fix(text-actions): Calculate accumulated transform matrix for text inside groups (#1082)

When editing text inside a group element, the text cursor was appearing
in the wrong position because only the text element's own transform was
being considered, ignoring transforms from parent groups.

This change:
- Adds a new #getAccumulatedMatrix() method that traverses up the DOM tree
  from the text element to the SVG content element, collecting and multiplying
  all transform matrices along the way
- Updates the init() method to use this accumulated matrix instead of just
  the text element's transform
- Updates test mock to include getSvgContent() method

Fixes the issue where editing text inside a transformed group would show
the cursor at the wrong position.

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

tests/unit/text-actions.test.js

@@ -18,12 +18,17 @@ describe('TextActions', () => {
     svgRoot.setAttribute('height', '480')
     document.body.append(svgRoot)
 
+    // Create svgContent element (container for SVG content)
+    const svgContent = document.createElementNS(NS.SVG, 'svg')
+    svgContent.id = 'svgcontent'
+    svgRoot.append(svgContent)
+
     textElement = document.createElementNS(NS.SVG, 'text')
     textElement.setAttribute('x', '100')
     textElement.setAttribute('y', '100')
     textElement.setAttribute('id', 'text1')
     textElement.textContent = 'Test'
-    svgRoot.append(textElement)
+    svgContent.append(textElement)
 
     // Mock text measurement methods
     textElement.getStartPositionOfChar = vi.fn((i) => ({ x: 100 + i * 10, y: 100 }))
@@ -55,6 +60,7 @@ describe('TextActions', () => {
     // Mock svgCanvas
     svgCanvas = {
       getSvgRoot: () => svgRoot,
+      getSvgContent: () => svgContent,
       getZoom: () => 1,
       setCurrentMode: vi.fn(),
       clearSelection: vi.fn(),