Use Milvus in DocsGPT

DocsGPT is an advanced open-source solution that simplifies finding information in project documentation by integrating powerful GPT models. It enables developers to get accurate answers to their questions about a project easily, eliminating time-consuming manual searches.

In this tutorial, we will show you how to use Milvus as the backend vector database for DocsGPT.

Requirements

Make sure you have Docker installed

Clone the repository

Clone the repository and navigate to it:

$ git clone https://github.com/arc53/DocsGPT.git
$ cd DocsGPT

Add dependency

Append langchain-milvus dependency to the requirements.txt file under the application folder:

$ echo "\nlangchain-milvus==0.1.6" >> ./application/requirements.txt

Set environment variables

Add VECTOR_STORE=milvus, MILVUS_URI=..., MILVUS_TOKEN=... to the environment variables for both the backend and worker services in the docker-compose.yaml file, just like this:

  backend:
    build: ./application
    environment:
      - VECTOR_STORE=milvus
      - MILVUS_URI=...
      - MILVUS_TOKEN=...

  worker:
    build: ./application
    command: celery -A application.app.celery worker -l INFO -B
    environment:
      - VECTOR_STORE=milvus
      - MILVUS_URI=...
      - MILVUS_TOKEN=...

For the MILVUS_URI and MILVUS_TOKEN, you can either use fully managed Zilliz Cloud(Recommended) service or manually started Milvus service.

• For fully managed Zillz Cloud service: We recommend using Zilliz Cloud service. You can sign up for a free trial account on Zilliz Cloud. After that, you will get the MILVUS_URI and MILVUS_TOKEN, which correspond to the Public Endpoint and API key.

• For manually started Milvus service: If you want to set up a Milvus service, you can follow the official Milvus documentation to set up a Milvus server, and then get the MILVUS_URI and MILVUS_TOKEN from the server. The MILVUS_URI and MILVUS_TOKEN should be in the format of http://<your_server_ip>:19530 and <your_username>:<your_password> respectively.

Start the services

Run: ./setup.sh

Then navigate to http://localhost:5173/.

You can play around with the UI and ask questions about your documents.

alt text

If you want to stop the services, run:

$ docker compose down

For further details and more advanced setups, please refer to the DocsGPT official documentation.