[3rdparty]: paperless-ngx AttributeError: 'PdfInfo' object has no attribute '_pages'

[remisharrock]

Simple sanity checks

• This is an issue with an app that uses OCRmyPDF for OCR
• I am using a recent version of the third party app
• I will include a file that reproduces the issuse

Third party app name and version

paperless-ngx 2.14.7

Describe the bug

[2025-02-20 23:51:47,676] [DEBUG] [paperless.parsing.tesseract] Calling OCRmyPDF with args: {'input_file': PosixPath('/tmp/paperless/paperless-ngx79qbunrp/DevisPellegrin_1_Plantation-1.pdf'), 'output_file': PosixPath('/tmp/paperless/paperless-2zyv5sx4/archive.pdf'), 'use_threads': True, 'jobs': 2, 'language': 'fra', 'output_type': 'pdfa', 'progress_bar': False, 'color_conversion_strategy': 'RGB', 'redo_ocr': True, 'clean': True, 'rotate_pages': True, 'rotate_pages_threshold': 12.0, 'sidecar': PosixPath('/tmp/paperless/paperless-2zyv5sx4/sidecar.txt')}

[2025-02-20 23:51:47,816] [DEBUG] [paperless.parsing.tesseract] Deleting directory /tmp/paperless/paperless-2zyv5sx4

[2025-02-20 23:51:47,819] [ERROR] [paperless.consumer] Error occurred while consuming document DevisPellegrin_1_Plantation-1.pdf: AttributeError: 'PdfInfo' object has no attribute '_pages'

DevisPellegrin_1_Plantation.pdf

Steps to reproduce

1. Import attached file into Paperless-ngx
2. Trigger OCR
3. Check log file

Files

I have added the pdf file above

OCRmyPDF version

ocrmypdf --version 16.8.0

Relevant log output

[shamoon]

FYI, I cannot reproduce this in our current paperless-ngx release with 16.8:

[kernie]

Hi, it is related to the OCR mode in paperless and after forcing ocrmypdf to do the OCR (and not skip), I could reproduce the error as @remisharrock described (paperless-ngx/paperless-ngx#9179 (comment)).

[remisharrock]

thank @kernie , so when @shamoon says that he cannot reproduce because the OCR was maybe not triggered ?

[kernie]

Yes, default is mode 'skip', if ocr'ed text is available in the PDF, the ocr step is not done. When I activated mode 'redo' the error appeared also in the demo instance.

[shamoon]

Yes apologies for not confirming, I can reproduce as above

[remisharrock]

Thank you for investigating that. I actually have around 5% of my documents that have this kind of error (approximately 200 documents) . I think they might come from the same "source" (a scanner or something) but I'm not sure....

[kernie]

My documents come frome several sources like banks and insurances.

[aphex008]

Same here. Documents are usually invoices from different web stores.

[kernie]

Don't know if it helps..

I installed the current head of ocrmypdf and run it with the PDF above

(.venv) id6945@nucleus ~/Development/OCRmyPDF (main) $ ocrmypdf --redo-ocr ~/Downloads/DevisPellegrin_1_Plantation.pdf out.pdf
Scanning contents     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━   0% 0/3 -:--:--
An exception occurred while executing the pipeline                                                                    _common.py:296
Traceback (most recent call last):                                                                                                  
  File "/home/id6945/Development/OCRmyPDF/src/ocrmypdf/_pipelines/_common.py", line 261, in cli_exception_handler                   
    return fn(options, plugin_manager)                                                                                              
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^                                                                                              
  File "/home/id6945/Development/OCRmyPDF/src/ocrmypdf/_pipelines/ocr.py", line 178, in _run_pipeline                               
    validate_pdfinfo_options(context)                                                                                               
  File "/home/id6945/Development/OCRmyPDF/src/ocrmypdf/_pipeline.py", line 257, in validate_pdfinfo_options                         
    context.plugin_manager.hook.validate(pdfinfo=pdfinfo, options=options)                                                          
  File "/home/id6945/Development/.venv/lib64/python3.11/site-packages/pluggy/_hooks.py", line 513, in __call__                      
    return self._hookexec(self.name, self._hookimpls.copy(), kwargs, firstresult)                                                   
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^                                                   
  File "/home/id6945/Development/.venv/lib64/python3.11/site-packages/pluggy/_manager.py", line 120, in _hookexec                   
    return self._inner_hookexec(hook_name, methods, kwargs, firstresult)                                                            
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^                                                            
  File "/home/id6945/Development/.venv/lib64/python3.11/site-packages/pluggy/_callers.py", line 139, in _multicall                  
    raise exception.with_traceback(exception.__traceback__)                                                                         
  File "/home/id6945/Development/.venv/lib64/python3.11/site-packages/pluggy/_callers.py", line 103, in _multicall                  
    res = hook_impl.function(*args)                                                                                                 
          ^^^^^^^^^^^^^^^^^^^^^^^^^                                                                                                 
  File "/home/id6945/Development/OCRmyPDF/src/ocrmypdf/builtin_plugins/tesseract_ocr.py", line 195, in validate                     
    tess_threads = clamp(options.jobs // len(pdfinfo), 1, 3)                                                                        
                                         ^^^^^^^^^^^^                                                                               
  File "/home/id6945/Development/OCRmyPDF/src/ocrmypdf/pdfinfo/info.py", line 1254, in __len__                                      
    return len(self._pages)                                                                                                         
               ^^^^^^^^^^^                                                                                                          
AttributeError: 'PdfInfo' object has no attribute '_pages'  

After that I run the particular pdfinfo command:

(.venv) id6945@nucleus ~/Development/OCRmyPDF (main) $ python3 /home/id6945/Development/OCRmyPDF/src/ocrmypdf/pdfinfo/info.py  ~/Downloads/DevisPellegrin_1_Plantation.pdf
<PdfInfo('...'), page count=3>
<PageInfo pageno=0 8.222222222222222222222222222"x11.65277777777777777777777778" rotation=0 dpi=112.069732×116.005394 has_text=True>
<ImageInfo '/I0' image 175×227 Colorspace.rgb 3 8 image 112.069732×116.005394>
<PageInfo pageno=1 8.222222222222222222222222222"x11.65277777777777777777777778" rotation=0 dpi=0.000000×0.000000 has_text=True>
<PageInfo pageno=2 8.222222222222222222222222222"x11.65277777777777777777777778" rotation=0 dpi=402.704049×396.689525 has_text=True>
<ImageInfo '/I1' image 786×466 Colorspace.rgb 3 8 image 402.704049×396.689525>

which tells me, that _pages was there (at a point in time).

With some guidance I could follow up on this.

[kernie]

Another observation: When I run ocrmypdf with --force-ocr instead of --redo-ocr, it works without problems.

[nareddyt]

Also running into this on 2.14.7. Confirmed changing OCR mode to skip fixed it for me.

[carsaig]

Same issue here with an invoice inbound from a US software vendor, nothing special so far. Other documents are consumed without errors. I tested all releases downwards to 16.5.0 --> which is the last working release for that specific document causing the error. Anything beyond that introduced the issue. As soon as I get another document causing issues, I would compare it against the first one.
The OCR setting in paperless is set to: "PAPERLESS_OCR_SKIP_ARCHIVE_FILE: with_text"