/llama_index_starter_pack

This repository provides very basic flask, streamlit, and docker examples for the llama_index (fka gpt_index) package

Primary LanguagePythonMIT LicenseMIT

llama_index_starter_pack

This repository provides very basic flask, Streamlit, and docker examples for the llama_index (FKA gpt_index) package.

If you need to quickly create a POC to impress your boss, start here!

If you are having trouble with dependencies, I dump my entire env into requirements_full.txt, but otherwise, use the base requirements.txt.

The basic demo includes the classic "Paul Graham Essay" from the original llama_index repo. Some good starting questions are

  • What did the author do growing up?
  • Tell me more about interleaf

Local Setup

conda create --name llama_index python=3.11
pip install -r requirements.txt

What is included?

There are two main example folders

  • flask_react (runs three services on localhost:5601/5602/3000)
    • sh launch_app.sh
    • creates a simple api that loads the text from the documents folder (if any), also launches the react frontend
    • the "/query" endpoint accepts requests that contain a "text" parameter, which is used to query the index
    • the "/upload" endpoint is a POST endpoint that inserts an attached text file into the index
    • the index is managed by a seperate server using locks, since inserting a document is a mutable operation and flask is multithreaded
    • I strongly recommend using a tool like Postman to test the api - there are example screenshots using postman in the postman_examples folder
    • react_frontend
      • A basic frontend using react, which calls the flask_api to insert, view, and query a GPTSimpleVectorIndex
      • The scope for a frontend is basically limitless - this is intended to give a small taste of what's possible for users less familiar with python

react_frontend screenshot

  • streamlit_vector (runs on localhost:8501)

    • streamlit run streamlit_demo.py
    • creates a simple UI using streamlit and GPTSimpleVectorIndex
    • loads text from the documents folder (using st.cache_resource, so it only loads once)
    • provides an input text-box and a button to run the query
    • the string response is displayed after it finishes
    • want to see this example in action? Check it out here
  • streamlit_sql_sandbox (runs on localhost:8501)

    • streamlit run streamlit_demo.py
    • creates a streamlit app using a local SQL database about restaurant inspections in San Francisco (data sample)
    • The "Setup" tab allows you to configure various LLM and LLama Index settings
    • The "Llama Index" tab demos some basic Text2SQL capabilities using only Llama Index
    • The "Langchain+Llama Index" tab uses a custom langchain agent, and uses the SQL index from Llama Index as a tool during conversations.
    • Check out the huggingface space here!
  • streamlit_term_definition (runs on localhost:8501)

    • streamlit run streamlit_demo.py
    • creates a small app that allows users to extract terms/definitions from documents and query against the extracted information
    • pre-loaded with information from the NYC Wikipedia page
    • supports reading text from image uploads
    • allows users to configure LLM settings
    • users can build their own knowledge base of terms/definitions
    • query against these terms as they are added

Docker

Each example contains a Dockerfile. You can run docker build -t my_tag_name . to build a python3.11-slim docker image inside your desired folder. It ends up being about 600MB-900MB depending on the example.

Inside the Dockerfile, certain ports are exposed based on which ports the examples need.

When running the image, be sure to include the -p option to access the proper ports (8501, 5601, or 3000).

Contributing

I welcome any suggestions or PRs, or more examples!