How ramkishan576/moive_recom works
This repository implements a movie-recommendation app: a user selects a movie in the Streamlit interface, which requests recommendations from a FastAPI endpoint and displays the returned titles. The API matches the title against its movie catalog and ranks movies using a similarity matrix. The source also references a similarity pickle that is not present in the supplied file tree; notebook contents and the similarity-building workflow were not inspected.
Main entry points: app.py Movie selector · api.py FastAPI endpoints · movies.pkl Movie catalog
Components 6
User interface
- Movie selector
app.py
Recommendation service
- FastAPI endpoints
api.py - Title matching and ranking
api.py
Recommendation data
- Movie catalog
movies.pkl - Similarity matrix
Other components
- User
Connections 8
- User → Movie selector: selects movie
- Movie selector → Movie catalog: loads titles
- Movie selector → FastAPI endpoints: requests recommendations
- FastAPI endpoints → Title matching and ranking: calls
- Title matching and ranking → Movie catalog: matches and reads
- Title matching and ranking → Similarity matrix: ranks with
- FastAPI endpoints → Movie selector: returns recommendations
- Movie selector → User: displays titles