feat: implement programmatic lasso selection

[glebmoisseyev]

• Add dedicated lassoSelect() method for programmatic polygon selection
• Accepts array of [x,y] coordinate pairs in data space (default) or GL space
• Supports merge, remove, and isGl options
• Auto-closes open polygons via verticesToPolygon() helper
• Add test suite (8 tests)
• Add programmatic-lasso.js example with interactive demo
• Update README with API documentation and examples
• Add to Examples menu in Tweakpane

Description

Adds scatterplot.lassoSelect(vertices, options) method to enable automated point selection without manual interaction. Vertices can be specified in data space (using xScale/yScale) or GL space (isGl: true). This provides a programmatic alternative to the interactive lasso tool.

What was changed in this pull request?

New API:

• lassoSelect(vertices, { merge, remove, isGl }) - dedicated method for programmatic lasso
• vertices - array of [x, y] coordinates (minimum 3 vertices)
• isGl: false (default) - vertices in data space, requires xScale/yScale
• isGl: true - vertices in GL space, works without scales

Helper functions:

• isVertices() - validates all vertices are valid [x, y] pairs
• verticesToPolygon() - auto-closes polygons if first/last vertices differ
• verticesFromDataToGl() - converts data space → GL space coordinates

Example

// Data space (requires xScale/yScale)
scatterplot.lassoSelect([
  [10, 20],
  [50, 80],
  [90, 30]
]);

// GL space (works without scales)
scatterplot.lassoSelect(
  [[-0.5, -0.5], [0.5, -0.5], [0.5, 0.5], [-0.5, 0.5]],
  { isGl: true }
);

// Merge and remove modes
scatterplot.lassoSelect(polygon, { merge: true });
scatterplot.lassoSelect(polygon, { remove: true });

Why is this PR necessary?

• Annotation replay: load an existing session and re-apply the saved lasso region so the same points get selected again.
• Predefined regions: select a known ROI polygon (e.g., “this cluster boundary”) as part of a workflow.
• Reproducible selections: store the polygon vertices and get consistent results across runs.

Checklist

• Provided a concise title as a semantic commit message (e.g. "fix: correctly handle undefined properties")
• CHANGELOG.md updated
• Tests added or updated
• Documentation in README.md added or updated
• Example(s) added or updated
• Screenshot, gif, or video attached for visual changes

[flekschas]

🙌 Thanks for putting this together! 🎉 I really like this feature 👏 but I would say we should expose it as a new function: scatterplot.lasso(). I tried to flesh out the details as PR comments but there's more docs adjustments needed.

Most critically, we need a test :) You could add one that's essentially identical to https://github.com/flekschas/regl-scatterplot/blob/main/tests/events.test.js#L312-L399 but uses the new lasso method instead of emitting events. We should be testing:

1. data space lasso polygon
2. GL space lasso polygon
3. an open polygon
4. edge cases like when less than 3 vertices are passed to the lasso() function or the user uses vertices in data space without xScale and yScale