feat(canvas): add CAD vector semantic context and non-destructive vector crop

Extend CAD conversion to emit bounded semantic metadata (units, layers,
blocks, entity types, labels, crop-indexed bounds) alongside the derived
SVG, tag imported drawings as cad-vector nodes, and carry semanticContext
through canvas nodes, snapshot normalization, and reference images. Agent
Chat forwards this metadata as hidden cad-context reference parts.

Add non-destructive vector cropping: cropping a vector/CAD node keeps the
source SVG and spawns a transparent WebP sibling with crop-scoped semantic
context, wired through the vectorized image toolbar and the workspace crop
tool. Include node --test coverage for CAD semantics and vector crop.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

frontend/tests/vectorImage.test.mjs

@@ -19,6 +19,7 @@ test("recognizes uploaded and CAD-derived SVG image nodes", () => {
   assert.equal(isVectorImageNode({ ...imageNode, content: "/uploads/plan.svg?version=2" }), true);
   assert.equal(isVectorImageNode({ ...imageNode, content: "data:image/svg+xml,%3Csvg/%3E" }), true);
   assert.equal(isVectorImageNode({ ...imageNode, layerRole: "vectorized-image" }), true);
+  assert.equal(isVectorImageNode({ ...imageNode, layerRole: "cad-vector" }), true);
   assert.equal(isVectorImageNode(imageNode), false);
 });