refactor: extract textFromContentStream into handler methods, remove dead code, add reference snapshot tests

[ymettier]

Summary

Refactor textFromContentStream (282 → 61 lines) by extracting 16 handler methods into a textRenderer struct. Each handler is under 50 lines per KISS.

Changes

• pdftotext: textRenderer struct with handler methods (handlePushCTM, handlePopCTM, handleConcatMatrix, handleSetFont, handleTextMove, handleSetTextMatrix, handleTextStar, handleQuoteSingle, handleQuoteDouble, handleShowText, renderTopString, etc.), parse6Numbers helper deduplicates cm/Tm matrix parsing, removed dead aw/ac parsing from "
• logger: removed unused FromCtx/WithCtx/ctxKey abstraction (dead code, only called in tests)
• main_test.go: extracted withArgs(t, args...) helper eliminating 57 lines of os.Args boilerplate across 19 tests
• Tests: 10 reference snapshot tests comparing PDF text extraction against expected output, 100% coverage in pdftotext package

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!