This is a Streamlit-based application that performs semantic search on a dataset of movie titles and overviews. It uses the SentenceTransformer model to create text embeddings, which are stored in a SQLite database. Users can search for movies by inputting a description, and the app will find the most similar movies based on cosine similarity.
- Load and display a dataset of the top 1000 IMDb movies.
- Clean and preprocess movie titles and overviews.
- Generate sentence embeddings using the
paraphrase-MiniLM-L6-v2model fromSentenceTransformer. - Store the embeddings and movie information in a SQLite database.
- Search for similar movies using cosine similarity on the embeddings.
Make sure you have the following installed on your system:
- Python 3.8+
- Streamlit
- Pandas
- Numpy
- Sentence-Transformers
- SQLite (bundled with Python via
sqlite3)
You can install the required libraries using pip:
pip install streamlit pandas numpy sentence-transformers