README.md

metadata

title: Feedback Chatbot
emoji: 🌖
colorFrom: purple
colorTo: green
sdk: docker
app_port: 7860
pinned: false

Quick start

1. Duplicate this Space on Hugging Face.
2. In your Space → Settings → Variables and secrets, add:
   • HF_TOKEN — required; use a token with write access so the app can create/sync the dataset.
   • APP_PASSWORD — strongly recommended if the Space is public/protected; default is password.
3. Optional overrides:
   • APP_USERNAME — defaults to student.
   • DATASET_ID — defaults to <SPACE_ID>-corpus on Spaces and is created automatically when needed.
   • ADMIN_PASSCODE — defaults to password; set this for real deployments.
   • (optional) LLM_MODEL — defaults to deepseek-ai/DeepSeek-V4-Flash; change this Space variable to switch models without a code push.
   • (optional) LLM_MAX_TOKENS — defaults to 8192.
   • (optional) LLM_FINAL_ANSWER_MAX_TOKENS — defaults to 8192.
   • (optional) EMBED_MODEL — defaults to BAAI/bge-small-en-v1.5.
4. Restart the Space. Open /admin, sign in with the passcode, and upload rubric files (.pdf, .docx, .txt, .md).
5. Share the root URL with students. They can paste or upload tutor feedback and upload coursework locally for the chat session.

Student-uploaded tutor feedback and coursework are parsed in the running app and sent as chat context only. They are not saved to the linked Hugging Face Dataset. The Dataset stores teacher-uploaded rubrics and the vector cache used to retrieve rubric excerpts.

Local development

Requires Python 3.13, Node 20+, and uv.

cp .env.example .env   # fill in HF_TOKEN; other values have defaults

# Backend
cd backend && uv sync && uv run uvicorn main:app --port 8000 --reload

# Frontend (in another shell)
cd frontend && npm install && npm run dev

Visit http://localhost:7860. Next.js proxies /api/* to 127.0.0.1:8000.

Or with Docker (matches the HF Space environment exactly)

docker build -t iamearthdev .
docker run --rm -p 7860:7860 --env-file .env iamearthdev

Customizing personas

Edit frontend/lib/personas.ts — each persona is just an id, name, description, and a prompt that gets prepended to the system message. Rebuild the Space to deploy.

Repo layout

/
├── Dockerfile          # multi-stage: Next build → Python+Node runtime
├── start.sh            # launches FastAPI (:8000) and Next.js (:7860)
├── .env.example
├── backend/
│   ├── pyproject.toml  # uv-managed deps
│   ├── main.py         # FastAPI app, /chat streams NDJSON
│   ├── rag.py          # rubric parsing → LlamaIndex → LanceDB
│   ├── hf_sync.py      # Dataset sync for rubrics and vector cache
│   └── auth.py         # passcode cookie middleware
└── frontend/
    ├── next.config.js  # /api/* → http://127.0.0.1:8000/*
    ├── package.json
    ├── app/
    │   ├── page.tsx           # chat UI
    │   ├── admin/page.tsx     # upload + file list
    │   └── admin/login/page.tsx
    └── components/            # chat-interface, persona-selector, file-uploader, …

License

This project is released under the MIT License.

MIT License

Copyright (c) 2026 Dr Noorhan Abbas and Sen Lin

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Citation

If you use or adapt this tool in your work, please cite:

Abbas, N. (under review). Investigating Student Perceptions of an AI-Powered Chatbot to Support Feedback Interpretation and Uptake in Higher Education. Manuscript submitted to Assessment and Evaluation in Higher Education.