Implement a functionality using FastAPI to periodically provide clients with briefings on the results of an AI Inference Pipeline. Enable interaction between the client and the server through request and response mechanisms.

Please refer to the attached "requirements.txt" for the versions of packages and libraries needed for environment installation.

pip install fastapi
pip install "uvicorn[standard]"

pip install websockets

pip install apscheduler

pip install pymongo

Install MongoDB following the instructions on the

official MongoDB website

pip install python-dotenv

Create a separate file named .env.production and add the following content:

DATABASE_ID="your_id"
DATABASE_PASSWORD="your_pw"
DATABASE_URI=mongodb+srv:https://<DATABASE_ID>:<DATABASE_PASSWORD>@"your_dbname".keizkxa.mongodb.net

Ensure that you keep this file secure and do not share sensitive information.

This file contains configuration details for your production environment, including database credentials. Keep in mind the importance of securing this information and not exposing it publicly.

• login

  • description

    • Utilize check_user in models.py to verify new and existing users.

      • In the case of an existing user, return {"message": "signup", "user_id": <user_id>}.
      • For a new user, return {"message": "login", "user_id": <user_id>}.

• websocket_endpoint

  • Description:
    • Handle WebSocket connections for continuous updates.
    • Establish WebSocket connection.
    • Store new request in the database and register with the scheduler.
    • Continuously send updates to the frontend through the WebSocket.

Parameters

• websocket_reloadBriefing

  • Description
    • Display the ongoing briefing.
    • Establish WebSocket connection.
    • Send reloaded briefing data to the frontend.
    • Get the latest request ID and continuously watch the database for updates.

Parameters

• endBriefing

  • description

    • Terminate the ongoing briefing.

    • The function endBriefing uses the endBriefing_worker from utils.py and end_scheduler from scheduler.py.

      • Upon success, it prints a success message and returns a success response with the message "Briefing successfully terminated".
      • In case of failure (e.g., an exception), it catches the "HTTPException", logs the error message, and returns an error response with the corresponding status code and details.

• getAllRequest

  • description
    • Implement a functionality to retrieve all requests for a specific username from the "request" Collection within the database.
    • getAllRequest utilizes the get_all_request function from utils.py to retrieve and return the list of requests for a specific username from the "request" Collection in the database.

• getBriefing

  • description
    • Retrieve briefing information based on the provided request_id.
    • Utilize the get_briefing function from utils.py to perform a functionality that retrieves all briefings for a specific request_id from the "briefing" Collection in the database.

• aiSummary

  • description

    • Generates an AI summary captions based on the provided request_id.

    • This API utilizes two methods: get_briefing in utils.py for querying all data from the database and chat_summary in utils.py for generating AI summaries based on the returned briefing data list.

    • To provide additional details on chat_summary, it involves using the OpenAI API key to generate messages based on a specified format prompt. This process utilizes the GPT-4-preview model and performs prompt engineering on the briefing data list retrieved earlier, adjusting hyperparameters such as "tone, writing style, temperature, top_p," and others.

• getCustom

  • description
    • this API retrieves the "custom" data for a specific username. The obtained custom_list is then used to generate an HTTP response in JSON format, encoded in UTF-8.
    • The get_custom in models.py process involves searching for the specified username within the "user" collection in the database and referring to it as a custom list.

• updateCustom

  • description
    • Utilizing the update_custom function in models.py, this API updates the "custom" field in the "user" table.
      • If the custom_list is successfully updated, it returns the message "Success to update Custom."
      • In case of exceptions or errors, it raises an HTTPException with a status code of 500 and the detail "Fail to update Custom."