sciplex-flow 0.1.7

Local web app for Sciplex visual programming - node-based data science workflows

Sciplex Flow

Local web application for Sciplex visual programming. Runs a FastAPI/uvicorn backend and serves the bundled React/Vite frontend. Depends on sciplex-core for models, controllers, and default nodes.

Installation

Python 3.11+ is required.

pip install sciplex-flow

Quick start

# Start the local server on the default port (8000)
sciplex-flow

# Choose a port
sciplex-flow --port 5173

What happens:

• FastAPI server starts locally (no cloud calls by default)
• Web UI served from the packaged frontend/dist bundle
• WebSocket channel powers live node execution and UI updates

Project layout

• cli/ — Entry point for the sciplex-flow CLI.
• backend/ — FastAPI app and websocket adapters.
• frontend/ — React/Vite source; dist/ is the built bundle shipped in the wheel.
• frontend/dist/ — Bundled assets included in the package (do not delete; rebuild when frontend changes).

Architecture (high level)

• FastAPI backend serving REST + WebSocket for live updates.
• React/Vite frontend bundled into frontend/dist and served by the backend.
• Depends on sciplex-core for models/controllers and default nodes.

Supported versions

• Python 3.11, 3.12
• Node 20.x (for frontend build)

Development

# Install backend in editable mode
pip install -e ".[dev]"

# Run backend
sciplex-flow --port 8000

# Frontend (from frontend/)
npm install
npm run dev         # dev server with HMR
npm run build       # produces frontend/dist for packaging
npm run lint        # ESLint (TS/React)
npm run typecheck   # TypeScript type-check only

To update the packaged assets, rebuild the frontend and ensure frontend/dist is present before creating a wheel/sdist. CI does this automatically (see .github/workflows/ci.yml).

License

MIT