Merge pull request #743 from myhloli/para-split-v3

refactor(para_split_v3): merge list and index block detection
opendatalab · Oct 15, 2024 · 0d83fb7 · 0d83fb7
2 parents 702b6ac + 244b868
commit 0d83fb7
Show file tree

Hide file tree

Showing 3 changed files with 111 additions and 82 deletions.
diff --git a/magic_pdf/libs/draw_bbox.py b/magic_pdf/libs/draw_bbox.py
@@ -237,6 +237,8 @@ def get_span_info(span):
                 BlockType.Text,
                 BlockType.Title,
                 BlockType.InterlineEquation,
+                BlockType.List,
+                BlockType.Index,
             ]:
                 for line in block['lines']:
                     for span in line['spans']:

diff --git a/magic_pdf/para/para_split_v3.py b/magic_pdf/para/para_split_v3.py
diff --git a/magic_pdf/pdf_parse_union_core_v2.py b/magic_pdf/pdf_parse_union_core_v2.py
@@ -360,7 +360,7 @@ def parse_page_core(pdf_docs, magic_model, page_id, pdf_bytes_md5, imageWriter,
                                                need_drop, drop_reason)
 
     '''将span填入blocks中'''
-    block_with_spans, spans = fill_spans_in_blocks(all_bboxes, spans, 0.3)
+    block_with_spans, spans = fill_spans_in_blocks(all_bboxes, spans, 0.5)
 
     '''对block进行fix操作'''
     fix_blocks = fix_block_spans(block_with_spans, img_blocks, table_blocks)