krishAIbackend Setup (Windows)

End-to-end setup for the Kerala Farming RAG Chatbot + Plant Disease Detection.

1. Prerequisites

• Python 3.13 (preferred; repo pycache shows 3.13)
• Git
• (Optional) Google Gemini API key for better answers
• (Alternative) Ollama installed + ollama pull llama2
• ~6 GB free disk (models + embeddings + deps)

2. Clone & Enter

git clone <repo-url>
cd krishAIbackend

3. Virtual Environment

py -3.13 -m venv .venv
./.venv/Scripts/Activate.ps1

If PowerShell blocks scripts: run (once, as admin):

Set-ExecutionPolicy RemoteSigned

4. Environment Variables

Create .env (or copy from example):

GEMINI_API_KEY=your_key_here   # leave blank to use Ollama llama2

If no key: ensure Ollama service running and model pulled:

ollama pull llama2

5. Install Dependencies

Both subprojects share nearly identical requirements. Install from one list (choose the larger — disease_detection):

pip install --upgrade pip
pip install -r disease_detection/requirements.txt

(First install may take several minutes: torch, sentence-transformers, chromadb.)

6. Build Vector Database (RAG Knowledge Base)

This uses PDFs in rag_chatbot/data/kb and persists embeddings to rag_chatbot/chroma_db.

python rag_chatbot/main.py --setup-db

Rebuild (e.g., after adding PDFs):

python rag_chatbot/main.py --setup-db --force-rebuild

7. Run Chatbot API (Flask)

python flask_app.py

Test endpoints:

curl http://127.0.0.1:8000/health
curl -X POST http://127.0.0.1:8000/chat -H "Content-Type: application/json" -d '{"query":"Best paddy practices in monsoon?"}'

8. CLI Chat (Optional)

python rag_chatbot/main.py --chat

Combine build + chat:

python rag_chatbot/main.py --setup-db --chat

9. Plant Disease Detection (Image + RAG Advice)

Image model: disease_detection/model/plant_disease_cnn_scripted.pt (CPU inference). Run analysis (uses disease_detection/chroma_db by default—point it to the one you built if needed):

python disease_detection/main.py --image disease_detection/test_images/CornCommonRust1.JPG --db-path ../rag_chatbot/chroma_db

Interactive mode:

python disease_detection/main.py --image disease_detection/test_images/PotatoHealthy2.JPG --interactive --db-path ../rag_chatbot/chroma_db

10. Common Issues

Problem	Fix
Module import errors	Ensure venv active & deps installed
Chatbot says DB missing	Re-run vector build step
Gemini errors	Check GEMINI_API_KEY and network
Ollama fallback fails	Start Ollama app/service; pull model
Slow first answer	Embedding model + DB warm-up; improves after

11. Updating Knowledge Base

1. Drop new PDFs into rag_chatbot/data/kb/
2. Rebuild: python rag_chatbot/main.py --setup-db --force-rebuild
3. Restart Flask server (if running)

12. Minimal Smoke Tests

# Health
curl http://127.0.0.1:8000/health
# Simple chat
curl -X POST http://127.0.0.1:8000/chat -H "Content-Type: application/json" -d '{"query":"Soil conservation methods"}'
# Disease detection
python disease_detection/main.py --image disease_detection/test_images/virus4.JPG --db-path ../rag_chatbot/chroma_db

13. Folder Roles (Quick)

• flask_app.py : REST API wrapper over RAG chatbot
• rag_chatbot/ : RAG pipeline, vector DB builder, PDFs
• disease_detection/ : Image classifier + RAG treatment adviser

14. Performance Notes

All configured for CPU; set model_kwargs={'device':'cuda'} in build_vector_db.py & adjust torch install for GPU acceleration.

15. Next Steps (Optional Enhancements)

• Unify requirements into root requirements.txt
• Add Dockerfile
• Add basic pytest health tests
• Share single chroma_db between both subsystems via config

---

Happy farming 🌱