Extending the usage of BoundingRectangle #346
Closed
+3,024
−213
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR introduces support for representing bounding boxes using 4-vertex convex polygons instead of axis-aligned rectangles. This allows for more accurate localization of elements within a document, especially in cases where the document is photographed under imperfect conditions, such as misalignment, bending, or stretching of the paper.
While the class name BoundingRectangle is currently used, it may be misleading and should potentially be renamed to better reflect the generalized shape (e.g., BoundingPolygon).
To support this new structure in SmolDocling, a dedicated localization token <rec_ has been added.
Significant effort has gone into ensuring backward compatibility within the Prov class and related methods, so existing workflows remain unaffected.