How tarunkaul35/ai-based-movies-recommendation-system works
This is a Streamlit movie recommender. The user selects a title, requests recommendations, and sees similar movies filtered by genre, rating, and year, with poster images where available. Offline preprocessing cleans the IMDb CSV, builds CountVectorizer features, computes cosine similarity, and saves data consumed by the recommendation function. The UI calls that function and retrieves posters from OMDb. The source index and excerpts support these main connections; generated pickle files are runtime artifacts, not listed in the tree.
Main entry points: app.py Streamlit app and filters · model.py Pickled movie data and similarity · main.py Top-five recommender · poster.py Poster lookup
Components 8
Data preparation
- IMDb movie data
IMDb_All_Genres_etf_clean1.csv - Feature and similarity builder
model.py - Pickled movie data and similarity
model.py
Recommendation
- Top-five recommender
main.py
User interface
- Streamlit app and filters
app.py - Poster lookup
poster.py
Other components
- Movie viewer
- OMDb API
Connections 9
- IMDb movie data → Feature and similarity builder: reads CSV
model.py - Feature and similarity builder → Pickled movie data and similarity: writes pickles
model.py - Top-five recommender → Pickled movie data and similarity: loads pickles
main.py - Streamlit app and filters → Pickled movie data and similarity: loads movie data
app.py - Movie viewer → Streamlit app and filters: selects title
app.py - Streamlit app and filters → Top-five recommender: requests recommendations
app.py - Streamlit app and filters → Poster lookup: requests poster
app.py - Poster lookup → OMDb API: queries API
poster.py - Streamlit app and filters → Movie viewer: shows filtered results
app.py