labrijisaad / llm-rag Goto Github PK

Things to be done:

• Organize the code more effectively (refactor the code, improve variable naming, remove unnecessary codes and files).
• Dockerize the application.
• Allow the possibility to mount volumes for secrets and markdown documents using Docker volume.
• Add more Make commands in the Makefile.
• Update the Readme file.

• It would be better to add the possibility to search in the Knowledge base using Faiss Similatity Search, this way the user can choose between a Normal Search or a Similatity Search 👌

• Should create the dev branch to develop the src code.

• Add CI worflow to test the code for FAISS Index Creation.
• Add tests/ folder and create tests with pytest for the workflow.
• Add OpenAI key as a repository secret and set up readme mock data for testing.
• Add other types of tests...

Add a README file to explain the main idea, and also provide explanations for the prototyping notebooks 🙌

• Should allow the user to delete specified documents in the Knowledge Base without dropping the entire Knowledge Base.
  • Update the method of saving/uploading embeddings and their corresponding files (utilize one embedding file per chunk of data).
  • Add a search mechanism that will locate the corresponding file for each saved embedding file.

• Update the Readme file.
• Organize the code more effectively (refactor the code, improve variable naming, remove unnecessary codes and files).
• Push the Docker Image in a Container Registry.
• Deploy the App in Cloud ( GCP, Azure.. )