EnterRAG

EnterRAG, refactored.

Structure

• app/
  • config/settings.py — loads env vars
  • services/
    • openai_client.py — OpenAI chat/embeddings
    • chroma_store.py — ChromaDB wrapper
    • mongodb.py — MongoDB connection and CRUD
  • utils/
    • pdf.py — PDF text extraction and chunking
    • json_tools.py — dict flatten helper
  • pages/
    • chatbot.py — chatbot UI + collection manager
    • pdf_to_mongo.py — PDF -> MongoDB
    • mongo_audit.py — AI-assisted edit
    • mongo_viewer.py — view Mongo docs
• index.py — Streamlit entry wiring pages
• requirements.txt — dependencies
• .env.example — copy to .env and fill

Setup

1. Create .env in project root (copy .env.example):

OPENAI_API_KEY=sk-...
MONGODB_URI=mongodb+srv://...
CHROMA_PERSIST_DIR=.chroma
EMBEDDING_MODEL=text-embedding-3-small
CHAT_MODEL=gpt-4o-mini

2. Install dependencies:

python -m venv .venv
.venv\Scripts\Activate.ps1
pip install -r requirements.txt

3. Run the app:

streamlit run index.py

Notes

• Secrets are no longer hard-coded; everything reads from env.
• The original features are preserved: Chroma collections, chatbot, PDF->Mongo, Mongo viewer and audit, finance dashboard.
• If you used a persistent Chroma directory before, set CHROMA_PERSIST_DIR accordingly.