Split the 1288-line info.py into smaller, single-responsibility modules:
- _types.py: Enums, type aliases, lookup dictionaries
- _contentstream.py: PDF content stream parsing, DPI calculation
- _image.py: ImageInfo class and image finding functions
- _worker.py: Concurrency/worker process handling
- info.py: PageInfo, PdfInfo classes (reduced to ~530 lines)
Public API unchanged - all existing imports continue to work.
Comprehensive test coverage for the new hocrtransform components:
- test_ocr_element.py: Tests for BoundingBox, Baseline, FontInfo,
OcrElement dataclass methods (iter_by_class, find_by_class,
get_text_recursive, words/lines/paragraphs properties)
- test_hocr_parser.py: Tests for parsing hOCR files including
page/paragraph/line/word extraction, RTL text, rotated text,
different line types (header, caption), font info, and edge cases
- test_pdf_renderer.py: Tests for PDF rendering including text
extraction verification, page sizing, multi-line content,
text direction, baseline handling, textangle rotation, word breaks,
debug options, and image overlay
Also fixes x_font regex pattern to not capture trailing semicolons.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Added use_cropbox parameter to rasterize_pdf_page hook to allow
choosing between MediaBox and CropBox rendering:
- Default is use_cropbox=False (MediaBox) for consistency with
Ghostscript's existing behavior
- Ghostscript: passes -dUseCropBox when use_cropbox=True
- pypdfium: calculates crop values to expand from CropBox to MediaBox
when use_cropbox=False
This aligns both rasterizers to produce the same output dimensions
by default, making the rasterizer choice transparent for page
geometry.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
pypdfium2/PDFium is not thread-safe - concurrent calls from different
threads can crash or corrupt the process. Added a module-level lock to
serialize all pdfium operations.
PIL image processing and file I/O are done outside the lock since they
are thread-safe, minimizing lock contention.
For maximum parallelism, users can use process-based parallelism
(use_threads=False) where each process has its own pdfium instance.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add test_rasterizer.py with tests covering:
- Basic rasterizer option validation ('auto', 'ghostscript', 'pypdfium')
- Rasterizer + --rotate-pages interaction
- PDFs with nonstandard MediaBox/TrimBox/CropBox
- Direct hook tests verifying plugins respect the option
Also fix pluggy parameter passing: make 'options' a required parameter
(no default) in the hookspec so pluggy forwards it to implementations.
Update test plugins and test_rotation.py to pass the new parameter.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Completes Phase 5 of the CLI refactoring plan by enabling nested
plugin option access (e.g., options.tesseract.timeout) alongside
the legacy flat access (options.tesseract_timeout).
Changes:
- Add module-level plugin option model registry in _options.py
- Add __getattr__ to OCROptions for dynamic namespace access
- Register plugin models in setup_plugin_infrastructure()
- Add test for nested plugin option access
Plugin option instances are lazily created from flat field values
and cached for subsequent access.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Instead the standard executor will fall back to threads.
semfree caused test failures with Py3.14:
https://github.com/ocrmypdf/OCRmyPDF/issues/1558
In retrospect and with emerging Python tech like freethreading, semfree is becoming less necessary. We can use threads for the time being.
A consequence is that performance may be lower on Lambda and Termux when we are using threads and not shelling out work.
strip_invisible_text resets the text render mode on each `BT` (begin text) command. However the text state is not actually reset for each text element, only for each page.
The pdf reference says:
> The text state operators can appear outside text objects, and the values they set
> are retained across text objects in a single content stream. Like other graphics
> state parameters, these parameters are initialized to their default values at the
> beginning of each page.
>
> -- https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandards/pdfreference1.7old.pdf#page=397
With the current implementation, a text object is only deleted if it contains a `3 Tr` command (setting the text rendering mode to invalid). However the rendering mode may be set once and then not changed for multiple text objects or set outside of a text object.
In that case only the first text object (which contains the `3 Tr`-command) is removed. This not only leaves the other text objects in the pdf, but also makes them visible, since the text object that contained the `3 Tr`-command is removed.
This PR updates `strip_invisible_text` to not reset the rendering mode for each object and to keep track of the rendering mode when the graphic state is pushed/popped.
Before 42ff7fc842, `make_rotate_test`
always used `resources / 'typewriter.png'`, but after the change the
second call accidentally used just `resources`, which is a directory,
and fails to open.
James R. Barlow
Kara Engelhardt
Elliott Sales de AndradeandGitHub