magic-pdf 1.2.1__py3-none-any.whl → 1.2.2__py3-none-any.whl

magic_pdf/libs/version.py

@@ -1 +1 @@
-__version__ = "1.2.1"
+__version__ = "1.2.2"

magic_pdf/post_proc/para_split_v3.py

@@ -108,29 +108,32 @@ def __is_list_or_index_block(block):
         ):
             multiple_para_flag = True
 
-        for line in block['lines']:
-            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] > 0.7 * 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:
-                center_close_num += 1
+        block_text = ''
 
+        for line in block['lines']:
             line_text = ''
 
             for span in line['spans']:
                 span_type = span['type']
                 if span_type == ContentType.Text:
                     line_text += span['content'].strip()
-
             # 添加所有文本，包括空行，保持与block['lines']长度一致
             lines_text_list.append(line_text)
             block_text = ''.join(lines_text_list)
-            block_lang = detect_lang(block_text)
-            # logger.info(f"block_lang: {block_lang}")
+
+        block_lang = detect_lang(block_text)
+        # logger.info(f"block_lang: {block_lang}")
+
+        for line in block['lines']:
+            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] > 0.7 * 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:
+                center_close_num += 1
 
             # 计算line左侧顶格数量是否大于2，是否顶格用abs(block['bbox_fs'][0] - line['bbox'][0]) < line_height/2 来判断
             if abs(block['bbox_fs'][0] - line['bbox'][0]) < line_height / 2:

{magic_pdf-1.2.1.dist-info → magic_pdf-1.2.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: magic-pdf
-Version: 1.2.1
+Version: 1.2.2
 Summary: A practical tool for converting PDF to Markdown
 Home-page: https://github.com/opendatalab/MinerU
 Requires-Python: >=3.9
@@ -9,7 +9,7 @@ License-File: LICENSE.md
 Requires-Dist: boto3>=1.28.43
 Requires-Dist: Brotli>=1.1.0
 Requires-Dist: click>=8.1.7
-Requires-Dist: fast-langdetect>=0.2.3
+Requires-Dist: fast-langdetect<0.3.0,>=0.2.3
 Requires-Dist: loguru>=0.6.0
 Requires-Dist: numpy<2.0.0,>=1.21.6
 Requires-Dist: pydantic>=2.7.2

{magic_pdf-1.2.1.dist-info → magic_pdf-1.2.2.dist-info}/RECORD

@@ -51,7 +51,7 @@ magic_pdf/libs/pdf_check.py,sha256=7GWWvDR6g_rj_fE6XJlbTq5AFVX11ngRIzT0N18F214,3
 magic_pdf/libs/pdf_image_tools.py,sha256=kjzSEbm7K0yiHv8kJ4VbZ9HHktM8qvAv3LhxRyDZEQk,1987
 magic_pdf/libs/performance_stats.py,sha256=BFi4NIsUYlanznYoTVq4hBpj4NOuShAlWBHzebBGVYM,1702
 magic_pdf/libs/safe_filename.py,sha256=ckwcM_eqoysTb5id8czp-tXq2G9da0-l3pshZDCHQtE,236
-magic_pdf/libs/version.py,sha256=Mlm4Gvmb_6yQxwUbv2Ksc-BJFXLPg9H1Vt2iV7wXrA4,22
+magic_pdf/libs/version.py,sha256=uuf4VNtTNA93fMhoAur9YafzaKJFnczY-H1SSCSuRVQ,22
 magic_pdf/model/__init__.py,sha256=sa-dO2k-TLy25I2gRrzjm_cQeYfzMf-pLwBJHkIxGo0,51
 magic_pdf/model/batch_analyze.py,sha256=sbrgOJWycb1Ep6e62CPi6jEyG6VSeklIxc4PmrqaLhM,11933
 magic_pdf/model/doc_analyze_by_custom_model.py,sha256=T0-h4QmSIDXRzgF5uWO4jQrwIot221l26PXU52xeKiA,7933
@@ -117,7 +117,7 @@ magic_pdf/operators/models.py,sha256=mRqbCVrxxaUVDpEBAsXaK7EL1M-goICkE1W0FYgewio
 magic_pdf/operators/pipes.py,sha256=XgBgisKQd_ruW-3Tw4v5LhqloZUHgn2aFcpi_q8LbCs,6767
 magic_pdf/post_proc/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
 magic_pdf/post_proc/llm_aided.py,sha256=yzhu2cCpUZjdwf3v0swYDgSs9VWIfMAoXepYIP1EMZs,6367
-magic_pdf/post_proc/para_split_v3.py,sha256=v4SdQn4OZdHRXpWQMfQ-FGJz_tglQ88uFUqpwY542Fo,16922
+magic_pdf/post_proc/para_split_v3.py,sha256=SPN_VVGvFX5KpFMGw9OzgoE-kTZq-FF036i0cIImGH8,16975
 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=NDzbxwD7z7Tb4uAxL4KR6LzURFdN1Tzr4nPvj-VmEqc,1225
@@ -139,9 +139,9 @@ magic_pdf/tools/common.py,sha256=1LfMeXBBsb3WlGeNAze_pPOYXQ8Qbfh-JgRXweojHKo,838
 magic_pdf/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 magic_pdf/utils/annotations.py,sha256=82ou3uELNbQWa9hOFFkVt0gsIskAKf5msCv5J2IJ5V0,211
 magic_pdf/utils/office_to_pdf.py,sha256=7aj-Ls2v8saD-Rgu_t3FIc-J3Ka9wnmiEH5zY-H1Vxs,729
-magic_pdf-1.2.1.dist-info/LICENSE.md,sha256=jVa0BUaKrRH4erV2P5AeJ24I2WRv9chIGxditreJ6e0,34524
-magic_pdf-1.2.1.dist-info/METADATA,sha256=oMxODNFW4g154Rrh9g2sE9_irmB6x7j_5KmF7dRYEPQ,40994
-magic_pdf-1.2.1.dist-info/WHEEL,sha256=tZoeGjtWxWRfdplE7E3d45VPlLNQnvbKiYnx7gwAy8A,92
-magic_pdf-1.2.1.dist-info/entry_points.txt,sha256=wXwYke3j8fqDQTocUspL-CqDUEv3Tfcwp09fM8dZAhA,98
-magic_pdf-1.2.1.dist-info/top_level.txt,sha256=J9I0AzmHWGkp9c6DL8Oe4mEx3yYphLzkRn4H25Lg1rE,10
-magic_pdf-1.2.1.dist-info/RECORD,,
+magic_pdf-1.2.2.dist-info/LICENSE.md,sha256=jVa0BUaKrRH4erV2P5AeJ24I2WRv9chIGxditreJ6e0,34524
+magic_pdf-1.2.2.dist-info/METADATA,sha256=FYzj0yWzmFAG4mQ22DH9F4KZfqexNg7YuhgiXMHc9Ug,41001
+magic_pdf-1.2.2.dist-info/WHEEL,sha256=tZoeGjtWxWRfdplE7E3d45VPlLNQnvbKiYnx7gwAy8A,92
+magic_pdf-1.2.2.dist-info/entry_points.txt,sha256=wXwYke3j8fqDQTocUspL-CqDUEv3Tfcwp09fM8dZAhA,98
+magic_pdf-1.2.2.dist-info/top_level.txt,sha256=J9I0AzmHWGkp9c6DL8Oe4mEx3yYphLzkRn4H25Lg1rE,10
+magic_pdf-1.2.2.dist-info/RECORD,,