|
1 | | -# deep-next |
| 1 | +# DeepNext |
| 2 | + |
| 3 | +Your AI-powered junior software engineering assistant that takes an issue description and converts it into a pull request. |
| 4 | + |
| 5 | +DeepNext is an advanced AI system that automatically analyzes a code repository, prepares a solution plan, and implements the necessary code for software engineering tasks using Large Language Models, saving developers hours of repetitive work. |
| 6 | + |
| 7 | +See [documentation](stxnext.github.io/deep-next) for more information. |
| 8 | + |
| 9 | +## Quick start |
| 10 | + |
| 11 | +### Installation |
| 12 | + |
| 13 | +```bash |
| 14 | +# Clone the repository |
| 15 | +git clone git@github.com:stxnext/deep-next.git |
| 16 | +cd deep-next |
| 17 | + |
| 18 | +# Install dependencies |
| 19 | +make install_venv |
| 20 | + |
| 21 | +# Test the installation |
| 22 | +make test_unit |
| 23 | +``` |
| 24 | + |
| 25 | +### Example usage |
| 26 | + |
| 27 | +```bash |
| 28 | +# Process a single issue |
| 29 | +poetry run python -m deep_next.core.entrypoint \ |
| 30 | + --problem-statement "Add type hints in file.py" \ |
| 31 | + --hints "The error occurs in file.py" \ |
| 32 | + --root-dir /path/to/repository |
| 33 | +``` |
| 34 | + |
| 35 | +## GitHub/GitLab integration |
| 36 | + |
| 37 | +Running as a service to automatically process issues: |
| 38 | + |
| 39 | +[See integration details](https://stxnext.github.io/deep-next/integration.html) |
| 40 | + |
| 41 | +## Configuration |
| 42 | + |
| 43 | +DeepNext supports multiple LLM providers: |
| 44 | +- OpenAI |
| 45 | +- AWS Bedrock (Claude, Mistral, and others) |
| 46 | + |
| 47 | +Create an `llm-config.yaml` file based on the example provided to configure model preferences for each pipeline stage. |
| 48 | + |
| 49 | +For tracking and metrics, DeepNext integrates with LangSmith. Set up your credentials in the `.env` file. |
| 50 | + |
| 51 | +[See configuration details](https://stxnext.github.io/deep-next/configuration.html) |
| 52 | + |
| 53 | +## Roadmap |
| 54 | + |
| 55 | +- **May 2025**: First public release |
| 56 | +- Future enhancements: |
| 57 | + - Improved code generation accuracy |
| 58 | + - Support for more programming languages |
| 59 | + |
| 60 | +## Why choose DeepNext? |
| 61 | + |
| 62 | +- **End-to-End Automation**: Complete pipeline from issue to merge request |
| 63 | +- **Multiple LLM Support**: Works with various models to fit your needs |
| 64 | +- **Flexible Integration**: Compatible with GitHub, GitLab and AWS |
| 65 | +- **Customizable**: Configure different models for different pipeline stages |
| 66 | +- **Modular**: Easy to modify, test, and improve by swapping small pieces of code |
| 67 | +- **Open Source**: MIT-licensed and community-driven |
0 commit comments