How mejji/movie-search-with-vectors---a-beginner-s-guide-to-qdrant--embeddings---rag works
A Django movie-semantic-search application offers CLI commands and a browser UI over a shared Qdrant/embedding core. The main workflow loads and embeds the TMDB catalogue, then searches it by meaning; RAG chat retrieves movie context before generating answers. Exploration and vector visualizations are additional product capabilities. The README documents the broader interfaces; sampled source confirms the principal search, ingest, chat, and exploration wiring, but not every view or command implementation.
Main entry points: qdrant_utils.py Qdrant utilities · manage.py Django CLI · views.py Web views · visualize_vectors.py Vector charts
Components 15
User interfaces
- Django CLI
manage.py - Web routes
tutorial/urls.py - Web views
reviews/views.py - Web pages
reviews/templates/reviews
Movie workflows
- Load movies
reviews/management/commands/load_movies.py - Search movies
reviews/management/commands/search_movies.py - RAG chat
reviews/management/commands/rag_chat.py - Vector explorer
reviews/management/commands/explore_vectors.py - Vector charts
reviews/management/commands/visualize_vectors.py
Vector search core
- Qdrant utilities
reviews/qdrant_utils.py
Data and services
- TMDB catalogue
data/tmdb_5000_movies.csv - Local Qdrant
- Embedding model
- Groq LLM API
Other components
- User
Connections 24
- User → Django CLI: runs commands
- User → Web routes: browses
- Django CLI → Load movies: dispatches
- Django CLI → Search movies: dispatches
- Django CLI → RAG chat: dispatches
- Django CLI → Vector explorer: dispatches
- Django CLI → Vector charts: dispatches
- Web routes → Web views: routes requests
- Web views → Web pages: renders pages
- Web views → Qdrant utilities: searches movies
- Web views → Local Qdrant: reads collection
- Web views → Groq LLM API: generates answers
- Load movies → TMDB catalogue: reads catalogue
- Load movies → Qdrant utilities: creates and embeds
- Qdrant utilities → Embedding model: encodes text
- Qdrant utilities → Local Qdrant: creates and queries
- Search movies → Qdrant utilities: searches movies
- RAG chat → Qdrant utilities: retrieves movies
- RAG chat → Groq LLM API: generates answers
- Vector explorer → Local Qdrant: inspects vectors
- Vector explorer → Embedding model: encodes queries
- Vector charts → Local Qdrant: reads vectors
- Vector charts → Qdrant utilities: searches matches
- Vector charts → Embedding model: encodes query