feat: update remove overlap

opendatalab · Apr 28, 2024 · 6a3d1f2 · 6a3d1f2
1 parent 96d17cb
commit 6a3d1f2
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 5 deletions.
diff --git a/magic_pdf/model/magic_model.py b/magic_pdf/model/magic_model.py
@@ -461,9 +461,6 @@ def __get_blocks_by_type(
                     blocks.append(block)
         return blocks
 
-    def get_model_list(self, page_no):
-        return self.__model_list[page_no]
-
 
 if __name__ == "__main__":
     drw = DiskReaderWriter(r"D:/project/20231108code-clean")

diff --git a/magic_pdf/pre_proc/remove_bbox_overlap.py b/magic_pdf/pre_proc/remove_bbox_overlap.py
@@ -3,9 +3,21 @@
 
 def _remove_overlap_between_bbox(spans):
     res = []
-    for v in spans:
+
+    keeps = [True] * len(spans)
+    for i in range(len(spans)):
+        for j in range(len(spans)):
+            if i == j:
+                continue
+            if _is_in(spans[i]["bbox"], spans[j]["bbox"]):
+                keeps[i] = False
+
+    for idx, v in enumerate(spans):
+        if not keeps[idx]:
+            continue
+
         for i in range(len(res)):
-            if _is_in(res[i]["bbox"], v["bbox"]) or _is_in(v["bbox"], res[i]["bbox"]):
+            if  _is_in(v["bbox"], res[i]["bbox"]):
                 continue
             if _is_in_or_part_overlap(res[i]["bbox"], v["bbox"]):
                 ix0, iy0, ix1, iy1 = res[i]["bbox"]