[reading order] Improve reading order and add exporters#2116
Open
felixdittrich92 wants to merge 13 commits into
Open
[reading order] Improve reading order and add exporters#2116felixdittrich92 wants to merge 13 commits into
felixdittrich92 wants to merge 13 commits into
Conversation
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #2116 +/- ##
==========================================
- Coverage 96.61% 96.08% -0.53%
==========================================
Files 168 169 +1
Lines 9094 9533 +439
==========================================
+ Hits 8786 9160 +374
- Misses 308 373 +65
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
felixdittrich92
added
topic: documentation
felixdittrich92
changed the title
There was a problem hiding this comment.
Pull request overview
This PR improves DocTR’s reading-order handling (including memory optimizations), introduces a dedicated doctr.io.exporters module with new Markdown/AsciiDoc/HTML exporters, and updates the builder/predictor paths and docs/tests to support these new export surfaces.
Changes:
- Refactors export logic into
doctr/io/exporters.py, adds Markdown/AsciiDoc/HTML exporters, and wires export mixins into core element classes. - Updates reading-order computation (DAG traversal + overlap ratio memory behavior) and adds extensive regression tests for reading-order and exporters.
- Improves table-region cropping by adding padding and adjusts table word ordering / geometry handling in the builder.
Reviewed changes
Copilot reviewed 12 out of 12 changed files in this pull request and generated 4 comments.
Show a summary per file
| File | Description |
|---|---|
| tests/common/test_models_reading_order.py | Adds regression tests and helpers for reading-order edge cases (columns, key/value rows). |
| tests/common/test_models_builder.py | Extends builder repr expectations and adds tests for keep_reading_order. |
| tests/common/test_io_exporters.py | New test suite covering page/document export to Markdown/AsciiDoc/HTML/XML and reading-order behavior. |
| doctr/models/reading_order/base.py | Optimizes overlap computations and refines topological reading-order traversal (incl. multi-column heuristic). |
| doctr/models/predictor/pytorch.py | Adds padding to table-region crops to improve downstream table recognition. |
| doctr/models/layout/lw_detr/pytorch.py | Updates layout model class list (removes checkbox classes). |
| doctr/models/builder.py | Adds keep_reading_order, improves rotated-page handling in line resolution, and refines table cell word ordering. |
| doctr/io/exporters.py | New module implementing exporters, reading-order caching, and export mixins. |
| doctr/io/elements.py | Moves export/render methods to mixins (delegating to doctr.io.exporters). |
| doctr/io/init.py | Re-exports new exporters in the doctr.io package namespace. |
| docs/source/using_doctr/using_models.rst | Documents new export APIs and keep_reading_order. |
| docs/source/modules/io.rst | Adds API docs for new export methods and exporter classes. |
Comments suppressed due to low confidence (1)
doctr/io/exporters.py:930
- Same issue as
export_as_markdown: document-level calls can reasonably passinclude_furniture, which currently breaks on KIE pages. Accept/ignore it here as well to avoidTypeErrorwhen exporting mixed document types with shared helpers.
def export_as_asciidoc(self, direction: str = "auto", escape: bool = True) -> str:
"""Export the KIE page as AsciiDoc, with the predictions of each class sorted in reading order.
Args:
direction: reading direction, one of 'auto', 'ltr', 'rtl', 'ttb-rtl' or 'ttb-ltr'
escape: whether the characters and line markers carrying a structural meaning in AsciiDoc should
be escaped
Returns:
an AsciiDoc string with one section per detection class
"""
return AsciiDocExporter().export_kie_page(cast("KIEPage", self), direction=direction, escape=escape)
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+906
to
+917
| def export_as_markdown(self, direction: str = "auto", escape: bool = True) -> str: | ||
| """Export the KIE page as Markdown, with the predictions of each class sorted in reading order. | ||
|
|
||
| Args: | ||
| direction: reading direction, one of 'auto', 'ltr', 'rtl', 'ttb-rtl' or 'ttb-ltr' | ||
| escape: whether the characters carrying a structural meaning in Markdown should be escaped | ||
|
|
||
| Returns: | ||
| a Markdown string with one section per detection class | ||
| """ | ||
| return MarkdownExporter().export_kie_page(cast("KIEPage", self), direction=direction, escape=escape) | ||
|
|
Comment on lines
+931
to
+934
| def export_as_html(self, direction: str = "auto") -> str: | ||
| """Export the KIE page as semantic HTML, with the predictions of each class sorted in reading order""" | ||
| return HTMLExporter().export_kie_page(cast("KIEPage", self), direction=direction) | ||
|
|
Comment on lines
+631
to
+636
| To export the output in reading order as Markdown, AsciiDoc, HTML or XML, you can use the `export_as_markdown`, | ||
| `export_as_asciidoc`, `export_as_html` and `export_as_xml` methods (also available on a single :class:`Page`). | ||
| Reading order is always applied by these exporters: the content is linearized column by column, the reading direction is inferred from the | ||
| recognized text (e.g. right-to-left for Arabic or Hebrew documents), and - when the predictor is run with | ||
| layout detection (``detect_layout=True``) - the layout regions are used to render headings, list items and | ||
| recognized tables, and to place the page furniture (headers, footers, footnotes): |
Comment on lines
+149
to
+151
| The reading-order-aware export of a :class:`Document` / :class:`Page` to Markdown, AsciiDoc is available | ||
| through the ``export_as_markdown`` / ``export_as_asciidoc`` / ``export_as_html`` / ``export_as_xml`` / ``export_as`` / ``export`` / ``render`` methods documented above, which | ||
| delegate to the exporters of :mod:`doctr.io.exporters`. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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:
Closes: #1586 #2085