Merge pull request #1534 from myhloli/dev

Update pdf_parse_union_core_v2.py

Merge pull request #1534 from myhloli/dev
Update pdf_parse_union_core_v2.py
1b0ef29a · Xiaomeng Zhao · GitHub · 2c64ff3d · 9f12c398 · 1b0ef29a
Unverified Commit 1b0ef29a authored Jan 14, 2025 by Xiaomeng Zhao Committed by GitHub Jan 14, 2025
Hide whitespace changes
Inline Side-by-side

Showing with 5 additions and 1 deletion

magic_pdf/pdf_parse_union_core_v2.py magic_pdf/pdf_parse_union_core_v2.py +5 -1

No files found.
--- a/magic_pdf/pdf_parse_union_core_v2.py
+++ b/magic_pdf/pdf_parse_union_core_v2.py
@@ -178,7 +178,11 @@ def txt_spans_extract_v2(pdf_page, spans, all_bboxes, all_discarded_blocks, lang
    # text_blocks_raw = pdf_page.get_text('rawdict', flags=fitz.TEXT_PRESERVE_WHITESPACE | fitz.TEXT_MEDIABOX_CLIP)['blocks']

    # cid用0xfffd表示，连字符不拆开
-    text_blocks_raw = pdf_page.get_text('rawdict', flags=fitz.TEXT_PRESERVE_LIGATURES | fitz.TEXT_PRESERVE_WHITESPACE | fitz.TEXT_MEDIABOX_CLIP)['blocks']
+    #text_blocks_raw = pdf_page.get_text('rawdict', flags=fitz.TEXT_PRESERVE_LIGATURES | fitz.TEXT_PRESERVE_WHITESPACE | fitz.TEXT_MEDIABOX_CLIP)['blocks']
+
+    # 自定义flags出现较多0xfffd，可能是pymupdf可以自行处理内置字典的pdf，不再使用
+    text_blocks_raw = pdf_page.get_text('rawdict', flags=fitz.TEXTFLAGS_TEXT)['blocks']
+    # text_blocks = pdf_page.get_text('dict', flags=fitz.TEXTFLAGS_TEXT)['blocks']
    all_pymu_chars = []
    for block in text_blocks_raw:
        for line in block['lines']: