mineru 2.6.2__py3-none-any.whl → 2.6.3__py3-none-any.whl

mineru/backend/pipeline/batch_analyze.py

@@ -281,28 +281,20 @@ class BatchAnalyze:
 
                 # 按分辨率分组并同时完成padding
                 # RESOLUTION_GROUP_STRIDE = 32
-                RESOLUTION_GROUP_STRIDE = 64  # 定义分辨率分组的步进值
+                RESOLUTION_GROUP_STRIDE = 64
 
                 resolution_groups = defaultdict(list)
                 for crop_info in lang_crop_list:
                     cropped_img = crop_info[0]
                     h, w = cropped_img.shape[:2]
-                    # 使用更大的分组容差，减少分组数量
-                    # 将尺寸标准化到32的倍数
-                    normalized_h = ((h + RESOLUTION_GROUP_STRIDE) // RESOLUTION_GROUP_STRIDE) * RESOLUTION_GROUP_STRIDE  # 向上取整到32的倍数
-                    normalized_w = ((w + RESOLUTION_GROUP_STRIDE) // RESOLUTION_GROUP_STRIDE) * RESOLUTION_GROUP_STRIDE
-                    group_key = (normalized_h, normalized_w)
+                    # 直接计算目标尺寸并用作分组键
+                    target_h = ((h + RESOLUTION_GROUP_STRIDE - 1) // RESOLUTION_GROUP_STRIDE) * RESOLUTION_GROUP_STRIDE
+                    target_w = ((w + RESOLUTION_GROUP_STRIDE - 1) // RESOLUTION_GROUP_STRIDE) * RESOLUTION_GROUP_STRIDE
+                    group_key = (target_h, target_w)
                     resolution_groups[group_key].append(crop_info)
 
                 # 对每个分辨率组进行批处理
-                for group_key, group_crops in tqdm(resolution_groups.items(), desc=f"OCR-det {lang}"):
-
-                    # 计算目标尺寸（组内最大尺寸，向上取整到32的倍数）
-                    max_h = max(crop_info[0].shape[0] for crop_info in group_crops)
-                    max_w = max(crop_info[0].shape[1] for crop_info in group_crops)
-                    target_h = ((max_h + RESOLUTION_GROUP_STRIDE - 1) // RESOLUTION_GROUP_STRIDE) * RESOLUTION_GROUP_STRIDE
-                    target_w = ((max_w + RESOLUTION_GROUP_STRIDE - 1) // RESOLUTION_GROUP_STRIDE) * RESOLUTION_GROUP_STRIDE
-
+                for (target_h, target_w), group_crops in tqdm(resolution_groups.items(), desc=f"OCR-det {lang}"):
                     # 对所有图像进行padding到统一尺寸
                     batch_images = []
                     for crop_info in group_crops:
@@ -310,49 +302,34 @@ class BatchAnalyze:
                         h, w = img.shape[:2]
                         # 创建目标尺寸的白色背景
                         padded_img = np.ones((target_h, target_w, 3), dtype=np.uint8) * 255
-                        # 将原图像粘贴到左上角
                         padded_img[:h, :w] = img
                         batch_images.append(padded_img)
 
                     # 批处理检测
-                    det_batch_size = min(len(batch_images), self.batch_ratio * OCR_DET_BASE_BATCH_SIZE)  # 增加批处理大小
-                    # logger.debug(f"OCR-det batch: {det_batch_size} images, target size: {target_h}x{target_w}")
+                    det_batch_size = min(len(batch_images), self.batch_ratio * OCR_DET_BASE_BATCH_SIZE)
                     batch_results = ocr_model.text_detector.batch_predict(batch_images, det_batch_size)
 
                     # 处理批处理结果
-                    for i, (crop_info, (dt_boxes, elapse)) in enumerate(zip(group_crops, batch_results)):
+                    for crop_info, (dt_boxes, _) in zip(group_crops, batch_results):
                         bgr_image, useful_list, ocr_res_list_dict, res, adjusted_mfdetrec_res, _lang = crop_info
 
                         if dt_boxes is not None and len(dt_boxes) > 0:
-                            # 直接应用原始OCR流程中的关键处理步骤
-
-                            # 1. 排序检测框
-                            if len(dt_boxes) > 0:
-                                dt_boxes_sorted = sorted_boxes(dt_boxes)
-                            else:
-                                dt_boxes_sorted = []
-
-                            # 2. 合并相邻检测框
-                            if dt_boxes_sorted:
-                                dt_boxes_merged = merge_det_boxes(dt_boxes_sorted)
-                            else:
-                                dt_boxes_merged = []
-
-                            # 3. 根据公式位置更新检测框（关键步骤！）
-                            if dt_boxes_merged and adjusted_mfdetrec_res:
-                                dt_boxes_final = update_det_boxes(dt_boxes_merged, adjusted_mfdetrec_res)
-                            else:
-                                dt_boxes_final = dt_boxes_merged
-
-                            # 构造OCR结果格式
-                            ocr_res = [box.tolist() if hasattr(box, 'tolist') else box for box in dt_boxes_final]
-
-                            if ocr_res:
+                            # 处理检测框
+                            dt_boxes_sorted = sorted_boxes(dt_boxes)
+                            dt_boxes_merged = merge_det_boxes(dt_boxes_sorted) if dt_boxes_sorted else []
+
+                            # 根据公式位置更新检测框
+                            dt_boxes_final = (update_det_boxes(dt_boxes_merged, adjusted_mfdetrec_res)
+                                              if dt_boxes_merged and adjusted_mfdetrec_res
+                                              else dt_boxes_merged)
+
+                            if dt_boxes_final:
+                                ocr_res = [box.tolist() if hasattr(box, 'tolist') else box for box in dt_boxes_final]
                                 ocr_result_list = get_ocr_result_list(
                                     ocr_res, useful_list, ocr_res_list_dict['ocr_enable'], bgr_image, _lang
                                 )
-
                                 ocr_res_list_dict['layout_res'].extend(ocr_result_list)
+
         else:
             # 原始单张处理模式
             for ocr_res_list_dict in tqdm(ocr_res_list_all_page, desc="OCR-det Predict"):

mineru/backend/pipeline/model_init.py

@@ -8,7 +8,7 @@ from ...model.layout.doclayoutyolo import DocLayoutYOLOModel
 from ...model.mfd.yolo_v8 import YOLOv8MFDModel
 from ...model.mfr.unimernet.Unimernet import UnimernetModel
 from ...model.mfr.pp_formulanet_plus_m.predict_formula import FormulaRecognizer
-from ...model.ocr.paddleocr2pytorch.pytorch_paddle import PytorchPaddleOCR
+from mineru.model.ocr.pytorch_paddle import PytorchPaddleOCR
 from ...model.ori_cls.paddle_ori_cls import PaddleOrientationClsModel
 from ...model.table.cls.paddle_table_cls import PaddleTableClsModel
 # from ...model.table.rec.RapidTable import RapidTableModel

mineru/backend/pipeline/model_json_to_middle_json.py

@@ -148,7 +148,7 @@ def page_model_info_to_page_info(page_model_info, image_dict, page, image_writer
     fix_discarded_blocks = fix_discarded_block(discarded_block_with_spans)
 
     """如果当前页面没有有效的bbox则跳过"""
-    if len(all_bboxes) == 0:
+    if len(all_bboxes) == 0 and len(fix_discarded_blocks) == 0:
         return None
 
     """对image/table/interline_equation截图"""

mineru/backend/pipeline/pipeline_middle_json_mkcontent.py

@@ -191,11 +191,20 @@ def merge_para_with_text(para_block):
 def make_blocks_to_content_list(para_block, img_buket_path, page_idx, page_size):
     para_type = para_block['type']
     para_content = {}
-    if para_type in [BlockType.TEXT, BlockType.LIST, BlockType.INDEX]:
+    if para_type in [
+        BlockType.TEXT,
+        BlockType.LIST,
+        BlockType.INDEX,
+    ]:
         para_content = {
             'type': ContentType.TEXT,
             'text': merge_para_with_text(para_block),
         }
+    elif para_type == BlockType.DISCARDED:
+        para_content = {
+            'type': para_type,
+            'text': merge_para_with_text(para_block),
+        }
     elif para_type == BlockType.TITLE:
         para_content = {
             'type': ContentType.TEXT,
@@ -268,15 +277,19 @@ def union_make(pdf_info_dict: list,
     output_content = []
     for page_info in pdf_info_dict:
         paras_of_layout = page_info.get('para_blocks')
+        paras_of_discarded = page_info.get('discarded_blocks')
         page_idx = page_info.get('page_idx')
         page_size = page_info.get('page_size')
-        if not paras_of_layout:
-            continue
         if make_mode in [MakeMode.MM_MD, MakeMode.NLP_MD]:
+            if not paras_of_layout:
+                continue
             page_markdown = make_blocks_to_markdown(paras_of_layout, make_mode, img_buket_path)
             output_content.extend(page_markdown)
         elif make_mode == MakeMode.CONTENT_LIST:
-            for para_block in paras_of_layout:
+            para_blocks = (paras_of_layout or []) + (paras_of_discarded or [])
+            if not para_blocks:
+                continue
+            for para_block in para_blocks:
                 para_content = make_blocks_to_content_list(para_block, img_buket_path, page_idx, page_size)
                 if para_content:
                     output_content.append(para_content)

mineru/backend/vlm/vlm_analyze.py

@@ -8,6 +8,7 @@ from .utils import enable_custom_logits_processors, set_default_gpu_memory_utili
 from .model_output_to_middle_json import result_to_middle_json
 from ...data.data_reader_writer import DataWriter
 from mineru.utils.pdf_image_tools import load_images_from_pdf
+from ...utils.check_mac_env import is_mac_os_version_supported
 from ...utils.config_reader import get_device
 
 from ...utils.enum_class import ImageType
@@ -47,7 +48,7 @@ class ModelSingleton:
             for param in ["batch_size", "max_concurrency", "http_timeout"]:
                 if param in kwargs:
                     del kwargs[param]
-            if backend in ['transformers', 'vllm-engine', "vllm-async-engine"] and not model_path:
+            if backend in ['transformers', 'vllm-engine', "vllm-async-engine", "mlx-engine"] and not model_path:
                 model_path = auto_download_and_get_model_root_path("/","vlm")
                 if backend == "transformers":
                     try:
@@ -75,6 +76,15 @@ class ModelSingleton:
                     )
                     if batch_size == 0:
                         batch_size = set_default_batch_size()
+                elif backend == "mlx-engine":
+                    mlx_supported = is_mac_os_version_supported()
+                    if not mlx_supported:
+                        raise EnvironmentError("mlx-engine backend is only supported on macOS 13.5+ with Apple Silicon.")
+                    try:
+                        from mlx_vlm import load as mlx_load
+                    except ImportError:
+                        raise ImportError("Please install mlx-vlm to use the mlx-engine backend.")
+                    model, processor = mlx_load(model_path)
                 else:
                     if os.getenv('OMP_NUM_THREADS') is None:
                         os.environ["OMP_NUM_THREADS"] = "1"

mineru/backend/vlm/vlm_middle_json_mkcontent.py

@@ -248,13 +248,16 @@ def union_make(pdf_info_dict: list,
         paras_of_discarded = page_info.get('discarded_blocks')
         page_idx = page_info.get('page_idx')
         page_size = page_info.get('page_size')
-        if not paras_of_layout:
-            continue
         if make_mode in [MakeMode.MM_MD, MakeMode.NLP_MD]:
+            if not paras_of_layout:
+                continue
             page_markdown = mk_blocks_to_markdown(paras_of_layout, make_mode, formula_enable, table_enable, img_buket_path)
             output_content.extend(page_markdown)
         elif make_mode == MakeMode.CONTENT_LIST:
-            for para_block in paras_of_layout+paras_of_discarded:
+            para_blocks = (paras_of_layout or []) + (paras_of_discarded or [])
+            if not para_blocks:
+                continue
+            for para_block in para_blocks:
                 para_content = make_blocks_to_content_list(para_block, img_buket_path, page_idx, page_size)
                 output_content.append(para_content)
 

mineru/cli/client.py

@@ -4,6 +4,7 @@ import click
 from pathlib import Path
 from loguru import logger
 
+from mineru.utils.check_mac_env import is_mac_os_version_supported
 from mineru.utils.cli_parser import arg_parse
 from mineru.utils.config_reader import get_device
 from mineru.utils.guess_suffix_or_lang import guess_suffix_by_path
@@ -11,6 +12,11 @@ from mineru.utils.model_utils import get_vram
 from ..version import __version__
 from .common import do_parse, read_fn, pdf_suffixes, image_suffixes
 
+
+backends = ['pipeline', 'vlm-transformers', 'vlm-vllm-engine', 'vlm-http-client']
+if is_mac_os_version_supported():
+    backends.append("vlm-mlx-engine")
+
 @click.command(context_settings=dict(ignore_unknown_options=True, allow_extra_args=True))
 @click.pass_context
 @click.version_option(__version__,
@@ -38,25 +44,28 @@ from .common import do_parse, read_fn, pdf_suffixes, image_suffixes
     '--method',
     'method',
     type=click.Choice(['auto', 'txt', 'ocr']),
-    help="""the method for parsing pdf:
-    auto: Automatically determine the method based on the file type.
-    txt: Use text extraction method.
-    ocr: Use OCR method for image-based PDFs.
+    help="""\b
+    the method for parsing pdf:
+      auto: Automatically determine the method based on the file type.
+      txt: Use text extraction method.
+      ocr: Use OCR method for image-based PDFs.
     Without method specified, 'auto' will be used by default.
-    Adapted only for the case where the backend is set to "pipeline".""",
+    Adapted only for the case where the backend is set to 'pipeline'.""",
     default='auto',
 )
 @click.option(
     '-b',
     '--backend',
     'backend',
-    type=click.Choice(['pipeline', 'vlm-transformers', 'vlm-vllm-engine', 'vlm-http-client']),
-    help="""the backend for parsing pdf:
-    pipeline: More general.
-    vlm-transformers: More general.
-    vlm-vllm-engine: Faster(engine).
-    vlm-http-client: Faster(client).
-    without method specified, pipeline will be used by default.""",
+    type=click.Choice(backends),
+    help="""\b
+    the backend for parsing pdf:
+      pipeline: More general.
+      vlm-transformers: More general, but slower.
+      vlm-mlx-engine: Faster than transformers.
+      vlm-vllm-engine: Faster(engine).
+      vlm-http-client: Faster(client).
+    Without method specified, pipeline will be used by default.""",
     default='pipeline',
 )
 @click.option(
@@ -66,7 +75,7 @@ from .common import do_parse, read_fn, pdf_suffixes, image_suffixes
     type=click.Choice(['ch', 'ch_server', 'ch_lite', 'en', 'korean', 'japan', 'chinese_cht', 'ta', 'te', 'ka', 'th', 'el',
                        'latin', 'arabic', 'east_slavic', 'cyrillic', 'devanagari']),
     help="""
-    Input the languages in the pdf (if known) to improve OCR accuracy.  Optional.
+    Input the languages in the pdf (if known) to improve OCR accuracy.
     Without languages specified, 'ch' will be used by default.
     Adapted only for the case where the backend is set to "pipeline".
     """,
@@ -119,7 +128,8 @@ from .common import do_parse, read_fn, pdf_suffixes, image_suffixes
     '--device',
     'device_mode',
     type=str,
-    help='Device mode for model inference, e.g., "cpu", "cuda", "cuda:0", "npu", "npu:0", "mps". Adapted only for the case where the backend is set to "pipeline". ',
+    help="""Device mode for model inference, e.g., "cpu", "cuda", "cuda:0", "npu", "npu:0", "mps".
+         Adapted only for the case where the backend is set to "pipeline" and "vlm-transformers". """,
     default=None,
 )
 @click.option(

mineru/cli/gradio_app.py

@@ -13,6 +13,7 @@ from gradio_pdf import PDF
 from loguru import logger
 
 from mineru.cli.common import prepare_env, read_fn, aio_do_parse, pdf_suffixes, image_suffixes
+from mineru.utils.check_mac_env import is_mac_os_version_supported
 from mineru.utils.cli_parser import arg_parse
 from mineru.utils.hash_utils import str_sha256
 
@@ -273,7 +274,7 @@ def to_pdf(file_path):
 
 # 更新界面函数
 def update_interface(backend_choice):
-    if backend_choice in ["vlm-transformers", "vlm-vllm-async-engine"]:
+    if backend_choice in ["vlm-transformers", "vlm-vllm-async-engine", "vlm-mlx-engine"]:
         return gr.update(visible=False), gr.update(visible=False)
     elif backend_choice in ["vlm-http-client"]:
         return gr.update(visible=True), gr.update(visible=False)
@@ -381,6 +382,8 @@ def main(ctx,
                         preferred_option = "vlm-vllm-async-engine"
                     else:
                         drop_list = ["pipeline", "vlm-transformers", "vlm-http-client"]
+                        if is_mac_os_version_supported():
+                            drop_list.append("vlm-mlx-engine")
                         preferred_option = "pipeline"
                     backend = gr.Dropdown(drop_list, label="Backend", value=preferred_option)
                 with gr.Row(visible=False) as client_options:

mineru/cli/models_download.py

@@ -21,7 +21,7 @@ def download_and_modify_json(url, local_filename, modifications):
     if os.path.exists(local_filename):
         data = json.load(open(local_filename))
         config_version = data.get('config_version', '0.0.0')
-        if config_version < '1.3.0':
+        if config_version < '1.3.1':
             data = download_json(url)
     else:
         data = download_json(url)

mineru/model/ocr/{paddleocr2pytorch/pytorch_paddle.py → pytorch_paddle.py}

@@ -134,7 +134,7 @@ def get_model_params(lang, config):
         raise Exception (f'Language {lang} not supported')
 
 
-root_dir = os.path.join(Path(__file__).resolve().parent.parent.parent, 'utils')
+root_dir = os.path.join(Path(__file__).resolve().parent.parent, 'utils')
 
 
 class PytorchPaddleOCR(TextSystem):

mineru/model/table/rec/RapidTable.py

@@ -11,7 +11,7 @@ from rapid_table import ModelType, RapidTable, RapidTableInput
 from rapid_table.utils import RapidTableOutput
 from tqdm import tqdm
 
-from mineru.model.ocr.paddleocr2pytorch.pytorch_paddle import PytorchPaddleOCR
+from mineru.model.ocr.pytorch_paddle import PytorchPaddleOCR
 from mineru.utils.enum_class import ModelPath
 from mineru.utils.models_download_utils import auto_download_and_get_model_root_path
 

mineru/utils/block_sort.py

@@ -179,13 +179,14 @@ def insert_lines_into_block(block_bbox, line_height, page_w, page_h):
 def model_init(model_name: str):
     from transformers import LayoutLMv3ForTokenClassification
     device_name = get_device()
+    device = torch.device(device_name)
     bf_16_support = False
     if device_name.startswith("cuda"):
-        bf_16_support = torch.cuda.is_bf16_supported()
+        if torch.cuda.get_device_properties(device).major >= 8:
+            bf_16_support = True
     elif device_name.startswith("mps"):
         bf_16_support = True
 
-    device = torch.device(device_name)
     if model_name == 'layoutreader':
         # 检测modelscope的缓存目录是否存在
         layoutreader_model_dir = os.path.join(auto_download_and_get_model_root_path(ModelPath.layout_reader), ModelPath.layout_reader)

mineru/utils/check_mac_env.py

@@ -0,0 +1,30 @@
+# Copyright (c) Opendatalab. All rights reserved.
+import platform
+
+from packaging import version
+
+
+# Detect if the current environment is a Mac computer
+def is_mac_environment() -> bool:
+    return platform.system() == "Darwin"
+
+
+# Detect if CPU is Apple Silicon architecture
+def is_apple_silicon_cpu() -> bool:
+    return platform.machine() in ["arm64", "aarch64"]
+
+
+# If Mac computer with Apple Silicon architecture, check if macOS version is 13.5 or above
+def is_mac_os_version_supported(min_version: str = "13.5") -> bool:
+    if not is_mac_environment() or not is_apple_silicon_cpu():
+        return False
+    mac_version = platform.mac_ver()[0]
+    if not mac_version:
+        return False
+    # print("Mac OS Version:", mac_version)
+    return version.parse(mac_version) >= version.parse(min_version)
+
+if __name__ == "__main__":
+    print("Is Mac Environment:", is_mac_environment())
+    print("Is Apple Silicon CPU:", is_apple_silicon_cpu())
+    print("Is Mac OS Version Supported (>=13.5):", is_mac_os_version_supported())

mineru/utils/llm_aided.py

@@ -84,16 +84,21 @@ Corrected title list:
     max_retries = 3
     dict_completion = None
 
+    # Build API call parameters
+    api_params = {
+        "model": title_aided_config["model"],
+        "messages": [{'role': 'user', 'content': title_optimize_prompt}],
+        "temperature": 0.7,
+        "stream": True,
+    }
+
+    # Only add extra_body when explicitly specified in config
+    if "enable_thinking" in title_aided_config:
+        api_params["extra_body"] = {"enable_thinking": title_aided_config["enable_thinking"]}
+
     while retry_count < max_retries:
         try:
-            completion = client.chat.completions.create(
-                model=title_aided_config["model"],
-                messages=[
-                    {'role': 'user', 'content': title_optimize_prompt}],
-                extra_body={"enable_thinking": False},
-                temperature=0.7,
-                stream=True,
-            )
+            completion = client.chat.completions.create(**api_params)
             content_pieces = []
             for chunk in completion:
                 if chunk.choices and chunk.choices[0].delta.content is not None:

mineru/version.py

@@ -1 +1 @@
-__version__ = "2.6.2"
+__version__ = "2.6.3"

{mineru-2.6.2.dist-info → mineru-2.6.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mineru
-Version: 2.6.2
+Version: 2.6.3
 Summary: A practical tool for converting PDF to Markdown
 License: AGPL-3.0
 Project-URL: homepage, https://mineru.net/
@@ -37,7 +37,7 @@ Requires-Dist: scikit-image<1.0.0,>=0.25.0
 Requires-Dist: openai<3,>=1.70.0
 Requires-Dist: beautifulsoup4<5,>=4.13.5
 Requires-Dist: magika<0.7.0,>=0.6.2
-Requires-Dist: mineru-vl-utils<1,>=0.1.14
+Requires-Dist: mineru-vl-utils<1,>=0.1.15
 Provides-Extra: test
 Requires-Dist: mineru[core]; extra == "test"
 Requires-Dist: pytest; extra == "test"
@@ -50,6 +50,8 @@ Requires-Dist: transformers<5.0.0,>=4.51.1; extra == "vlm"
 Requires-Dist: accelerate>=1.5.1; extra == "vlm"
 Provides-Extra: vllm
 Requires-Dist: vllm<0.12,>=0.10.1.1; extra == "vllm"
+Provides-Extra: mlx
+Requires-Dist: mlx-vlm<0.4,>=0.3.3; extra == "mlx"
 Provides-Extra: pipeline
 Requires-Dist: matplotlib<4,>=3.10; extra == "pipeline"
 Requires-Dist: ultralytics<9,>=8.3.48; extra == "pipeline"
@@ -76,6 +78,7 @@ Requires-Dist: mineru[vlm]; extra == "core"
 Requires-Dist: mineru[pipeline]; extra == "core"
 Requires-Dist: mineru[api]; extra == "core"
 Requires-Dist: mineru[gradio]; extra == "core"
+Requires-Dist: mineru[mlx]; sys_platform == "darwin" and extra == "core"
 Provides-Extra: all
 Requires-Dist: mineru[core]; extra == "all"
 Requires-Dist: mineru[vllm]; extra == "all"
@@ -127,6 +130,10 @@ Dynamic: license-file
 </div>
 
 # Changelog
+- 2025/10/31 2.6.3 Release
+  - Added support for a new backend `vlm-mlx-engine`, enabling MLX-accelerated inference for the MinerU2.5 model on Apple Silicon devices. Compared to the `vlm-transformers` backend, `vlm-mlx-engine` delivers a 100%–200% speed improvement.
+  - Bug fixes: #3849, #3859
+
 - 2025/10/24 2.6.2 Release
   - `pipeline` backend optimizations
     - Added experimental support for Chinese formulas, which can be enabled by setting the environment variable `export MINERU_FORMULA_CH_SUPPORT=1`. This feature may cause a slight decrease in MFR speed and failures in recognizing some long formulas. It is recommended to enable it only when parsing Chinese formulas is needed. To disable this feature, set the environment variable to `0`.
@@ -666,7 +673,7 @@ https://github.com/user-attachments/assets/4bea02c9-6d54-4cd6-97ed-dff14340982c
 - Automatically recognize and convert formulas in the document to LaTeX format.
 - Automatically recognize and convert tables in the document to HTML format.
 - Automatically detect scanned PDFs and garbled PDFs and enable OCR functionality.
-- OCR supports detection and recognition of 84 languages.
+- OCR supports detection and recognition of 109 languages.
 - Supports multiple output formats, such as multimodal and NLP Markdown, JSON sorted by reading order, and rich intermediate formats.
 - Supports various visualization results, including layout visualization and span visualization, for efficient confirmation of output quality.
 - Supports running in a pure CPU environment, and also supports GPU(CUDA)/NPU(CANN)/MPS acceleration
@@ -703,41 +710,70 @@ A WebUI developed based on Gradio, with a simple interface and only core parsing
 > In non-mainline environments, due to the diversity of hardware and software configurations, as well as third-party dependency compatibility issues, we cannot guarantee 100% project availability. Therefore, for users who wish to use this project in non-recommended environments, we suggest carefully reading the documentation and FAQ first. Most issues already have corresponding solutions in the FAQ. We also encourage community feedback to help us gradually expand support.
 
 <table>
-    <tr>
-        <td>Parsing Backend</td>
-        <td>pipeline</td>
-        <td>vlm-transformers</td>
-        <td>vlm-vllm</td>
-    </tr>
-    <tr>
-        <td>Operating System</td>
-        <td>Linux / Windows / macOS</td>
-        <td>Linux / Windows</td>
-        <td>Linux / Windows (via WSL2)</td>
-    </tr>
-    <tr>
-        <td>CPU Inference Support</td>
-        <td>✅</td>
-        <td colspan="2">❌</td>
-    </tr>
-    <tr>
-        <td>GPU Requirements</td>
-        <td>Turing architecture and later, 6GB+ VRAM or Apple Silicon</td>
-        <td colspan="2">Turing architecture and later, 8GB+ VRAM</td>
-    </tr>
-    <tr>
-        <td>Memory Requirements</td>
-        <td colspan="3">Minimum 16GB+, recommended 32GB+</td>
-    </tr>
-    <tr>
-        <td>Disk Space Requirements</td>
-        <td colspan="3">20GB+, SSD recommended</td>
-    </tr>
-    <tr>
-        <td>Python Version</td>
-        <td colspan="3">3.10-3.13</td>
-    </tr>
+    <thead>
+        <tr>
+            <th rowspan="2">Parsing Backend</th>
+            <th rowspan="2">pipeline <br> (Accuracy<sup>1</sup> 82+)</th>
+            <th colspan="4">vlm (Accuracy<sup>1</sup> 90+)</th>
+        </tr>
+        <tr>
+            <th>transformers</th>
+            <th>mlx-engine</th>
+            <th>vllm-engine / <br>vllm-async-engine</th>
+            <th>http-client</th>
+        </tr>
+    </thead>
+    <tbody>
+        <tr>
+            <th>Backend Features</th>
+            <td>Fast, no hallucinations</td>
+            <td>Good compatibility, <br>but slower</td>
+            <td>Faster than transformers</td>
+            <td>Fast, compatible with the vLLM ecosystem</td>
+            <td>Suitable for OpenAI-compatible servers<sup>5</sup></td>
+        </tr>
+        <tr>
+            <th>Operating System</th>
+            <td colspan="2" style="text-align:center;">Linux<sup>2</sup> / Windows / macOS</td>
+            <td style="text-align:center;">macOS<sup>3</sup></td>
+            <td style="text-align:center;">Linux<sup>2</sup> / Windows<sup>4</sup> </td>
+            <td>Any</td>
+        </tr>
+        <tr>
+            <th>CPU inference support</th>
+            <td colspan="2" style="text-align:center;">✅</td>
+            <td colspan="2" style="text-align:center;">❌</td>
+            <td>Not required</td>
+        </tr>
+        <tr>
+            <th>GPU Requirements</th><td colspan="2" style="text-align:center;">Volta or later architectures, 6 GB VRAM or more, or Apple Silicon</td>
+            <td>Apple Silicon</td>
+            <td>Volta or later architectures, 8 GB VRAM or more</td>
+            <td>Not required</td>
+        </tr>
+        <tr>
+            <th>Memory Requirements</th>
+            <td colspan="4" style="text-align:center;">Minimum 16 GB, 32 GB recommended</td>
+            <td>8 GB</td>
+        </tr>
+        <tr>
+            <th>Disk Space Requirements</th>
+            <td colspan="4" style="text-align:center;">20 GB or more, SSD recommended</td>
+            <td>2 GB</td>
+        </tr>
+        <tr>
+            <th>Python Version</th>
+            <td colspan="5" style="text-align:center;">3.10-3.13</td>
+        </tr>
+    </tbody>
 </table>
+ 
+<sup>1</sup> Accuracy metric is the End-to-End Evaluation Overall score of OmniDocBench (v1.5), tested on the latest `MinerU` version.   
+<sup>2</sup> Linux supports only distributions released in 2019 or later.  
+<sup>3</sup> MLX requires macOS 13.5 or later, recommended for use with version 14.0 or higher.  
+<sup>4</sup> Windows vLLM support via WSL2(Windows Subsystem for Linux).  
+<sup>5</sup> Servers compatible with the OpenAI API, such as local or remote model services deployed via inference frameworks like `vLLM`, `SGLang`, or `LMDeploy`.
+
 
 ### Install MinerU
 

{mineru-2.6.2.dist-info → mineru-2.6.3.dist-info}/RECORD

@@ -1,28 +1,28 @@
 mineru/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
-mineru/version.py,sha256=53Sii4w6BIWn-1RhaTyqUO46gDe4nDCRQDAcpsWFH24,22
+mineru/version.py,sha256=uJ6TLK18jhCrL0aclBja7NzlAGLAyZjVpX-gq3d461k,22
 mineru/backend/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
 mineru/backend/utils.py,sha256=GLJU3IznDmhE1_qNmkU1UOtsuskIHBezgsEVO6Uar-Y,698
 mineru/backend/pipeline/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
-mineru/backend/pipeline/batch_analyze.py,sha256=dOnktvOMjfg84w1H34YlJg6N9_x6Yfvf14NIpOQcZqQ,22221
-mineru/backend/pipeline/model_init.py,sha256=OfB2MMjNmZcHl4fkqS1fT5R8I3LVoSKAHGtl8PcBfBs,9372
-mineru/backend/pipeline/model_json_to_middle_json.py,sha256=DtB7kE_7CtxwOMcb6QYeKzY6vMwUJNpavc5fn9z9oiI,10916
+mineru/backend/pipeline/batch_analyze.py,sha256=gnilKhFlMe8-55X2PJnb-ZSVeZIS-5DxIbMpHnwLne8,20889
+mineru/backend/pipeline/model_init.py,sha256=OAylOcQD9gu5TBcX7nMt7X5NpJMtQICI5IvEQ648lpI,9358
+mineru/backend/pipeline/model_json_to_middle_json.py,sha256=reXkUR_wKmJD64d7vRNXMxFviwkzDlGjRshpdwsVquI,10951
 mineru/backend/pipeline/model_list.py,sha256=7cXMBfZrP0K6qWueg1D_-WoUANeSINzkn_ic9E7YQLs,222
 mineru/backend/pipeline/para_split.py,sha256=Kq95MmvkPm7rKxlCSGiTvVKyF7CErHI2eGGAs5sLl0Q,17119
 mineru/backend/pipeline/pipeline_analyze.py,sha256=rbO5AetOdnxR5ctkoDzFCFoElkz7Jgb7gi2Ct596NK8,6655
 mineru/backend/pipeline/pipeline_magic_model.py,sha256=w8jGx8f6yZN0Wf2yPP3L9rYKc9rogxreZCrUJzJvPO8,14974
-mineru/backend/pipeline/pipeline_middle_json_mkcontent.py,sha256=xWWOFmYL6hB8PLrxQFyRJ72dAmTIDHtqiWV-WFUfR44,14081
+mineru/backend/pipeline/pipeline_middle_json_mkcontent.py,sha256=YlnEbbUnkniZXS13aLo5mjfFQvQM5SrIVvTAGBZsLmw,14478
 mineru/backend/vlm/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
 mineru/backend/vlm/model_output_to_middle_json.py,sha256=AqYX44gS9crUO_t7SuUatD71EVjow6pI6yA2Ik3gQ0s,5139
 mineru/backend/vlm/utils.py,sha256=woGqyRI4S7p69daLCU07XNXWTV27aLf7YBjjVH1x-5o,2794
-mineru/backend/vlm/vlm_analyze.py,sha256=nzwTGndwZFfTEvHppakyDKZxph7SYOuUZW3johY5F8c,8154
+mineru/backend/vlm/vlm_analyze.py,sha256=7c5_JN1F9YTDNNgA_Rmw6xX1PI7gcIT4A4ujtGQHH9Q,8792
 mineru/backend/vlm/vlm_magic_model.py,sha256=Pd0sOr7G1crAJIVeq6h_03gNSuxmV5U8dvGTGT_rrjs,23452
-mineru/backend/vlm/vlm_middle_json_mkcontent.py,sha256=Ie95XpwTgi7EmidcwE_scvXMRQjE2xASU_Rm_F8EP-I,13377
+mineru/backend/vlm/vlm_middle_json_mkcontent.py,sha256=5V-AU9KkxxMn0DDSQBrb15I4GVpEyiQy8uNI_tQhS6M,13498
 mineru/cli/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
-mineru/cli/client.py,sha256=uo7db9Wqj1Mc11MYuaM-bi54BfKKU3SFB9Urc8md5X4,6641
+mineru/cli/client.py,sha256=ArnoT2psOQRnTqLpsFwPaoi-l444iIVkbBn90Pm16n8,6915
 mineru/cli/common.py,sha256=jxFJMdc-02UMO3SXAtcZ6aIdPrakAE6DCccZ9kDlPKc,14276
 mineru/cli/fast_api.py,sha256=t5bda769VbM5iokAboiJfPIOnm-r5GTFReE-KQy8L3g,10941
-mineru/cli/gradio_app.py,sha256=8rMdW7grwBUn0MdXyG4eOTQUzKWq6nErtMWl-vGdWbU,14525
-mineru/cli/models_download.py,sha256=7KA-Boe-eIt3WW6eyaxM1HfubTXLsQ8sMmT1H1X7vAc,4815
+mineru/cli/gradio_app.py,sha256=6dA0ARpdOoewFeXmHrleF1amCgBV9ilY_nkWAmAmN8A,14731
+mineru/cli/models_download.py,sha256=LNfoIpUlJM7m7qb2SiCxtjMDw4jILBQtZwNP2JoY81U,4815
 mineru/cli/vlm_vllm_server.py,sha256=fQJyD-gIPQ41hR_6aIaDJczl66N310t0CiZEBAfX5mc,90
 mineru/data/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
 mineru/data/data_reader_writer/__init__.py,sha256=9qnGNrsuGBMwwfsQy6oChdkz--a_LPdYWE0VZZr0yr4,490
@@ -62,8 +62,7 @@ mineru/model/mfr/unimernet/unimernet_hf/unimer_swin/configuration_unimer_swin.py
 mineru/model/mfr/unimernet/unimernet_hf/unimer_swin/image_processing_unimer_swin.py,sha256=a9kCvwzJJSRrKQNtW2oOpTwrapzep8BjGFWLhLF1T0k,6036
 mineru/model/mfr/unimernet/unimernet_hf/unimer_swin/modeling_unimer_swin.py,sha256=Q_fdmFHUBtEoAfWp9aowdwTCE2MIFMOPbYjoSyXK2iU,48929
 mineru/model/ocr/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
-mineru/model/ocr/paddleocr2pytorch/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
-mineru/model/ocr/paddleocr2pytorch/pytorch_paddle.py,sha256=wZOw82q1NARNHBW2Lk5zumjdAqzPZqnhV6rvMULvLs8,9207
+mineru/model/ocr/pytorch_paddle.py,sha256=cHMTl5sKyn4BY2207-7GQ4eZl9BQUcs5ucxw_NFezII,9200
 mineru/model/ori_cls/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
 mineru/model/ori_cls/paddle_ori_cls.py,sha256=VIS22IerHST7g60AC9r2PEQIG6NQWeQaH1OrXIxNTsg,11943
 mineru/model/reading_order/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
@@ -72,7 +71,7 @@ mineru/model/reading_order/xycut.py,sha256=ezNSq_Y4UXiztB58hbXJsjTJlOBqWIjuW5A2u
 mineru/model/table/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
 mineru/model/table/cls/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
 mineru/model/table/cls/paddle_table_cls.py,sha256=5PtieKQnAzgMNRTZFgnqQsGWKTEQ3yyFWQnBRIjfQ4A,5781
-mineru/model/table/rec/RapidTable.py,sha256=FxO3dLNKfQrgcQU7gRI0kLAxllnoHWZptCtyyHNuMpM,5973
+mineru/model/table/rec/RapidTable.py,sha256=2dNdGJsVdsGfRm6r3deERUMst5RIxH0YuiGALkQbNTw,5955
 mineru/model/table/rec/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
 mineru/model/table/rec/slanet_plus/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 mineru/model/table/rec/slanet_plus/main.py,sha256=vfrcvQ9JBf32YZU9eNoetoqdpcrFNsA1WNqQBsG8i2o,7646
@@ -152,8 +151,9 @@ mineru/resources/header.html,sha256=PUselBXLBn8gfeP3zwEtj6zIxfhcCN4vN_B796nQFNQ,
 mineru/resources/fasttext-langdetect/lid.176.ftz,sha256=jzRyz-hzintgmejpmcPL-uDc0VaWqsfXc4qAOdtgPoM,938013
 mineru/utils/__init__.py,sha256=8CRrCQVuExa0BttRFh3Z40lFy2K5jN0sp67KWjOlj5c,50
 mineru/utils/block_pre_proc.py,sha256=uGBmxf2MR9bplTnQI8xHjCI-kj3plRhJr0hcWKidbOQ,9632
-mineru/utils/block_sort.py,sha256=mViceDw3O2ksBDFxt-wmX67bCZOwKyp68yZnEjS3Ijc,12934
+mineru/utils/block_sort.py,sha256=5e1mOLB3W7xu5Y1hmhvGSHPL_aQ41R_4VXcP4vjYAOU,12976
 mineru/utils/boxbase.py,sha256=moP660AmZq_udHEsfvFkTQdJ4gjrrBwN7t0Enx7CIL8,6903
+mineru/utils/check_mac_env.py,sha256=pbmbcnS60zZRqNRBxFJbbPrCosU1lC464b0v6JUlgaE,1031
 mineru/utils/cli_parser.py,sha256=4seFAu1kulsYnw6WM2q_cxgEOt2tErZVkI-LNEF_kGw,1445
 mineru/utils/config_reader.py,sha256=IRVWTpBnbnRpck6eXZUKw-fcLt7hon5S4uqWW-RBb1w,4075
 mineru/utils/cut_image.py,sha256=g3m4nfcJNWlxi-P0kpXTtlmspXkMcLCfGwmYuQ-Z2hE,751
@@ -163,7 +163,7 @@ mineru/utils/format_utils.py,sha256=2s89vHcSISjuolk8Hvg3K-5-rRbiT3Us7eFLzUKrNKs,
 mineru/utils/guess_suffix_or_lang.py,sha256=nznyQpUn1BSA8JNw9HuG3pVV-xtVAtrtcGuHZ-VXt9M,856
 mineru/utils/hash_utils.py,sha256=UPS_8NRBmVumdyOv16Lmv6Ly2xK8OVDJEe5gG6gKIFk,857
 mineru/utils/language.py,sha256=7RT3mxSa7jdpoC5ySd7ZddHA7TO7UsnmDOWiYZAxuyg,1433
-mineru/utils/llm_aided.py,sha256=eBGKCD7cJBjkyn38yqCdh0S-fgRG9fLuQCByLDQuyWs,4983
+mineru/utils/llm_aided.py,sha256=9WUytvxenSAuaWR4sTQhVPQ5h8pY0wVOH1O2sj_6dLs,5149
 mineru/utils/magic_model_utils.py,sha256=2xOvi4oqg3MSw1FUrJTnYDtWeFrrm6qbmlEorLZSaYs,5650
 mineru/utils/model_utils.py,sha256=6OsgFLsABX5JuShSzCMSNHWV-yi-1cjwHweafyxIgRo,18448
 mineru/utils/models_download_utils.py,sha256=UfjvwhxO6BkJHa5JSpEVNZ71GoLMPMmJpym3THET2T4,2957
@@ -176,9 +176,9 @@ mineru/utils/run_async.py,sha256=rPeP4BCZerR8VByRDhiYzfZiahLVqoZEBVAS54dAjNg,128
 mineru/utils/span_block_fix.py,sha256=0eVQjJCrT03woRt9hoh6Uu42Tp1dacfGTv2x3B9qq94,8797
 mineru/utils/span_pre_proc.py,sha256=h41q2uQajI0xQbc_30hqaju1dv3oVYxBAlKgURl8HIc,13692
 mineru/utils/table_merge.py,sha256=d98zNbM1ZQ8V1kUt6RugParNUNPv7DGL-XKIzR3iJVQ,15360
-mineru-2.6.2.dist-info/licenses/LICENSE.md,sha256=jVa0BUaKrRH4erV2P5AeJ24I2WRv9chIGxditreJ6e0,34524
-mineru-2.6.2.dist-info/METADATA,sha256=QGCp0YLuKymDMYmMZuOn8IYM-kpbKas5nKF7yl3la_0,68440
-mineru-2.6.2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-mineru-2.6.2.dist-info/entry_points.txt,sha256=luXmbhPiZK_tKlRgWuYOaW_V6EFpG-yJcAevVv9MEqE,252
-mineru-2.6.2.dist-info/top_level.txt,sha256=zuGQfZcbsHv4I4oKI9gaKPqEWBFm6xJroKuug2LnKP8,7
-mineru-2.6.2.dist-info/RECORD,,
+mineru-2.6.3.dist-info/licenses/LICENSE.md,sha256=jVa0BUaKrRH4erV2P5AeJ24I2WRv9chIGxditreJ6e0,34524
+mineru-2.6.3.dist-info/METADATA,sha256=mUBTxFG5tgdyQ1caZVRNrk4MhIX36PECY09dwCNYXZ4,70689
+mineru-2.6.3.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+mineru-2.6.3.dist-info/entry_points.txt,sha256=luXmbhPiZK_tKlRgWuYOaW_V6EFpG-yJcAevVv9MEqE,252
+mineru-2.6.3.dist-info/top_level.txt,sha256=zuGQfZcbsHv4I4oKI9gaKPqEWBFm6xJroKuug2LnKP8,7
+mineru-2.6.3.dist-info/RECORD,,

mineru/model/ocr/paddleocr2pytorch/__init__.py

@@ -1 +0,0 @@
-# Copyright (c) Opendatalab. All rights reserved.