +- Beautiful CLI using Typer and Rich libraries |
| 21 | +- `refactor` command - Generate and apply code patches |
| 22 | +- `status` command - Check run status and provenance |
| 23 | +- `list-runs` command - View recent generation runs |
| 24 | +- Progress bars and spinners for long operations |
| 25 | +- Rich table output for results |
| 26 | +- Dry-run mode to preview changes |
| 27 | +- Auto-apply for low-risk patches |
| 28 | +- Risk scoring (0.0-1.0) for all patches |
| 29 | +- Windows (`ouroboros.bat`) and Unix (`ouroboros.sh`) launchers |
| 30 | + |
| 31 | +#### Provenance Logging |
| 32 | +- Complete audit trail for every generation run |
| 33 | +- Model usage tracking (Reasoner, Compressor, Generator) |
| 34 | +- Safety check logging with timestamps |
| 35 | +- File modification tracking with SHA256 hashes |
| 36 | +- JSON export to `artifact_metadata.json` |
| 37 | +- Metadata includes: run_id, models_used, safety_checks, file_modifications |
| 38 | + |
| 39 | +#### Builder Enhancements |
| 40 | +- Integrated syntax validator into generation pipeline |
| 41 | +- Retry mechanism with enhanced error feedback |
| 42 | +- Validation result tracking in metadata |
| 43 | +- Configurable max retry attempts |
| 44 | + |
| 45 | +#### Pipeline Integration |
| 46 | +- Provenance logger initialized for every run |
| 47 | +- Model usage tracked across all phases |
| 48 | +- Safety checks logged after generation |
| 49 | +- File modifications tracked on apply |
| 50 | +- Error logging with failure provenance |
| 51 | + |
| 52 | +#### Documentation |
| 53 | +- `PHASE_5_COMPLETE.md` - Comprehensive Phase 5 guide |
| 54 | +- `CLI_QUICK_REFERENCE.md` - Quick CLI reference |
| 55 | +- `PHASE_5_SUMMARY.md` - Implementation summary |
| 56 | +- Updated main `README.md` with Phase 5 info |
| 57 | +- `CONTRIBUTING.md` - Contribution guidelines |
| 58 | +- `CHANGELOG.md` - This file |
| 59 | + |
| 60 | +#### Dependencies |
| 61 | +- Added `typer>=0.9.0` for CLI framework |
| 62 | +- Added `rich>=13.7.0` for beautiful terminal output |
| 63 | +- Added `click>=8.1.0` as typer dependency |
| 64 | + |
| 65 | +### Changed |
| 66 | +- Updated `.gitignore` to ignore test files and artifacts |
| 67 | +- Moved all documentation to `docs/` folder |
| 68 | +- Updated `README.md` with new project structure |
| 69 | +- Enhanced error handling across all components |
| 70 | + |
| 71 | +### Fixed |
| 72 | +- Builder now validates syntax before applying changes |
| 73 | +- Proper error context in validation failures |
| 74 | +- Provenance saved even on generation failures |
| 75 | + |
| 76 | +## [1.5.0] - 2025-12-10 |
| 77 | + |
| 78 | +### Added - Phase 4: The Builder |
| 79 | + |
| 80 | +#### Discrete Diffusion |
| 81 | +- Discrete diffusion model for code generation |
| 82 | +- Cosine noise schedule (1000 steps) |
| 83 | +- Classifier-free guidance (CFG scale 7.5) |
| 84 | +- Temperature control for generation |
| 85 | +- Multiple backbone support (Mock, GPT, Claude) |
| 86 | + |
| 87 | +#### Builder Component |
| 88 | +- High-level orchestrator for code generation |
| 89 | +- Batch patch generation |
| 90 | +- Risk scoring for patches |
| 91 | +- Unified diff generation |
| 92 | +- Provenance metadata tracking |
| 93 | + |
| 94 | +#### AST Masking |
| 95 | +- Tree-sitter based AST masking |
| 96 | +- Target-specific function/class masking |
| 97 | +- Masked span tracking |
| 98 | +- Multi-language support |
| 99 | + |
| 100 | +#### Configuration |
| 101 | +- Fast mode (10 steps) |
| 102 | +- Balanced mode (50 steps) |
| 103 | +- Quality mode (100 steps) |
| 104 | +- Mock mode for testing |
| 105 | + |
| 106 | +#### Documentation |
| 107 | +- `PHASE_4_COMPLETE.md` - Complete Phase 4 documentation |
| 108 | +- Architecture diagrams |
| 109 | +- Usage examples |
| 110 | +- API reference |
| 111 | + |
| 112 | +### Changed |
| 113 | +- Refactored generation pipeline |
| 114 | +- Enhanced metadata tracking |
| 115 | +- Improved error handling |
| 116 | + |
| 117 | +## [1.0.0] - 2025-12-08 |
| 118 | + |
| 119 | +### Added - Phase 1: The Librarian |
| 120 | + |
| 121 | +#### Knowledge Graph |
| 122 | +- Neo4j graph database integration |
| 123 | +- Tree-sitter code parsing (Python, JS, TS) |
| 124 | +- Provenance metadata tracking |
| 125 | +- Code checksum validation |
| 126 | + |
| 127 | +#### Graph Construction |
| 128 | +- File node creation |
| 129 | +- Function/class node extraction |
| 130 | +- Dependency relationship mapping |
| 131 | +- Import tracking |
| 132 | + |
| 133 | +#### Retrieval |
| 134 | +- Graph traversal queries |
| 135 | +- Context extraction |
| 136 | +- Related node discovery |
| 137 | + |
| 138 | +#### Infrastructure |
| 139 | +- Docker Compose setup for Neo4j |
| 140 | +- Database schema initialization |
| 141 | +- Ingestion pipeline |
| 142 | +- Synthetic benchmarks |
| 143 | + |
| 144 | +### Added - Phase 2: The Reasoner |
| 145 | + |
| 146 | +#### Dependency Analysis |
| 147 | +- Code dependency tracking |
| 148 | +- Impact assessment |
| 149 | +- Change propagation analysis |
| 150 | + |
| 151 | +#### LLM Integration |
| 152 | +- Claude 3.5 Sonnet support |
| 153 | +- GPT-4 support |
| 154 | +- Anthropic API client |
| 155 | +- Plan generation |
| 156 | + |
| 157 | +### Added - Phase 3: The Compressor |
| 158 | + |
| 159 | +#### Context Encoding |
| 160 | +- AI21 Jamba 1.5 Mini integration |
| 161 | +- 256k token context window |
| 162 | +- Context compression |
| 163 | +- Token optimization |
| 164 | + |
| 165 | +#### Configuration |
| 166 | +- AI21 API setup |
| 167 | +- Compression parameters |
| 168 | +- Validation pipeline |
| 169 | + |
| 170 | +### Infrastructure |
| 171 | +- Python 3.10+ support |
| 172 | +- Virtual environment setup |
| 173 | +- Requirements management |
| 174 | +- Environment variables (.env) |
| 175 | + |
| 176 | +## [Unreleased] |
| 177 | + |
| 178 | +### Planned Features |
| 179 | +- Semantic validation (beyond syntax) |
| 180 | +- Security scanning |
| 181 | +- Performance profiling |
| 182 | +- Web UI |
| 183 | +- VS Code extension |
| 184 | +- GitHub Actions integration |
| 185 | +- CI/CD pipeline |
| 186 | +- Additional language support |
| 187 | +- Improved retry strategies |
| 188 | +- Rollback automation |
| 189 | + |
| 190 | +--- |
| 191 | + |
| 192 | +## Version History |
| 193 | + |
| 194 | +- **v2.0.0** (Dec 21, 2025) - Phase 5: Integration Loop (Safety, CLI, Provenance) |
| 195 | +- **v1.5.0** (Dec 10, 2025) - Phase 4: The Builder (Discrete Diffusion) |
| 196 | +- **v1.0.0** (Dec 08, 2025) - Phase 1-3: Foundation (Graph, Reasoner, Compressor) |
| 197 | + |
| 198 | +## Contributing |
| 199 | + |
| 200 | +See [CONTRIBUTING.md](CONTRIBUTING.md) for contribution guidelines. |
| 201 | + |
| 202 | +## License |
| 203 | + |
| 204 | +MIT License - see [LICENSE](LICENSE) for details. |
0 commit comments