fra 2.3.5

Simple Recommendation System, content based.

FRA

Fra is a Recommendation based framework. It was created with the idea of have an option for creating content-based recommendations easily via API.

Architecture

Requirements

• pyenv (optional but recommended)
• poetry
• python 3.9+
• postgres
• redis
• docker (optional)

Installation

• create a virtualenv

  pyenv virtualenv 3.9.15 fra pyenv activate fra

• create a database and update the config.py file with the name and host

  psql; create database fra; \q

• install the requirements

  poetry install

• migrate the models

  flask --app server/app db upgrade

Running the app

• activate the environment

  pyenv activate fra

• run the app

  flask --app server/app run

-- For async jobs that process recommendations

> celery -A motor.tasks worker  --loglevel=info  

Running the app with docker

• Run the docker-compose

  docker-compose up

That command will start all the services and the web server can be accessed via browser on http://localhost:5000

OpenAPI docs will be online on YOURDOMAIN/api/swagger/ui

Follow this order to be able to get recommendations:

1. create an organization
2. add users to the organization
3. add the files with the data (some examples on example_data folder)
4. add file mapping
5. add user ratings
6. get recommendations

TODO

• Build and publish on GCP
• Add more data set to the examples