shoya140 / chatgpt-llamaindex-demo

chatgpt-llamaindex-demo

Getting Started

1. Add OPENAI_API_KEY to environment variables

   # .env.local
   OPENAI_API_KEY=...
   

2. Install dependencies

   $ pip install -r backend/requirements.txt
   $ yarn install

3. Start the server.

   $ yarn build && yarn start

Build Your Own Index

0. (Optional) Copy wikipedia dump files into 1-wikipedia and extract documents containing a keyword.

   .
   ├── README.md
   └── backend
       └── 1-wikipedia
           └── AA
               ├── wiki_00
               ├── wiki_01
               └── wiki_02 ...
   
   $ python backend/extract.py [keyword]

1. Create embeddings by using an OpenAI API.

   $ python backend/index.py [keyword]

2. Add a new option for using the index in src/page.tsx.

   <option value="gpt-3.5-turbo">GPT-3.5 Turbo</option>
   <option value="gpt-4">GPT-4</option>
   <option value="llama-keyword">Llama Keyword</option>
   ...

3. Start the server.

   $ yarn build && yarn start