How lovishaggarwal45/movie_recommendation works
This is a content-based movie recommender with a Streamlit interface and a FastAPI recommendation backend. The user searches for a title, chooses a recommendation count, and receives similar movie names. The README documents the model-preparation stages; runtime wiring is based on the sampled Python files, so the notebook’s internal implementation is not inferred.
Main entry points: movie_recommend.ipynb Preprocessing and model · app.py Title validation and lookup · streamlit_app.py Search and results UI · df.pickle Processed movie data
Components 12
Model preparation
- Movie dataset
Data.csv - Preprocessing and model
movie_recommend.ipynb
Model artifacts
- Text vectorizer
vectorizer.pkl - Feature vectors
X.pkl - Processed movie data
df.pickle - Movie index mapping
indices.pkl
Recommendation backend
- Recommendation API
app.py - Title validation and lookup
app.py - Cosine similarity ranking
app.py - Recommendations
app.py
Web interface
- Search and results UI
streamlit_app.py
Other components
- Movie seeker
Connections 18
- Movie dataset → Preprocessing and model: loads and preprocesses
- Preprocessing and model → Text vectorizer: produces
- Preprocessing and model → Feature vectors: produces
- Preprocessing and model → Processed movie data: produces
- Preprocessing and model → Movie index mapping: produces
- Movie seeker → Search and results UI: searches
- Search and results UI → Processed movie data: loads titles
- Search and results UI → Recommendation API: requests recommendations
- Recommendation API → Title validation and lookup: dispatches
- Title validation and lookup → Processed movie data: checks title
- Title validation and lookup → Movie index mapping: finds index
- Title validation and lookup → Cosine similarity ranking: calls ranking
- Cosine similarity ranking → Feature vectors: scores vectors
- Cosine similarity ranking → Recommendations: ranks titles
- Title validation and lookup → Recommendations: returns not-found response
- Recommendation API → Recommendations: returns recommendations
- Recommendations → Search and results UI: HTTP response
- Search and results UI → Movie seeker: displays results