pathway-viz 0.1.5

The visualization layer for Pathway - real-time dashboards for streaming data

PathwayViz

Real-time dashboards for streaming pipelines. No frontend code required.

Build once in Python, then:

• run a live dashboard at http://localhost:3000
• embed individual live widgets in your app via iframes at http://localhost:3000/embed/{widget_id}

The Problem

You're building a streaming pipeline. You want a dashboard to see what's happening. Your options:

1. Jupyter notebooks — Great for exploration, but you can't deploy them or embed them in your app
2. Grafana/Prometheus — Works, but now you need to set up metrics exporters, another service, learn PromQL
3. Build a React app — Now you're maintaining frontend code, WebSocket connections, state management

PathwayViz is a simpler option: add a few lines of Python, get a production-ready dashboard.

How It Works

import pathway as pw
import pathway_viz as pv

# Read from Kafka/Redpanda (or any Pathway source)
orders = pw.io.kafka.read(...)
totals = orders.reduce(revenue=pw.reducers.sum(pw.this.amount))

# Widgets auto-update when Pathway updates the table
pv.stat(totals, "revenue", title="Revenue", unit="$")
pv.start()
pw.run()

Run it, open http://localhost:3000, done.

<iframe src="http://localhost:3000/embed/revenue"></iframe>

See Embedding docs for React and Svelte components.

Architecture

flowchart LR
    A[Pathway Pipeline] --> B[PathwayViz Python]
    B --> C[Rust WebSocket Server]
    C --> D[Browser]
    C --> E[Ring Buffers]

The Rust server handles high-throughput WebSockets and caches recent history in ring buffers so new clients can instantly catch up.

Why Rust?

PathwayViz uses a small Rust extension module (built with pyo3) for the hot path:

• WebSocket fan-out to many browser clients (Tokio + Axum)
• in-memory ring buffers for recent history per widget
• serving the embedded frontend assets from the binary

This keeps the runtime responsive under load and avoids Python bottlenecks in the networking/event-loop layer.

Demo

Install

pip install pathway-viz

Documentation

• Getting Started
• Widgets
• Embedding
• Deployment
• Concepts

Roadmap

• Core widgets: stat, chart, gauge, table
• Pathway integration
• Embed mode for iframes
• Theme customization
• Prometheus metrics endpoint
• Authentication

License

MIT