This project provides an implementation of a chatbot that interacts with various GPT-based models through a FastAPI interface and a script for command-line interactions.

1. Clone the repository: Clone this repository using git clone.
2. Create Virtual Env: Create a Python Virtual Environment venv to download the required dependencies and libraries.
3. Download Dependencies: Download the required dependencies into the Virtual Environment venv using pip.

git clone https://github.com/grisha765/chatbot_py.git
cd chatbot_py
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt

You can either run the FastAPI server or interact with the chatbot directly from the command line.

• To start the interactive mode, use the following command:

  python -m chatbot --model gpt-3.5-duck

• To start the FastAPI server, use the following command:

  python -m chatbot --api

This will start the server on 127.0.0.1:8000.

To interact with the chatbot directly from the command line, use:

• One message:

  python -m chatbot --model gpt-3.5-duck "Your message here"

• To save a screenshot of the interaction, use:

  python -m chatbot --model gpt-3.5-duck "Your message here" --screenshot "path/to/screenshot.png"

• Install with pip

  pip install git+https://github.com/grisha765/chatbot_py.git@main#egg=chatbot

• Use as python lib.

  # import libs
  import asyncio, chatbot
  
  # use async func
  async def main():
      # init model class
      model = chatbot.Model()
      # set options
      await model.options(model='gpt-3.5-duck', chrome_path='/usr/bin/chromium')
      # send message to chatbot without saving message history
      response = await model.send("Hello, world!", temp=False)
      # print response
      print(response)
      # close model
      await model.close()
  
  if __name__ == "__main__":
      # run async func
      asyncio.run(main())

• Request:

  {
    "model": "gpt-3.5-duck",
    "messages": [
      {"role": "user", "content": "Hello, GPT-3.5!"}
    ],
    "temperature": 0.5
  }

• Response:

  {
      "id": "chatcmpl-12345",
      "object": "chat.completion",
      "created": 1677631234,
      "model": "model",
      "choices": [
          {
              "index": 0,
              "message": {
                  "role": "assistant",
                  "content": "response"
              },
              "finish_reason": "stop"
          }
      ]
  }

• Headers:

• Authorization: Bearer <API_KEY>

• Curl request:

  curl http://127.0.0.1:8000/v1/chat/completions \
           -H "Content-Type: application/json" \
           -H "Authorization: Bearer adminapi007" \
           -d '{
          "model": "gpt-3.5-duck",
          "messages": [{"role": "user", "content": "Hello World!"}],
          "temperature": 0.7
        }'

• Use in python:

  import requests # pip install requests
  import json
  
  text = 'Hello World!'
  url = "http://127.0.0.1:8000/v1/chat/completions"
  headers = {
      "Content-Type": "application/json",
      "Authorization": "Bearer adminapi007"
  }
  data = {
      "model": "gpt-3.5-duck",
      "messages": [{"role": "user", "content": f"{text}"}],
      "temperature": 0.7
  }
  
  response = requests.post(url, headers=headers, data=json.dumps(data))
  
  if response.status_code == 200:
      response_data = response.json()
      message = response_data['choices'][0]['message']['content']
      print("Message:", message)
  else:
      print("Failed to get response:", response.status_code)

• Browser Automation: Utilizes pyppeteer to automate browser interactions, simulating user actions to chat with the GPT model.
• Interactive Mode: Provides an interactive mode for chatting directly via the command line, with special commands for taking screenshots.
• API Key Authentication: Secure API endpoint with key-based authentication for authorized access.
• Error Handling: Robust error handling to manage and log exceptions during execution.
• Configurable Settings: Easily adjustable settings for logging details and API configurations.

Contributions are welcome! Please fork the repository and submit pull requests.