Merge remote-tracking branch 'origin/dev' into dev

magic_pdf/libs/version.py

-__version__ = "1.1.0"
+__version__ = "1.2.2"

magic_pdf/model/doc_analyze_by_custom_model.py

@@ -165,12 +165,14 @@ def doc_analyze(
         import torch_npu
         if torch_npu.npu.is_available():
             npu_support = True
+            torch.npu.set_compile_mode(jit_compile=False)
 
     if torch.cuda.is_available() and device != 'cpu' or npu_support:
         gpu_memory = int(os.getenv("VIRTUAL_VRAM_SIZE", round(get_vram(device))))
         if gpu_memory is not None and gpu_memory >= 8:
-
-            if gpu_memory >= 16:
+            if gpu_memory >= 20:
+                batch_ratio = 16
+            elif gpu_memory >= 15:
                 batch_ratio = 8
             elif gpu_memory >= 10:
                 batch_ratio = 4

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/pre_proc/ocr_dict_merge.py

@@ -62,7 +62,15 @@ def merge_spans_to_line(spans, threshold=0.6):
 
 def span_block_type_compatible(span_type, block_type):
     if span_type in [ContentType.Text, ContentType.InlineEquation]:
-        return block_type in [BlockType.Text, BlockType.Title, BlockType.ImageCaption, BlockType.ImageFootnote, BlockType.TableCaption, BlockType.TableFootnote]
+        return block_type in [
+            BlockType.Text,
+            BlockType.Title,
+            BlockType.ImageCaption,
+            BlockType.ImageFootnote,
+            BlockType.TableCaption,
+            BlockType.TableFootnote,
+            BlockType.Discarded
+        ]
     elif span_type == ContentType.InterlineEquation:
         return block_type in [BlockType.InterlineEquation, BlockType.Text]
     elif span_type == ContentType.Image:

requirements.txt

 boto3>=1.28.43
 Brotli>=1.1.0
 click>=8.1.7
-fast-langdetect>=0.2.3
+fast-langdetect>=0.2.3,<0.3.0
 loguru>=0.6.0
 numpy>=1.21.6,<2.0.0
 pydantic>=2.7.2

signatures/version1/cla.json

@@ -159,6 +159,30 @@
       "created_at": "2025-02-22T07:15:35Z",
       "repoId": 765083837,
       "pullRequestNo": 1743
+    },
+    {
+      "name": "nadahlberg",
+      "id": 58701810,
+      "comment_id": 2676309097,
+      "created_at": "2025-02-22T17:04:14Z",
+      "repoId": 765083837,
+      "pullRequestNo": 1748
+    },
+    {
+      "name": "BetterAndBetterII",
+      "id": 141388234,
+      "comment_id": 2680567709,
+      "created_at": "2025-02-25T05:21:05Z",
+      "repoId": 765083837,
+      "pullRequestNo": 1767
+    },
+    {
+      "name": "luckymore",
+      "id": 5390013,
+      "comment_id": 2684392503,
+      "created_at": "2025-02-26T09:23:25Z",
+      "repoId": 765083837,
+      "pullRequestNo": 1785
     }
   ]
 }
\ No newline at end of file