divya-ai / hybrid-recommender-api

Hybrid recommender API

Recommender API for Hybrid recommender APP

Installation

First of all, install virtual environment on your machine. Then, create virtual environment (more details in this guide).

virtualenv venv

Then, activate virtual environment

source venv/bin/activate

Getting required packages

pip install -r requirements.txt

Configuration

For project configuration, you can use environment variables on your machine, or you can change configuration of project directly in app.py.

db_user = env('DB_USER', default='root')
db_password = env('DB_PASSWORD', default='')
db_name = env('DB_NAME', default='recommender')
db_host = env('DB_HOST', default='localhost')
db_port = env('DB_PORT', default=3306)
db_dialect = env('DB_DIALECT', default='mysql')

Also, you need to install and configure MongoDB. Configuration for MongoDB is in app.py too.

app.config["MONGO_URI"] = "mongodb://localhost:27017/recommender"

Database setup, data preparation, database population

All of these steps are performed by Recommender APP. Follow instructions in APP repo to prepare data for this API.

Training recommender API

To train recommender API, run command:

flask train

It will take some time.

You can also start training by accessing route /train.

Running recommender API

To start the API, run command

python app.py

It's not recommended to use this command in production mode. For that, use package uwsgi and configure your server properly (you can find more information here).