This repo is the development and experiment pipeline for recommendation model.
As default, scikit-surprise is wrapped and you can make use case for your development.

First, update some libraries.

pip install --upgrade setuptools wheel

On the top directory of this repo

pip install -e .

As example, development and experiment with MovieLens dataset is there.
To run, you need to download the data, ml-latest-small.zip, from https://grouplens.org/datasets/movielens/ . There, you can find ratings.csv. On the recommender_flow/use_case/movie_lens/development/setting.py, you need to set the path to this ratings.csv.
From the top of the repo, you can run

python recommender_flow/interface/script/development_movie_lens.py

It will run the SVD model and show simple evaluation.