Hi everyone,
I created an open-source semantic code search tool using OpenAI's embedding API. This tool indexes your project and enables fast searching of relevant functions through natural language queries. Currently, I have added support for Python projects, and working to add other languages too. You can perform all your queries through a simple web-based UI.
It stores the embeddings in a local Postgres database and uses FAISS to perform nearest neighbor searching for a given query. You just need to enter the path to your project and it will look for all the code files, and index all the functions inside those. Once done, you can then perform natural language queries to get your results.
I would love to hear your feedback.
Link: https://github.com/rahuldan/codesearch
Thanks
Semantic Code Search tool. Query your codebases using natural language - GitHub - rahuldan/codesearch: Semantic Code Search tool. Query your codebases using natural language