v0.5.1

Added

• GraphWidget for programmatic force-directed graph visualization. Nodes and edges can be supplied from Python with optional names, sizes, colors, and metadata; browser interactions support zoom, pan, drag, hover tooltips, bounded or unbounded layouts, and selection synced back to Python.
• GraphWidget.attach_node(...) for adding a new or existing node together with its connecting edge, plus detach_node(...) for removing attached edges while optionally deleting the node. The frontend now initializes newly connected nodes near their existing endpoint and restarts the force simulation gently, avoiding full-layout jumps for incremental graph updates.
• GraphWidget.hovered_node traitlet exposes the id of the node currently under the cursor (None when no node is hovered), so Python can react to hover without needing a click.
• Added the McNugget sums demo (demos/mcnugget_graph.py) — a GraphWidget + PlaySlider notebook that breadth-first-searches reachable totals from a configurable set of box sizes.
• Added a "Random growth graph" section to demos/graphwidget.py — at each step a new node is added and connected to a random existing node, scrubbed with PlaySlider.
• Added the Hypercube example (demos/hypercube.py, linked from the docs gallery Examples section) showing how the force layout unfurls $Q_n$ — $Q_3$ snaps into a cube, $Q_4$ projects into a tesseract. Sliders pick dimension $n$ and node size; nodes are colored by popcount, edges by which bit position the endpoints differ in.