| title | Bioimage Analysis Triage Bot |
|---|---|
| emoji | 🔬 |
| colorFrom | blue |
| colorTo | indigo |
| sdk | gradio |
| sdk_version | 6.19.0 |
| python_version | 3.14.6 |
| app_file | app.py |
| pinned | false |
A conversational chatbot for triaging bioimage analysis requests for the Image Analysis Team at the Francis Crick Institute.
BIA BOT is an AI-powered triage system that interviews users about their bioimage analysis projects and automatically generates:
- Triage Reports — comprehensive summaries of user requirements and project details
- Python Scripts — executable analysis scripts tailored to the specific analysis task
The chatbot uses a structured 20-question framework to gather information efficiently while minimizing user fatigue.
Try it here: https://huggingface.co/spaces/djpbarry/bia-bot
- 🤖 Conversational Interview — Natural language dialogue guided by a 20-question framework
- 📋 Automated Report Generation — Comprehensive markdown reports summarizing the triage
- 🐍 Script Generation — Clean, commented Python code for the proposed analysis
- 📧 Email Delivery — Reports and scripts zipped and sent to the facility team
- 💾 Interaction Logging — Full conversation transcripts saved for reference
- ♿ User-Friendly UI — Clean Gradio interface with progress feedback
- User starts a conversation with the chatbot
- AI asks targeted questions about their bioimage analysis project
- User provides information about images, objectives, constraints, and requirements
- When complete, the bot automatically:
- Saves the interaction transcript
- Generates a comprehensive triage report
- Creates a Python analysis script
- Sends both as a zip file to the facility team
- Facility team reviews and follows up with detailed recommendations
- LLM: Claude Haiku (Anthropic API)
- Interface: Gradio
- Backend: Python
- Deployment: Hugging Face Spaces
- Email: Gmail SMTP
- Python 3.14+
pixi(or pip for manual installation)- Anthropic API key
- Gmail app password (for email functionality)
# Clone the repo
git clone https://github.com/djpbarry/biabot.git
cd biabot
# Install dependencies with pixi
pixi install
# Or with pip
pip install -r requirements.txtCreate a .env file in the project root:
ANTHROPIC_API_KEY=your-anthropic-key
GMAIL_ADDRESS=your-email@gmail.com
GMAIL_APP_PASSWORD=your-app-password
python app.pyThe app will launch at http://127.0.0.1:7860
.
├── app.py # Main Gradio application
├── papers/Twenty_Questions.pdf # Triage framework document
├── requirements.txt # Python dependencies
├── interactions/ # Saved conversation logs (JSON)
├── reports/ # Generated triage reports (markdown)
├── scripts/ # Generated analysis scripts (Python)
└── README.md # This file
- The chatbot will ask for your email address so the facility team can contact you with recommendations
- Reports and scripts are sent as a single zip file to bypass email client filters
- All conversations are logged locally for quality assurance
- The framework is based on the "Twenty Questions" standard for bioimage analysis