You are building an application for MovieInsights, a platform that helps movie enthusiasts discover and analyze films. Your application should do the following:
- Find films from a search string
- List 20 most relevant movies per page, up to 3 pages
- Allow optional year filters
- For a movie, show:
- Movie data (title, release date, genre, etc.)
- Top 15 most similar movies
- Explorer: Given a genre and optional release year, identify the most popular movies
- Recommender: Given a viewing occasion, perform a search & recommend a movie from the database
- Provide statistics on the dataset
- Total object count; count by year
The application uses a curated movie dataset containing just under 20,00 movies.
You will learn how to connect a FastAPI app to Weaviate to:
- Implement hybrid search
- With pagination
- With filtering
- View individual object data
- Perform a NearObject search
- Perform retrieval augmented generation
- Manage Weaviate collections
- Create collections with vector configurations
- Delete and rebuild collections
- Handle batch data ingestion
You will be given a skeleton FastAPI-based application, along with scripts to:
- Populate collection (
populate.py): Add movie data to the Weaviate collection - Delete collection (
delete_collection.py): Remove the collection when needed - Main application (
main.py): FastAPI app with endpoints for search, recommendations, and exploration