We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent 4b74228 commit 0f1da38Copy full SHA for 0f1da38
1 file changed
docling_eval/dataset_builders/doclingsdg_builder.py
@@ -52,6 +52,7 @@
52
_MIN_PAGE_IMAGE_DIM = 32
53
_MAX_PAGE_IMAGE_DIM = 4096
54
_REGION_OVERLAP_IOU_THRESHOLD = 0.99
55
+_SKIP_ROTATED_90_TABLES = True
56
57
_TABLE_REGION_CATEGORY_IDS: Dict[str, int] = {
58
"table": 1,
@@ -1736,6 +1737,16 @@ def iterate(self) -> Iterable[DatasetRecordWithBBox]:
1736
1737
)
1738
1739
if has_rotated_90:
1740
+ if _SKIP_ROTATED_90_TABLES:
1741
+ skipped_filtered += 1
1742
+ filter_reason_counts["rotated_90_table"] = (
1743
+ filter_reason_counts.get("rotated_90_table", 0) + 1
1744
+ )
1745
+ _log.warning(
1746
+ "Skipping table sample %s: 90-degree rotated table detected.",
1747
+ doc_id,
1748
1749
+ continue
1750
tags.append("90_degree")
1751
else:
1752
ground_truth_bboxes = self._extract_top_level_bboxes(
0 commit comments