Skip to content

Latest commit

 

History

History
55 lines (37 loc) · 1.27 KB

README.md

File metadata and controls

55 lines (37 loc) · 1.27 KB
Raw

Movie App

This is a movie app that uses 680 movies between 1990 and 2024 from TMDB stored in a vector database. The app uses a simple cosine similarity to find the most similar movies (recommendation) to a given movie.

How to run ?

Make sure you have docker installed on your machine!

  1. Clone the repo
git clone https://github.com/Vicba/movies-recommendation.git
  1. Run docker-compose up in the root directory
docker-compose up
  1. Populate the database with the movies.
curl -X GET http://localhost:5000/populate
  1. Open http://localhost:3000 in your browser
  2. Browse around!

Technologies

  • Nextjs (typescript, Tailwindcss)
  • Flask
  • Weaviate
  • Docker
  • Huggingface API

The embedding model

The embedding model used is sentence-transformers/paraphrase-MiniLM-L6-v2 from huggingface. It has 384 dimensions.

If you want to use something else, you can change it in the /api/build_knowledge_base/embed.py file. Run the python script to generate the csv with embeddings csv in datasets folder.

cd api/build_knowledge_base
python embed.py

Learnings

  • Learned how to use Weaviate
  • Refresh my knowledge in nextjs & docker
  • Usign huggingface API
  • Project went super smooth with the research and pre-defined scope