fix: address copilot and coderabbit review feedback

- Use named returns for IsAlreadyConverted and WriteChapterToCBZ so
  errs.Capture defers propagate close errors to the caller.
- Add context.Context to IsAlreadyConverted and ExtractChapter for
  cancellation/timeout support; thread it from Optimize.
- Filter non-image files (__MACOSX, Thumbs.db, .DS_Store, etc.) and
  only extract supported image extensions in ExtractChapter.
- Validate TempDir is non-empty in ConvertChapter before creating output dir.
- Separate fatal conversion errors from PageIgnoredError in the WebP
  converter result aggregation so real failures aren't masked.
- Fix error message "failed to load chapter" → "failed to extract chapter".
- Use %w instead of %v in fmt.Errorf for proper error wrapping.
- Fix LoadChapter doc comment to match its actual behavior.
- Extract shared parseConvertedComment/parseConvertedCommentTime helpers
  to avoid duplicated zip-comment parsing logic.
- Convert tests to use testify assertions (assert/require) per guidelines.

Co-authored-by: Belphemur <197810+Belphemur@users.noreply.github.com>

internal/cbz/cbz_creator.go

@@ -14,7 +14,7 @@ import (
 
 // WriteChapterToCBZ creates a CBZ file from a Chapter by streaming page files
 // from disk directly into the zip archive. No image data is held in memory.
-func WriteChapterToCBZ(chapter *manga.Chapter, outputFilePath string) error {
+func WriteChapterToCBZ(chapter *manga.Chapter, outputFilePath string) (err error) {
 	log.Debug().
 		Str("chapter_file", chapter.FilePath).
 		Str("output_path", outputFilePath).