refactor: deduplicate image extraction logic

Extract shared image processing and retrieval logic into reusable helper functions, eliminating ~150 lines of duplicate code between extractImagesFromPage and extractPageContent.

**Changes:**
- Add processImageData() - converts raw PDF.js image data to ExtractedImage
- Add retrieveImageData() - handles image retrieval strategy (commonObjs -> sync -> async with timeout)
- Refactor extractImagesFromPage to use shared helpers
- Refactor extractPageContent to use shared helpers while preserving yPosition

**Benefits:**
- Reduces code duplication by ~150 lines
- Improves maintainability - fixes/improvements in one place
- Increases test coverage from 90.7% to 95.37%
- Consistent error handling and timeout behavior across both functions

Author: shtse8