Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add Inline Math Support #517

Open
wants to merge 26 commits into
base: dev
Choose a base branch
from
Open

Add Inline Math Support #517

wants to merge 26 commits into from

Conversation

tarun-menta
Copy link
Collaborator

Inline math is now detected from surya and spliced into the existing provider lines

Complete refactor of OCRBuilder and LayoutBuilder, OCR decisions are now made per-line, instead of for a full page.

VikParuchuri
VikParuchuri requested changes Jan 30, 2025
View reviewed changes
Copy link
Owner

@VikParuchuri VikParuchuri left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  • High level, I think an approach where we don't combine all the logic into the OCRBuilder is more maintainable (see comments)
  • Needs tests

marker/processors/equation.py Show resolved Hide resolved
marker/builders/ocr.py Show resolved Hide resolved
marker/util.py Outdated Show resolved Hide resolved
@tarun-menta
Update hashing
6eb7e83
@tarun-menta
Refactor fully
834c097 
All line detection and splitting is now done in a separate LineBuilder
class. Lines which require OCR are later picked up by OCRBuilder
@tarun-menta
Cleanup
b075c9d
@tarun-menta
Refactor splitting logic and add support for character level splitting
fd351dd 
Character level splitting still not working perfectly
@tarun-menta
Final refactor - Remove char processing logic
cdd30ab
@tarun-menta
Update test config with new model
1ed58bf
@tarun-menta
Cleanup
34381e6
@tarun-menta
Char level splicing of inline math into provider lines
ce867b8
@tarun-menta
Merge remote-tracking branch 'origin/texify' into inline-math
af2c09e
@tarun-menta
Update equation processor for inline
1871067
@tarun-menta
Update OCR and Line Builders
96c95a2
@tarun-menta
Better thresholds
ddd0616
@tarun-menta
Fix tests; Include new builder in args
9cbaae7
@tarun-menta
Fix OCR bug - Don't sort lines in RecognitionPredictor
92e75f9
@tarun-menta
Fix line merging
a2457e4
@tarun-menta
Update text_extraction_method for blocks in OCRBuilder
861fa91
@tarun-menta
Fix OCR+Provider line merging for rotated PDFs
ffdad8c
@tarun-menta
Set page extraction method in LineBuilder
43ddda5
@tarun-menta
Copy link
Collaborator Author

Fails two tests - test_garbed_pdf.py and test_rotated_bboxes.py, but the markdown and HTML outputs are the same as before after a diff check with dev.

@tarun-menta
Add flag for inline math instead of running every time
f3d5e5b
@tarun-menta
Update test
82668f3
@tarun-menta
Fix tests
bbdae4b 
Line counts changed because of different behavior of new line det model
Picks up some lines within figures which weren't before, hence the
increased count. Outputs are still unchanged
@tarun-menta
Update to new inline detection API
f0e4745
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@VikParuchuri VikParuchuri Awaiting requested review from VikParuchuri

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@tarun-menta @VikParuchuri