feat: take transformed coordinates, which are more precise in animation

[abelonogov-ld]

Summary

Take transformed coordinates, which are more precise in animation for both XML and Compose case using 4 points method.

How did you test this change?

Are there any deployment considerations?

---

Note

Switch masking to transformed 4-point polygons for Compose and Views, refactor collection via MaskContext, render via Path, and add a rotating credit card section in the demo form.

• SDK/Masking:
  • Use transformed 4-point polygon masks (points) for both Compose and native views; draw via Path in CaptureSource.drawMasks.
  • Refactor APIs: MaskTarget.mask(context) signature; Mask now holds rect, viewId, and optional points (removed matrix).
  • Introduce MaskCollector.MaskContext (shared Matrix, rootX/Y, matchers) and update traversal to pass context.
  • Compute transformed corners in ComposeMaskTarget.points and NativeMaskTarget.points (API 29+).
• E2E Demo:
  • Group credit card fields into credit_card_section and add continuous Y-rotation animation in XMLUserFormActivity to exercise transformed masking.

^{Written by Cursor Bugbot for commit b6e9440. This will update automatically on new commits. Configure here.}