Decision support api

Description

The Decision Support API module is designed for a startup acceleration platform, assisting in ranking help measures for the current project using sBERT embeddings and cosine distance.

Main features:

• Asynchronous API with swagger documentation
• Caching with Redis
• Interaction with the database using SQLalchemy and async driver aiomysql
• Migrations using Alembic
• Module for converting data from a database into text with ORM
• Module for text preprocessing
• Module for text ranking using sBERT embedding and cosine distance
• Autotests for API and database with pytest (with full database creation)
• Logging with Python JSON logger
• Sentry integration
• Monitoring with Grafana
• Linters with GitHub CI
• Fully containerized application

Preparing environment and docker

• Change .env data to your development database configuration

• Change .env.prod data to your production database configuration

• Replace docker-compose db service with a real database

Running server

Using Docker compose

• Run services

  docker-compose up

• Bring services down

  docker-compose down

go to the local host address: http://127.0.0.1:8000/

Using python virtual machine

1. Create a Python virtual environment and activate it.

2. Go to the folder with main.py

   cd backend

3. Install requirements:

   python3 -m pip install --user -r requirements.txt

4. Run pytest

   pytest

5. Run server using gunicorn:

   uvicorn main:app --reload

go to the local host address: http://127.0.0.1:8000/