How ramasamy-24-t/precision-farming-assistant works
Agricultural RAG Assistant ingests agricultural PDFs into a persistent vector store, then answers crop-advisory questions using multi-query retrieval and GPT-generated responses. Users can query through the notebook chat loop or the Flask REST API; the documented API also provides a health check. All application components are implemented in RAG.ipynb. Relationships below follow the README’s architecture and feature descriptions; no notebook source excerpts were available to verify finer control flow.
Main entry points: RAG.ipynb Query Paraphrases
Components 16
PDF Ingestion
- PDF Partitioning
RAG.ipynb - Title Chunking
RAG.ipynb - Chunk Summaries
RAG.ipynb - Text Embeddings
RAG.ipynb
Knowledge Store
- ChromaDB Store
RAG.ipynb
Answering
- Query Paraphrases
RAG.ipynb - RRF Retrieval
RAG.ipynb - Answer Generation
RAG.ipynb - Conversation History
RAG.ipynb
User Interfaces
- Notebook Chat
RAG.ipynb - Flask API
RAG.ipynb - Health Check
RAG.ipynb
Other components
- Agricultural PDFs
- OpenAI API
- Question
- API Client
Connections 21
- Agricultural PDFs → PDF Partitioning: provides documents
README.md - PDF Partitioning → Title Chunking: partitions into elements
README.md - Title Chunking → Chunk Summaries: creates chunks
README.md - Chunk Summaries → OpenAI API: requests summaries
README.md - Chunk Summaries → Text Embeddings: enriches chunks
README.md - Text Embeddings → OpenAI API: requests embeddings
README.md - Text Embeddings → ChromaDB Store: stores vectors
README.md - Question → Query Paraphrases: provides question
README.md - Query Paraphrases → OpenAI API: generates queries
README.md - Query Paraphrases → RRF Retrieval: supplies query variants
README.md - RRF Retrieval → ChromaDB Store: searches vectors
README.md - RRF Retrieval → Answer Generation: provides retrieved context
README.md - Answer Generation → OpenAI API: generates answer
README.md - Conversation History → Notebook Chat: supports multi-turn chat
README.md - Question → Notebook Chat: asks question
README.md - Notebook Chat → Query Paraphrases: runs retrieval
README.md - Answer Generation → Notebook Chat: returns answer
README.md - API Client → Flask API: sends question
README.md - Flask API → Query Paraphrases: runs retrieval
README.md - Answer Generation → Flask API: returns answer
README.md - API Client → Health Check: checks status
README.md