CHECK24 GenDev Messenger Challenge

Welcome to my submission for the 4th round of the GenDev Scholarship! 🚀 I'm excited to present my implementation of the CHECK24 Messenger Challenge, building a seamless chat experience for service providers and customers.

Overview

This messenger application allows users to communicate seamlessly after a service provider sends a quote. The challenge involves creating both the backend and frontend components, providing a robust and enjoyable user experience.

Stack

Backend: Pythons FastAPI with PostgresDB for data persistence. Frontend: React.js (Vite) for a responsive and dynamic UI.

This are the minimal functional requirements I have implemented:

✅ Chat view: the same conversation should be viewable from both perspectives: customer and service provider (they should each have their own site/route, so that hypothetically a service provider on one device is able to chat with the customer on the other device)

✅ Overview over all chats: think about implementing simple(!) chat overviews for both perspectives to click into chats, too.

✅ Persistance: chats are mutable, so they should be persisted in a database/store of your choice and accessible via an unique identifier

✅ Scrolling pagination: the customer/service provider should be able to scroll through a chat (and its history) with a hypothetical length of 10000s of messages per chat (think about Pagination)

✅ Image and PDF attachments: The customer/service provider should be able to send messages with optional attachments like pictures or documents (PDF)

✅ Masking of sensible information: Any contactable data (phone, email, URL) within all messages should be masked as long as the conversation status is "quoted"

✅ Show sent message in chat without refreshing: Messages should appear within the chat as they were sent (without refreshing the page)

✅ Hide message field if rejected state: If a conversation status is "rejected" the message input field should disappear and actions should not be possible any longer

✅ Request customer reviews: Within the chat the service provider should be able to request a review from the customer once they've worked together (feature explicitly explained within our provided datasets)

✅ A quick screencast 🎥 of your final working app within your GitHub repository (+ optional a link to a live version to check out or an explanation how to start your app on our machines, see "How do I hand in my project?")

The other ideas:

✅ Unread banner behaviour (highly appreciated)

✅ An extendable/generic message API format (consider whether your API design is also suitable for new product requirements e.g. new message types)

✅ Thumbnail processing (to improve UX)

✅ Live-Updates (Socket/Polling/...) - (in my case SSE)

❌ Start the conversation at the scroll offset of the first unread message

✅ Dockerize your system so that anyone can run it on their machine

✅ OpenAPI for backend <-> frontend communication (+ code generation)

Other implemented ideas:

✅ OAuth2-Integration

📚 Message Reactions

📚 User Search with Keywords from all users

✅ Paginated Chat view

(📚) On Database Level: Support for Thread Messages

Explanation

✅ Implemented in Client and Server

📚 Only implemented in Server

❌ Not implemented

How to Run Locally

Via docker-compose

You can run the application via docker-compose. Therefore, you need access to the docker image of the server and client on ghcr.io.

docker login ghcr.io

Navigate to the project directory and export the following environment variables:

export POSTGRES_PASSWORD=...
export SECRET_KEY=...
export TOKEN_EXPIRE_MINUTES

After that you can run:

docker-compose up

For development

Clone the repository. Navigate to the project directory.

For the server

❗️ Do this first if you do not have generated the client

Navigate to '/server' and export/put the following env variable in an .env file:

DATABASE_HOSTNAME=localhost
DATABASE_PORT=5432
DATABASE_PASSWORD=...
DATABASE_NAME=gendev_db
DATABASE_USERNAME=postgres
SECRET_KEY=...
ALGORITHM=HS256
ACCESS_TOKEN_EXPIRE_MINUTES=1000

If you have no running db you could also export the following env variable:

export SKIP_DB_CONNECTION=True

Afterwards, run the following command to generate the client:

python extract-openapi.py app.main:app

Before installing the requirements, think about creating a virtual environment.

python -m venv venv

Activate the virtual environment. Afterwards, install the requirements and run the server.

pip install -r requirements.txt

uvicorn app.main:app --reload --host 0.0.0.0

For the client

Navigate to '/client' and put the following env variable in an .env file:

VITE_PROXY_API_URL=http://0.0.0.0:8000

Afterwards, install the requirements and run the server.

yarn install

yarn run generateAPI

yarn run dev

Demo

Short drafts and a demo video can be found under my nextcloud.

Questions?

For any questions or clarifications, feel free to reach out to [email protected].

Thank you for the opportunity, and I look forward to your feedback! 😊