:blue_book: :book: :orange_book: :neckbeard: Books recommendation project :neckbeard: :orange_book: :book: :blue_book:
Project description: Recommendation system for books.
Project install steps:
- clone repo
git clone https://github.com/naz2001r/recsys_project_MAUN.git - create virtual environment
conda create --name py311 python=3.11 - activate virtual environment
conda activate py311(conda)
We are using poetry to manage dependencies.
- install poetry
pip install poetry - install dependecies using poetry. It could take some time
poetry install - pull needed data using dvc (it can take up to 30 minutes, please be patient, to see the progress we use verbose flag in here)
dvc pull -v - reproduce training pipeline (actually all steps should be skipped, that will mean that you have latest data locally)
dvc repro -v - to run single step of pipeline
dvc repro -v -sf STEPNAME - to run all steps of pipeline after some step
dvc repro -v -f STEPNAME --downstream - to run all steps of pipeline after some step without running them actually
dvc repro -v -f STEPNAME --downstream --dry
EDA jupyter notebook is in /data/notebooks folder. Note: EDA will be available after dvc pull -v
The final report and Online Methodology are in one file in the folder report.
Models developed:
- Baseline model
- Collaborative filtering
- Content-base filtering
- Matrix factorization
- Content-base Sentence Transformer recomender
- Super-duper hybrid NN + Sentence Transformer recommender
- Hybrid Matrix factorization + Sentence Transformer recommender
Inference app !NB Pull dvc pipeline first.
- To run inference app
python -m inference - Follow the instructions.
Inference usage example:
Contributors:
- Nazarii Drushchak
- Oleksandr Vashchuk
- Marianna Kovalova
- Uliana Zbezhkhovska
For more information reach us via slack