magic-pdf 0.10.3__py3-none-any.whl → 0.10.5__py3-none-any.whl

magic_pdf/data/utils.py

@@ -20,8 +20,8 @@ def fitz_doc_to_image(doc, dpi=200) -> dict:
     mat = fitz.Matrix(dpi / 72, dpi / 72)
     pm = doc.get_pixmap(matrix=mat, alpha=False)
 
-    # If the width or height exceeds 9000 after scaling, do not scale further.
-    if pm.width > 9000 or pm.height > 9000:
+    # If the width or height exceeds 4500 after scaling, do not scale further.
+    if pm.width > 4500 or pm.height > 4500:
         pm = doc.get_pixmap(matrix=fitz.Matrix(1, 1), alpha=False)
 
     img = Image.frombytes('RGB', (pm.width, pm.height), pm.samples)

magic_pdf/dict2md/ocr_mkcontent.py

@@ -5,6 +5,7 @@ from loguru import logger
 from magic_pdf.config.make_content_config import DropMode, MakeMode
 from magic_pdf.config.ocr_content_type import BlockType, ContentType
 from magic_pdf.libs.commons import join_path
+from magic_pdf.libs.language import detect_lang
 from magic_pdf.libs.markdown_utils import ocr_escape_special_markdown_char
 from magic_pdf.para.para_split_v3 import ListLineTag
 
@@ -135,18 +136,19 @@ def __replace_ligatures(text: str):
 
 
 def merge_para_with_text(para_block):
+    block_text = ''
+    for line in para_block['lines']:
+        for span in line['spans']:
+            if span['type'] in [ContentType.Text]:
+                block_text += span['content']
+    block_lang = detect_lang(block_text)
+
     para_text = ''
     for i, line in enumerate(para_block['lines']):
 
         if i >= 1 and line.get(ListLineTag.IS_LIST_START_LINE, False):
             para_text += '  \n'
 
-        line_text = ''
-        for span in line['spans']:
-            span_type = span['type']
-            if span_type == ContentType.Text:
-                line_text += span['content'].strip()
-
         for j, span in enumerate(line['spans']):
 
             span_type = span['type']
@@ -159,15 +161,24 @@ def merge_para_with_text(para_block):
                 content = f"\n$$\n{span['content']}\n$$\n"
 
             content = content.strip()
-            if content != '':
-                if span_type in [ContentType.Text, ContentType.InlineEquation]:
-                    # 如果span是line的最后一个且末尾带有-连字符，那么末尾不应该加空格,同时应该把-删除
-                    if j == len(line['spans'])-1 and __is_hyphen_at_line_end(content):
-                        para_text += content[:-1]
-                    else:  # content间需要空格分隔
+
+            if content:
+                langs = ['zh', 'ja', 'ko']
+                # logger.info(f'block_lang: {block_lang}, content: {content}')
+                if block_lang in langs: # 中文/日语/韩文语境下，换行不需要空格分隔
+                    if j == len(line['spans']) - 1:
+                        para_text += content
+                    else:
                         para_text += f'{content} '
-                elif span_type == ContentType.InterlineEquation:
-                    para_text += content
+                else:
+                    if span_type in [ContentType.Text, ContentType.InlineEquation]:
+                        # 如果span是line的最后一个且末尾带有-连字符，那么末尾不应该加空格,同时应该把-删除
+                        if j == len(line['spans'])-1 and span_type == ContentType.Text and __is_hyphen_at_line_end(content):
+                            para_text += content[:-1]
+                        else:  # 西方文本语境下 content间需要空格分隔
+                            para_text += f'{content} '
+                    elif span_type == ContentType.InterlineEquation:
+                        para_text += content
             else:
                 continue
     # 连写字符拆分

magic_pdf/libs/version.py

@@ -1 +1 @@
-__version__ = "0.10.3"
+__version__ = "0.10.5"

magic_pdf/para/para_split_v3.py

@@ -112,8 +112,8 @@ def __is_list_or_index_block(block):
             line_mid_x = (line['bbox'][0] + line['bbox'][2]) / 2
             block_mid_x = (block['bbox_fs'][0] + block['bbox_fs'][2]) / 2
             if (
-                line['bbox'][0] - block['bbox_fs'][0] > 0.8 * line_height
-                and block['bbox_fs'][2] - line['bbox'][2] > 0.8 * line_height
+                line['bbox'][0] - block['bbox_fs'][0] > 0.7 * line_height
+                and block['bbox_fs'][2] - line['bbox'][2] > 0.7 * line_height
             ):
                 external_sides_not_close_num += 1
             if abs(line_mid_x - block_mid_x) < line_height / 2:

magic_pdf/pre_proc/cut_image.py

@@ -12,12 +12,12 @@ def ocr_cut_image_and_table(spans, page, page_id, pdf_bytes_md5, imageWriter):
     for span in spans:
         span_type = span['type']
         if span_type == ContentType.Image:
-            if not check_img_bbox(span['bbox']):
+            if not check_img_bbox(span['bbox']) or not imageWriter:
                 continue
             span['image_path'] = cut_image(span['bbox'], page_id, page, return_path=return_path('images'),
                                            imageWriter=imageWriter)
         elif span_type == ContentType.Table:
-            if not check_img_bbox(span['bbox']):
+            if not check_img_bbox(span['bbox']) or not imageWriter:
                 continue
             span['image_path'] = cut_image(span['bbox'], page_id, page, return_path=return_path('tables'),
                                            imageWriter=imageWriter)

magic_pdf/pre_proc/ocr_detect_all_bboxes.py

@@ -117,8 +117,8 @@ def ocr_prepare_bboxes_for_layout_split_v2(
     all_bboxes = remove_overlaps_min_blocks(all_bboxes)
     all_discarded_blocks = remove_overlaps_min_blocks(all_discarded_blocks)
     """将剩余的bbox做分离处理，防止后面分layout时出错"""
-    all_bboxes, drop_reasons = remove_overlap_between_bbox_for_block(all_bboxes)
-
+    # all_bboxes, drop_reasons = remove_overlap_between_bbox_for_block(all_bboxes)
+    all_bboxes.sort(key=lambda x: x[0]+x[1])
     return all_bboxes, all_discarded_blocks
 
 

{magic_pdf-0.10.3.dist-info → magic_pdf-0.10.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: magic-pdf
-Version: 0.10.3
+Version: 0.10.5
 Summary: A practical tool for converting PDF to Markdown
 Home-page: https://github.com/opendatalab/MinerU
 Requires-Python: >=3.9

{magic_pdf-0.10.3.dist-info → magic_pdf-0.10.5.dist-info}/RECORD

@@ -16,7 +16,7 @@ magic_pdf/data/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 magic_pdf/data/dataset.py,sha256=n8rGw1-wizABR8giSk_XWPCXzx3478u5DK2Z0wOCOeI,5089
 magic_pdf/data/read_api.py,sha256=hGpSVg9EcyM2mIlOsDIwsl7Y_ybWf9kkoxRumIXSzQQ,3566
 magic_pdf/data/schemas.py,sha256=oIUTBzK8Wq8Wuy8A_uilWAbVhucRvOs9_f3lSKYgcmQ,664
-magic_pdf/data/utils.py,sha256=dJZiqygwNier0UG5tbt5jAPjgwcnfsAN6-m-G1kVPLQ,917
+magic_pdf/data/utils.py,sha256=uaSHprh80D_puPUmd1slQDoE4uecNn4zZMzYWY0-a-8,917
 magic_pdf/data/data_reader_writer/__init__.py,sha256=QtevUaeSivv9dQKi3Tomfn4Z0E4To0cB8qXTnglxaHc,705
 magic_pdf/data/data_reader_writer/base.py,sha256=gUrHCMTHYBrWpqgHdIc-hN7HHwUC2ApK_VXrDUrnfdg,1320
 magic_pdf/data/data_reader_writer/filebase.py,sha256=iVjBT1M_89F5HB2uMsBzLfCNhsTVnmurJk4VUuAA2tw,2111
@@ -27,7 +27,7 @@ magic_pdf/data/io/base.py,sha256=SqNQqe30ZvoVvg7GVv-hLMCjN6yBgDyQQWeLgGsTfhQ,111
 magic_pdf/data/io/http.py,sha256=XlKB0DNf4a_uUnfgcclvaaOtmE7lmddx0DnK8A-emAM,958
 magic_pdf/data/io/s3.py,sha256=hyA7sbNriQy64xd_uyJ7acN_oneQ1Pdmoc7_xcvkue8,3606
 magic_pdf/dict2md/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-magic_pdf/dict2md/ocr_mkcontent.py,sha256=hwcHTEx1tbIlM9ukmPBOAyH0G6rmbOTu87nVtZ1gE6k,12354
+magic_pdf/dict2md/ocr_mkcontent.py,sha256=pE-lEUsYAhZC3nSmbgYO42Kvk_bW8Ds-AL-QMPHFu8c,12941
 magic_pdf/filter/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 magic_pdf/filter/pdf_classify_by_type.py,sha256=spmDO-f2ihAl1d6-EP-j271Yi50oyu6mw4X2kRd_m0s,42320
 magic_pdf/filter/pdf_meta_scan.py,sha256=3ba7SxXu1z2r5N97Dxmp_L10Lo7llsrBlvtEAJeIJBQ,17403
@@ -53,7 +53,7 @@ magic_pdf/libs/path_utils.py,sha256=Hykw_l5CU736b2egHV9P7B-qh3QNKO4nZSGCbsi0Z8E,
 magic_pdf/libs/pdf_check.py,sha256=wCVOcwEPeMRcHW5OGN-GSQnPT5qNXUYHWWowoUknxF4,3178
 magic_pdf/libs/pdf_image_tools.py,sha256=kjzSEbm7K0yiHv8kJ4VbZ9HHktM8qvAv3LhxRyDZEQk,1987
 magic_pdf/libs/safe_filename.py,sha256=ckwcM_eqoysTb5id8czp-tXq2G9da0-l3pshZDCHQtE,236
-magic_pdf/libs/version.py,sha256=0C8KcY1dzs3hdkAre06v0NCQ0Uxcqv6g9a93bRcVLW0,23
+magic_pdf/libs/version.py,sha256=c61d5YjslqtpItkzB2NGlURm177H2racruHXV9G6u6s,23
 magic_pdf/model/__init__.py,sha256=1QcfMKET0xQhSaZMjNQHi_TjzSSDR6PI5mjkmaXHPe8,52
 magic_pdf/model/doc_analyze_by_custom_model.py,sha256=YZwlhIgidy1_MUyTM_MRSLfKR_rpi508Bra6Vpj8PJ4,7125
 magic_pdf/model/magic_model.py,sha256=ppMkMqtP7sKncHTZ2SbXuPOoR988iRPexBEMA6QeiIc,42208
@@ -110,7 +110,7 @@ magic_pdf/model/sub_modules/table/structeqtable/struct_eqtable.py,sha256=SrNPm-u
 magic_pdf/model/sub_modules/table/tablemaster/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 magic_pdf/model/sub_modules/table/tablemaster/tablemaster_paddle.py,sha256=QEQ-56AzoIAU7UWsEidWW_KDOY5r16qm2kSpox8cxq4,2755
 magic_pdf/para/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-magic_pdf/para/para_split_v3.py,sha256=UOQe0HUVX7FAlMbJp1OkGfdM7JECWeqscv3s8Hge7ps,16922
+magic_pdf/para/para_split_v3.py,sha256=v4SdQn4OZdHRXpWQMfQ-FGJz_tglQ88uFUqpwY542Fo,16922
 magic_pdf/pipe/AbsPipe.py,sha256=jPtAa0pz_vPddya3ZpUk6UrGqp8PcBdLONO1spzavQo,4371
 magic_pdf/pipe/OCRPipe.py,sha256=nuN-zpUzu--gyrC0_vsvvilAyK7Mp3Tom_UOnsur1ps,2158
 magic_pdf/pipe/TXTPipe.py,sha256=5OFo2e8U5Y24wJrFDEJghBDpklnKFEnzKTYVnnhQssE,2159
@@ -118,8 +118,8 @@ magic_pdf/pipe/UNIPipe.py,sha256=ik0xXPdsHo7Un0gFpLC5ul04BP3Omd2mp5gqem40deE,480
 magic_pdf/pipe/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 magic_pdf/pre_proc/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 magic_pdf/pre_proc/construct_page_dict.py,sha256=OFmq5XRKi6fYIo-lmGlL-NB16Sf0egzsfEx-fT2uYrc,660
-magic_pdf/pre_proc/cut_image.py,sha256=U-ttnl3lAhhmgtkR1GGyPAVm0i0-6VscXf3E2EDy3lE,1187
-magic_pdf/pre_proc/ocr_detect_all_bboxes.py,sha256=xQ1L6pwQjN4xBSKEXslheip1aMFaiB0grqlX3BF-kh0,9282
+magic_pdf/pre_proc/cut_image.py,sha256=NDzbxwD7z7Tb4uAxL4KR6LzURFdN1Tzr4nPvj-VmEqc,1225
+magic_pdf/pre_proc/ocr_detect_all_bboxes.py,sha256=nt88ttXCEI_1ihAF7HU15SQjwM69V-iJmk-L_nyzA6o,9328
 magic_pdf/pre_proc/ocr_dict_merge.py,sha256=Ycgz2whzotL7kwl0-mHNV48QOQ2j4tRXqLSQrJRojYg,4847
 magic_pdf/pre_proc/ocr_span_list_modify.py,sha256=a5OmIwtkXkz6fTQg6p8R-f1nA_w0rgMwKFQjfs_HwrE,2864
 magic_pdf/pre_proc/remove_bbox_overlap.py,sha256=8eXNdsz9s06LX0kS0AxYSkaY1tWQQMkJfVtVSdjTQNE,3090
@@ -139,9 +139,9 @@ magic_pdf/tools/cli_dev.py,sha256=3RbubfTIagWoFYdu8wSDanr-BJDjFGeDet55jTy7He0,39
 magic_pdf/tools/common.py,sha256=ILTv8YjnK-XTVV5nzak3Sm-EJJXjG1hJJghlYKgYVBQ,6809
 magic_pdf/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 magic_pdf/utils/annotations.py,sha256=82ou3uELNbQWa9hOFFkVt0gsIskAKf5msCv5J2IJ5V0,211
-magic_pdf-0.10.3.dist-info/LICENSE.md,sha256=jVa0BUaKrRH4erV2P5AeJ24I2WRv9chIGxditreJ6e0,34524
-magic_pdf-0.10.3.dist-info/METADATA,sha256=R86XDaSfj1tcu3etkvhQfg3FSoARv8mKW2KpwjsdqWs,36992
-magic_pdf-0.10.3.dist-info/WHEEL,sha256=tZoeGjtWxWRfdplE7E3d45VPlLNQnvbKiYnx7gwAy8A,92
-magic_pdf-0.10.3.dist-info/entry_points.txt,sha256=wXwYke3j8fqDQTocUspL-CqDUEv3Tfcwp09fM8dZAhA,98
-magic_pdf-0.10.3.dist-info/top_level.txt,sha256=J9I0AzmHWGkp9c6DL8Oe4mEx3yYphLzkRn4H25Lg1rE,10
-magic_pdf-0.10.3.dist-info/RECORD,,
+magic_pdf-0.10.5.dist-info/LICENSE.md,sha256=jVa0BUaKrRH4erV2P5AeJ24I2WRv9chIGxditreJ6e0,34524
+magic_pdf-0.10.5.dist-info/METADATA,sha256=TIb8C_MrpU0_XwZc2dLfKpH5wQtE8G8Q0w56OPWYG30,36992
+magic_pdf-0.10.5.dist-info/WHEEL,sha256=tZoeGjtWxWRfdplE7E3d45VPlLNQnvbKiYnx7gwAy8A,92
+magic_pdf-0.10.5.dist-info/entry_points.txt,sha256=wXwYke3j8fqDQTocUspL-CqDUEv3Tfcwp09fM8dZAhA,98
+magic_pdf-0.10.5.dist-info/top_level.txt,sha256=J9I0AzmHWGkp9c6DL8Oe4mEx3yYphLzkRn4H25Lg1rE,10
+magic_pdf-0.10.5.dist-info/RECORD,,