Merge pull request #50 from myhloli/master

统一使用ocr组装markdown
opendatalab · Apr 22, 2024 · d95f545 · d95f545
2 parents f6d8f6c + 1340a97
commit d95f545
Showing 1 changed file with 3 additions and 2 deletions.
diff --git a/magic_pdf/pipe/AbsPipe.py b/magic_pdf/pipe/AbsPipe.py
@@ -106,8 +106,9 @@ def mk_markdown(compressed_pdf_mid_data: str, img_buket_path: str) -> list:
         parse_type = pdf_mid_data["_parse_type"]
         pdf_info_list = pdf_mid_data["pdf_info"]
         if parse_type == AbsPipe.PIP_TXT:
-            content_list = mk_universal_format(pdf_info_list, img_buket_path)
-            md_content = mk_mm_markdown(content_list)
+            # content_list = mk_universal_format(pdf_info_list, img_buket_path)
+            # md_content = mk_mm_markdown(content_list)
+            md_content = ocr_mk_mm_markdown_with_para(pdf_info_list, img_buket_path)
         elif parse_type == AbsPipe.PIP_OCR:
             md_content = ocr_mk_mm_markdown_with_para(pdf_info_list, img_buket_path)
         return md_content