Apppublic

Awais234/las-cms-bot

sourceHugging Faceupdated 4mo agoView on Hugging Face

0likes

App README

LAS CMS Bot

RAG chatbot over LAS Case Management System data (3,800+ cases).

Stack

—Backend: FastAPI (Python 3.11)
—Vector store: ChromaDB with all-MiniLM-L6-v2 embeddings
—LLM: OpenAI gpt-4o-mini
—Auth: SQLite + bcrypt + JWT cookies
—Hosting: Hugging Face Spaces (Docker SDK)

Deploy on Hugging Face

1.Create a new Space at https://huggingface.co/new-space (Docker SDK)
2.Push the CMSBot/app/ folder as the Space root
3.Set Secrets in Settings → Variables and secrets:

Name	Value
`OPENAI_API_KEY`	Your OpenAI API key
`JWT_SECRET`	Long random string (`python -c "import secrets; print(secrets.token_hex(32))"`)
`SECURE_COOKIES`	`1`
`BOOTSTRAP_USER`	`admin`
`BOOTSTRAP_PASSWORD`	Choose a strong password

1.Before first run, copy cms_cases_rag.jsonl into data/ folder
2.The Space will auto-deploy. Login as admin / your password.

Local dev

bash

pip install -r requirements.txt
cp .env.example .env    # add your OpenAI key
mkdir -p data
# Copy cms_cases_rag.jsonl to data/
python ingest.py
python manage_users.py add admin "Admin" --admin
uvicorn backend.main:app --reload

User management

bash

python manage_users.py add client1 "Client Name"
python manage_users.py list
python manage_users.py delete client1

Folder structure

CMSBot/
├── app/
│   ├── backend/       # Python: auth, config, chatbot, main
│   ├── frontend/      # HTML: login, chat UI, admin
│   ├── data/          # RAG data + ChromaDB (runtime)
│   ├── Dockerfile
│   ├── ingest.py      # Load JSONL to ChromaDB
│   ├── manage_users.py
│   └── requirements.txt
└── README.md