Fix pytorch bug

There was a bug with pytorch 2.6 and MPS that caused errors in inference - this has been fixed.

New LaTeX OCR model; block visualizer; better links/references

Improved LaTeX OCR

We trained a new LaTeX OCR model that works a lot better overall. It will reliably output KaTeX-compatible math. It also operates on longer sequences than before.

The rendered output is on the right, original document on the left:

Block visualization

You can now visualize blocks in the streamlit app, thanks to @jazzido . By selecting json output and checking "show blocks", you get a nice visualization where you can see how marker parsed the page. Clicking on blocks will show the HTML.

Links and references

We fixed a bug with links and references, they now render as one block. You can see the extracted references here:

Misc bugfixes

• Fixed some bugs with tables and row splitting
• Escaped $ inside text and tables so we don't accidentally render things as equations

What's Changed

• [streamlit_app] Visualize extracted blocks by @jazzido in #502
• Texify by @VikParuchuri in #513
• Update texify by @VikParuchuri in #514

New Contributors

• @jazzido made their first contribution in #502

Full Changelog: v1.3.2...v1.3.3

Fix table bugs

• Issue where some blocks were hidden when they were around tables
• Fix span id issue with --use_llm and tables
• Fix problem with tables not OCRing when needed

Improved equations, bugfixes

• Equations in tables now render properly with --use_llm
• Fix how block equations render
• Fix bug with markdown table rendering and --use_llm
• Fix bug with convert.py CLI script

Improved tables; links and references

Table improvements

• Tables now handle colspans and rowspans properly
• Improved table model with better accuracy
• Tables merge across pages if you pass --use_llm
• New table benchmarks

Links and references

• Links and references are now pulled out of the pdf, and are clickable
• Anchors are placed on elements as targets

Better configuration

• Any configuration option can now be passed on the CLI

Misc

• With --use_llm, handwriting is now recognized (if the layout detects it)
• Better llm mode overall

What's Changed

• Dynamic option parsing and updated documentation by @iammosespaulr in #457
• Add context manager support for PdfProvider by @tarun-menta in #481
• Table benchmarks by @tarun-menta in #480
• PDF Link and References Support by @iammosespaulr in #474
• Improved References by @iammosespaulr in #491
• New table model by @VikParuchuri in #472
• Vik dev by @VikParuchuri in #497
• OCR Error Detection Hotfix by @tarun-menta in #498
• Vik dev by @VikParuchuri in #500
• Dev by @VikParuchuri in #499
• Table improvements by @VikParuchuri in #503

Full Changelog: v1.2.7...v1.3.0

Remove code from new version

Remove some code that came from the dev branch in the cli scripts.

Reorganize imports

Fix issue with needing server dependencies to run other CLI scripts.

Hotfix scripts

CLI scripts were broken on some systems with 1.2.4. This fixes it.

Fix section header bug

Fix a bug with nested section headers.

Fix math delimiter issue

Handle mismatched delimiters.