BioLM JupyterLab Extension

A JupyterLab extension that provides a graphical interface for browsing BioLM models, inserting SDK code snippets, and managing authentication - all from within JupyterLab's sidebar.

Features

🧠 Models Tab

• Browse all available BioLM models
• Search and filter models by name, description, or tags
• Click tags to insert code snippets with model + action pre-filled
• Copy model IDs to clipboard
• Expandable model descriptions

⚙️ Operations Tab

• Quick access to common SDK operation examples
• Searchable list of operations (Generate, Predict, Embed, etc.)
• One-click code insertion into notebook cells
• Organized by category

🔑 Settings Tab

• Manage multiple API key profiles
• Test API key connections
• Set default model and action preferences
• Respects BIOLM_API_KEY environment variable

Installation

From PyPI (when available)

pip install jupyterlab-biolm

From Source

# Clone the repository
git clone https://github.com/yourusername/jupyterlab-biolm.git
cd jupyterlab-biolm

# Install the extension
pip install -e .

# Build the extension
jlpm install
jlpm build

# Rebuild JupyterLab
jupyter lab build

Usage

1. Open the Extension: The BioLM sidebar will appear in JupyterLab's left sidebar. If not visible, use the command palette (Cmd/Ctrl + Shift + C) and search for "Open BioLM".

2. Configure API Key:

   • Go to the Settings tab
   • Add a new profile with your API key
   • Or use the BIOLM_API_KEY environment variable

3. Browse Models:

   • Switch to the Models tab
   • Search or filter by tags
   • Click a tag button to insert code for that model + action

4. Use Operations:

   • Switch to the Operations tab
   • Browse common SDK operations
   • Click "Insert" to add code to your active notebook cell

Development

Prerequisites

• Node.js >= 16
• Python >= 3.8
• JupyterLab >= 4.0.0

Setup

# Install dependencies
jlpm install

# Build TypeScript
jlpm build:lib

# Build extension in development mode
jlpm build:labextension:dev

# Watch for changes
jlpm watch

Project Structure

jupyterlab-biolm/
├── src/                    # TypeScript source
│   ├── index.ts           # Extension entry point
│   ├── widget.tsx         # Main sidebar widget
│   ├── api/               # API client
│   ├── components/        # React components
│   ├── services/          # Business logic
│   └── data/              # Static data
├── style/                  # CSS styles
├── schema/                 # Settings schema
└── jupyterlab_biolm/       # Python package

Configuration

User Settings

The extension stores settings in JupyterLab's settings system. You can configure:

• API Key Profiles: Multiple named profiles with different API keys
• Active Profile: Which profile to use (or environment variable)
• Default Model: Default model ID for code generation
• Default Action: Default action (predict, generate, etc.)

Environment Variables

The extension supports several environment variables for deployment-level configuration:

BIOLM_USE_PROXY (default: true)

Controls whether to use the server-side proxy for API requests or make direct API calls from the browser.

• true (default): Use server proxy to avoid CORS issues. Recommended for most deployments.
• false: Make direct API calls from the browser. Only use when CORS is properly configured on the API server.

# Use proxy (default, recommended)
export BIOLM_USE_PROXY=true

# Use direct API calls (requires CORS configuration)
export BIOLM_USE_PROXY=false

BIOLM_API_BASE_URL (default: https://biolm.ai/api/ui/community-api-models/)

Sets the base URL for the BioLM API. Useful for:

• Testing against staging environments
• Using self-hosted BioLM instances
• API version changes

# Use default production API
export BIOLM_API_BASE_URL=https://biolm.ai/api/ui/community-api-models/

# Use staging environment
export BIOLM_API_BASE_URL=https://staging.biolm.ai/api/ui/community-api-models/

Note: The URL should end with a trailing slash. If not provided, one will be added automatically.

BIOLM_REQUEST_TIMEOUT (default: 10)

Sets the timeout (in seconds) for API requests made by the server extension.

# Use default 10 second timeout
export BIOLM_REQUEST_TIMEOUT=10

# Increase timeout for slow networks
export BIOLM_REQUEST_TIMEOUT=30

BIOLM_CACHE_TTL (default: 300000)

Sets the cache time-to-live in milliseconds for model metadata. The default is 5 minutes (300000 ms).

# Use default 5 minute cache
export BIOLM_CACHE_TTL=300000

# Disable caching (0) or set custom duration
export BIOLM_CACHE_TTL=600000  # 10 minutes

BIOLM_API_KEY (optional)

Sets the default API key to use if no profile is selected in the UI. This is a user-facing setting that can also be configured in the Settings tab.

export BIOLM_API_KEY=your_api_key_here

Example: Docker Deployment

FROM jupyter/scipy-notebook

# Install extension
RUN pip install jupyterlab-biolm

# Configure environment variables
ENV BIOLM_USE_PROXY=true
ENV BIOLM_API_BASE_URL=https://biolm.ai/api/ui/community-api-models/
ENV BIOLM_REQUEST_TIMEOUT=10
ENV BIOLM_CACHE_TTL=300000

Example: Codespaces / Cloud Deployment

Set environment variables in your deployment configuration:

# .github/workflows/deploy.yml or similar
env:
  BIOLM_USE_PROXY: "false"  # If CORS is configured
  BIOLM_API_BASE_URL: "https://biolm.ai/api/ui/community-api-models/"
  BIOLM_REQUEST_TIMEOUT: "15"

API Integration

The extension fetches model metadata from https://api.biolm.ai/models. Models are cached for 5 minutes to reduce API calls.

Requirements

• JupyterLab >= 4.0.0
• React >= 18.0.0

License

BSD-3-Clause

Support

For issues and feature requests, please visit the GitHub repository.