DocWire SDK 2025.05.22 Released – A Big Leap for PDF, OCR, and Core Architecture

Hey everyone!

It’s been a while since our last post — and that’s because we’ve been deep in the engine room, rewriting, refactoring, and rebuilding. And now, we’re thrilled to share the results:

 DocWire SDK 2025.05.22 is now on GitHub

https://github.com/docwire/docwire/releases/tag/2025.05.22

This is not just a regular update. It’s one of the biggest internal shifts DocWire has seen to date — and it lays the foundation for where we’re headed next.

               What’s New in 2025.05.22

PDF Image Extraction + OCR

You can now extract images directly from PDF files and process them through downstream chain elements, including content type detection and OCR for embedded text. It’s a powerful step toward full-stack document intelligence.

HTML & Plain Text Writers Improved

• Image tags are now supported.
• Data URLs and OCR-derived text are embedded where applicable.

Smarter, Stronger Testing

• New automatic tests for image extraction & OCR.
• Windows-specific test discovery fixes (no more silent ctest misses).

• Improved CI structure: better reporting, tighter test execution.

      Internal Overhaul - Chain & Parser Refactors
  

A major internal modernization took place under the hood:

• Core Chain Elements were refactored for clarity and flexibility. Processing stages can now clearly continue, skip, or stop — making the flow easier to follow and debug.
• Parsers were unified under a cleaner structure by directly implementing ChainElement and improving data_source checks.
• Cleaner Tag Emission Logic : chain elements can now reprocess data more easily, enabling more advanced flows in the future.

• HTML components were reorganized for clarity and modularity.

                         Fixes
  

• Thread-safe parser MIME initialization

• Test discovery working on Windows with a custom main()

• Clean linking of docwire_html and mailio (v0.25.1 via vcpkg)

• PSTParser threading fix

                      Developer Summary (TL;DR)
  

• From PDFs, images now take flight,

• Through refined chains, data flows bright.

• With steadier tests and safer threads,

• DocWire advances, new paths it treads.

This release brings visual data, smoother architecture, and tighter CI discipline — all while staying true to our mission: giving C++ developers modern, robust tools for complex data extraction and AI-ready pipelines.

                    Try It Out

Install the latest release and check out the new features:

https://github.com/docwire/docwire/releases/tag/2025.05.22

Use case demos, test coverage, and upgrade guidance are included.

                !!! We Want to Hear From You !!!

• Found it useful? Drop a ⭐ on GitHub.
• See a bug? Open an issue.

• Have a wild idea? Let’s talk.

   Stay tuned — more frequent updates coming your way.
  

Happy hacking,
The DocWire Team