mac本地部署4.x版本，解析ppt只能选择flash解析等级，大文件还会报错

[lilychen9908-hash]

mac本地部署mineru后，发现3.x版本能解析966M的ppt成功，但是同样的ppt，如果用最新的4.x版本，只能选择flash解析等级，问题还报错：docvortex.analyzers.native.office.errors.LegacyOfficeResourceLimitError: image equation records exceed max_picture_records=100000，怎么感觉ppt的解析退化了呢

[lilychen9908-hash]

报错日志如下：（其中/xxxxx是掩盖原有路径，这不重要)
| WARNING | mineru.integrations.docvortex:read_source_properties:41 - Source metadata unavailable: Metadata input exceeds 512 MiB
ERROR: Parse-server job file failed: job_id=job_dcc8541e525d5a94c500c07c file='1.pptx' tier=flash page_range=''
Traceback (most recent call last):
File "/xxxxx/venv/lib/python3.12/site-packages/mineru/parser/api_server.py", line 1482, in _run_job
result = await parse_async(
^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/mineru/parser/init.py", line 57, in parse_async
return await parser.parse_async(path, page_range=page_range, source_context=source_context)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/mineru/parser/mineru_parser.py", line 104, in parse_async
middle_json, model_output = await self._arun_analysis(prepared)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/mineru/parser/mineru_parser.py", line 129, in _arun_analysis
return await aio_doc_analyze(
^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/mineru/backend/analyze.py", line 111, in aio_doc_analyze
return await run_sync(
^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/mineru/utils/async_utils.py", line 30, in run_sync
return await asyncio.shield(work)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/.local/share/uv/python/cpython-3.12.14-macos-aarch64-none/lib/python3.12/asyncio/threads.py", line 25, in to_thread
return await loop.run_in_executor(None, func_call)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/.local/share/uv/python/cpython-3.12.14-macos-aarch64-none/lib/python3.12/concurrent/futures/thread.py", line 59, in run
result = self.fn(*self.args, **self.kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/mineru/backend/analyze.py", line 79, in doc_analyze
result = analyze_office(file_bytes, cast(OfficeSuffix, file_suffix))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/mineru/backend/analysis/office.py", line 40, in analyze_office
model_list = office_model.predict(BytesIO(file_bytes))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/models.py", line 136, in predict
converter.convert(file_binary)
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/pptx/pptx_converter.py", line 72, in convert
self._retry_convert_package_bytes_after_normalization(file_bytes, exc)
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/pptx/pptx_converter.py", line 117, in _retry_convert_package_bytes_after_normalization
raise exc
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/pptx/pptx_converter.py", line 68, in convert
self._convert_package_stream(file_stream)
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/pptx/pptx_converter.py", line 105, in _convert_package_stream
self._walk_linear(self.pptx_obj)
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/pptx/pptx_converter.py", line 134, in _walk_linear
shape_blocks = self._collect_shape_blocks(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/pptx/shapes.py", line 149, in _collect_shape_blocks
image_latex = self._decode_pptx_shape_image_equation(shape)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/pptx/resources.py", line 260, in _decode_pptx_shape_image_equation
return self._image_equation_decoder.decode(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/equation/image.py", line 440, in decode
candidates, recognized = _gif_mtef_candidates(image_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/equation/image.py", line 331, in _gif_mtef_candidates
_image_payload, cursor = _gif_subblocks(image_data, cursor, counter)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/equation/image.py", line 281, in _gif_subblocks
_charge_picture_record(counter)
File "/xxxxx/venv/lib/python3.12/site-packages/docvortex/analyzers/native/office/equation/image.py", line 46, in _charge_picture_record
raise LegacyOfficeResourceLimitError(f"image equation records exceed max_picture_records={MAX_PICTURE_RECORDS}")
docvortex.analyzers.native.office.errors.LegacyOfficeResourceLimitError: image equation records exceed max_picture_records=100000

[myhloli]

请提供样本文档供开发人员调试复测

[lilychen9908-hash]

升级了，但还是同样的报错

[myhloli]

那就确实需要您提供样本文件供我们测试才行了

[lilychen9908-hash]

那我怎么提供这个文件呢，当前聊天窗现在文件太大，无法上传上去。

另外还有一个问题，当前mac上进行解析一个pdf时候，发现有大量重复的解析内容，似乎陷入了重复循环：

[myhloli]

1. 可以使用网盘
2. 可以使用basic模式

[lilychen9908-hash]

1、文件在：
链接: https://pan.baidu.com/s/1SC16amC88BjqppsyzskLHw?pwd=fn3m
2、如果选用basic模式，解析后的效果就达不到要求

[myhloli]

pptx大文件的问题修复了，重复的问题需要使用vllm作为推理后端