refactor: optimize image handling and simplify module architecture

[gupsammy]

Summary

This refactor consolidates image handling logic and eliminates unnecessary module abstraction layers. The primary improvement is moving image analysis into extractContent() to return pre-analyzed image data, eliminating redundant HTML parsing when handling images later.

• Single-pass image extraction: analyzeImagesInHtml() replaces separate extractImagesFromHtml() and countImagesInHtml() calls
• Pre-analyzed images in extraction result: Images are analyzed once during content extraction and passed through to content-script.js
• Unified naming function: generateBaseName() consolidates generateFilename() and generateSubfolder() logic
• Removed platform.js wrapper: Direct use of EzyCopyFiles module in content-script.js, reducing module layers
• Fixed popup.js toggle synchronization: Both checkboxes now sync correctly when enforcement rules apply

Key Changes

• lib/ezycopy.js: Replaced image extraction functions with unified analyzeImagesInHtml(), integrated image analysis into extractContent() flow, consolidated naming functions
• content-script.js: Updated to consume pre-analyzed images from extraction result, removed redundant image extraction call, uses EzyCopyFiles directly
• file-helpers.js: Removed generatePageSubfolder() function (replaced by generateBaseName)
• injection-files.js: Removed lib/platform.js from injection list
• popup.js: Fixed toggle state synchronization in both change handlers

Files Changed

• lib/ezycopy.js: Core refactoring of image and naming functions
• content-script.js: Updated to work with pre-analyzed image data
• file-helpers.js: Removed duplicate naming function
• injection-files.js: Simplified injection list
• popup.js: UI state synchronization fix
• lib/platform.js: Deleted (functionality moved to direct EzyCopyFiles usage)
• CLAUDE.md: Updated documentation

This reduces complexity, improves performance by eliminating redundant parsing, and streamlines the data flow from extraction to download.

[claude]

Claude finished @gupsammy's task —— View job

---

Code Review

Todo List:

• Review lib/ezycopy.js refactoring
• Check content-script.js integration with pre-analyzed images
• Verify platform.js removal doesn't break functionality
• Review popup.js toggle fix
• Check for critical bugs or Chrome Extension API issues

---

Review Summary

This PR looks good to ship after fixing the image count bug below. The refactoring achieves its goal of eliminating redundant HTML parsing while simplifying the module architecture.

---

CRITICAL Issues

1. Image count calculation bug in analyzeImagesInHtml()

Location: lib/ezycopy.js:91

Issue: The function returns count: imgElements.length but the actual images array may be shorter due to filtering (data URIs, duplicates). This causes incorrect image loss ratio calculations in extractContent().

// Current code returns total count, but images array is filtered
return { count: imgElements.length, images };

Impact: The image loss ratio calculation at line 335-337 will be wrong:

const containerImageCount = container.querySelectorAll('img').length;
const readabilityAnalysis = analyzeImagesInHtml(article.content);
const imageLossRatio = containerImageCount > 0
  ? (containerImageCount - readabilityAnalysis.count) / containerImageCount  // <- uses total count, not filtered count
  : 0;

If a page has 10 images but 5 are data URIs, readabilityAnalysis.count will be 10 but readabilityAnalysis.images.length will be 5. The image loss calculation will be wrong, potentially triggering the fallback incorrectly.

Fix: Return the filtered count instead:

return { count: images.length, images };

I've left an inline comment with a code suggestion on the specific line.

---

Architecture Notes (non-blocking)

The refactoring successfully:

• ✅ Eliminates redundant HTML parsing by doing single-pass image extraction
• ✅ Removes unnecessary abstraction layer (platform.js)
• ✅ Consolidates naming functions (generateFilename + generateSubfolder → generateBaseName)
• ✅ Fixes popup toggle synchronization bug
• ✅ Passes pre-analyzed images through extraction result

The code is well-structured and the performance optimization is solid. Once the count bug is fixed, this is ready to merge.

---

[claude]

Claude encountered an error —— View job

---

Re-reviewing PR Changes

Todo List:

• Check if critical image count bug was fixed
• Verify image loss ratio calculation is correct
• Scan for any new issues
• Provide final review

---

✅ Critical bug fixed! Line 91 now correctly returns count: images.length (filtered count)

[claude]

Claude encountered an error —— View job

---

I'll analyze this and get back to you.

[claude]

Claude finished @gupsammy's task —— View job

---

Re-reviewing PR After Latest Push

Todo List:

• Verify critical image count bug fix at line 91
• Review lib/ezycopy.js for other critical issues
• Check content-script.js integration
• Scan other changed files for bugs
• Provide final review verdict

---

Final Review: ✅ Ready to Merge

The critical bug has been fixed correctly. Line 91 in lib/ezycopy.js now returns count: images.length (the filtered count), which matches the actual images array length. This ensures the image loss ratio calculation at lines 335-337 works properly.

---

Verification Summary

Critical bug fix confirmed:

• ✅ analyzeImagesInHtml() at line 91 now returns filtered count
• ✅ Image loss ratio calculation will work correctly
• ✅ Fallback logic will trigger appropriately

Architecture review:

• ✅ Single-pass image extraction eliminates redundant parsing
• ✅ Pre-analyzed images correctly consumed in content-script.js:77-92
• ✅ Direct EzyCopyFiles usage (platform.js removed) works fine
• ✅ Popup toggle synchronization fixed (lines 52-54, 64-67 in popup.js)
• ✅ All integration points check out

No critical/high issues found. The refactoring achieves its performance goals without introducing bugs or Chrome Extension API violations.

---