/Chat-Bot-Backend

Primary LanguagePython

Chat Bot Backend

Description

A chatbot backend that provides some services.

Usage

Here are some ways to use it:

  1. From Docker image.
  2. From source code.

From docker image

Pull the image

docker pull {IMAGE_URL}

From source code

Clone the repository

git clone {REPOSITORY_URL}

Install dependencies

pip install -r requirements.txt

Prepare Redis

Please install and run Redis server by yourself.

Prepare LLM

Currently, only OpenAI is supported.

Please prepare the OpenAI API key by yourself and config it in the following .env file.

Prepare the vector database

If you want to use the RAG model, you need to prepare the vector database.

Currently, only Chroma databases is supported.

Here are some ways to use it:

  1. From the database file.
  2. From the corpus.

From the database file

Please config VECTOR_DATABASE_PATH in the following .env file.

From the corpus

Please config CORPUS_PATH in the following .env file.

Then vectorize the corpus:

from src.vector import vectorize_corpus

vectorize_corpus(corpus_path={CORPUS_PATH})

Prepare .env file

Here is the template:

# OpenAI
OPENAI_API_KEY=
# Corpus
CORPUS_PATH=
# Vector database
VECTOR_DATABASE_PATH=
# Redis
REDIS_URL=
# Server
HOST=
PORT=

Run the server

Here are some ways to run it:

  1. From Docker image.
  2. From source code.

From docker image

Run Docker Compose

Here is an example of docker-compose.yaml:

services:
  backend:
    image: chat-bot-backend:latest
    ports:
      - "8000:8000"
    volumes:
      - {VECTOR_DATABASE_PATH}:/app/database
      - {.ENV_PATH}:/app/.env

From source code

Run the server

Run script/run.sh.