This project serves as a core engine for communicating with various Large Language Model (LLM) backends, with a primary focus on integration with Sublime Text. Currently, network communication is implemented for the OpenAI gateway, with plans to support OpenAI-alike models and Claude.
Core of OpenAI-sublime-text package.
- Functionality: This core is designed to interface with LLMs, manage cache entries, handle network requests, and process streaming data.
- Integrations: Primarily targeted at Sublime Text, with support for Python bindings via PyO3 to facilitate usage in Python environments.
- Rust: Ensure
cargoandrustcare installed.
-
Clone the repository:
git clone [email protected]:yaroslavyaroslav/llm_runner.git cd llm_runner
-
Build the project:
Note
python 3.8
maturin build --release- Install the it as dependency for ST (==under python 3.8 venv==):
Note
python 3.8
pip install target/wheels/llm_runner-0.2.0-cp38-cp38-macosx_11_0_arm64.whl --target '/path/to/Sublime Text/Lib/python38/' --upgrade- Integrate with Sublime Text using the provided Python classes or directly using Rust API.
- Configure and use in your own extensions or projects that require LLM interaction.
- Assistant Settings: Modify settings in
AssistantSettingsstruct for your specific LLM configurations and preferences. - Cache Handling: Manage cache directory and file paths as per your application's needs.
All contributions are welcome! To get started with development:
- Fork the repository.
- Make your changes in a new branch.
- Create a pull request when ready.
Run tests with:
cargo testUtilize the extensive test suite included for validation of different components, including network client and cache handling.
- Additional LLM backend support.
- Enhance cache management.
- Expand Python bindings for broader application compatibility.
This project is licensed under custom license. See the LICENSE file for more details.