PAPERLESS_PRE_CONSUME_SCRIPT not being invoked

[geomule]

Hi,
while trying to automate the issues around ExifOrientationError: Invalid rotation (0) I am having when uploading jpg images that i take with my Android phone using Open Note Scanner (the dam app assigns incorrect ExifOrientation tag), I though I would try to find a work around by using the PAPERLESS_PRE_CONSUME_SCRIPT.

So I added the following to paperless.conf:

PAPERLESS_PRE_CONSUME_SCRIPT="/opt/paperless-ng/pre-cosume-script.sh"

The /opt/paperless-ng/pre-cosume-script.sh file looks like this:

#!/bin/bash

find /mnt/zeus1vd2/paperless/consume -iname "*jpg" | \
  while read I; do
   exiftool -Orientation=1 -n "$I"
  done

The above script works correctly when I run it in bash. In fact when I drop the resulting files on the dashboard, they are consumed. However, if I drop the non corrected files on the dashboard, I get the ExifOrientationError: Invalid rotation (0) error reported elsewhere on this forum.

The paperless log file suggests that the PAPERLESS_PRE_CONSUME_SCRIPT is not even being invoked. Note, I did restart the paperless-webserver and paperless-consumer services.

Any thoughts on why Paperless seems to be ignoring the PRE_CONSUME script?

2021-04-02 09:21:05,686] [DEBUG] [paperless.consumer] Parsing 03-20210331-04.jpg...
[2021-04-02 09:21:05,820] [DEBUG] [paperless.parsing.tesseract] Estimated DPI 296 based on image width 2448
[2021-04-02 09:21:05,821] [DEBUG] [paperless.parsing.tesseract] Detected DPI for image /mnt/zeus1vd2/paperless/consume/03-20210331-03.jpg: 72
[2021-04-02 09:21:05,821] [DEBUG] [paperless.parsing.tesseract] Calling OCRmyPDF with args: {'input_file': '/mnt/zeus1vd2/paperless/consume/03-20210331-03.jpg', 'output_file': '/tmp/paperless/paperless-udaxq7nf/archive.pdf', 'use_threads': True, 'jobs': 2, 'language': 'eng+ita', 'output_type': 'pdfa', 'progress_bar': False, 'skip_text': True, 'clean': True, 'deskew': True, 'rotate_pages': True, 'rotate_pages_threshold': 12.0, 'sidecar': '/tmp/paperless/paperless-udaxq7nf/sidecar.txt', 'image_dpi': 72}
[2021-04-02 09:21:05,830] [DEBUG] [paperless.parsing.tesseract] Estimated DPI 296 based on image width 2448
[2021-04-02 09:21:05,830] [DEBUG] [paperless.parsing.tesseract] Detected DPI for image /mnt/zeus1vd2/paperless/consume/03-20210331-04.jpg: 72
[2021-04-02 09:21:05,831] [DEBUG] [paperless.parsing.tesseract] Calling OCRmyPDF with args: {'input_file': '/mnt/zeus1vd2/paperless/consume/03-20210331-04.jpg', 'output_file': '/tmp/paperless/paperless-sm0keepl/archive.pdf', 'use_threads': True, 'jobs': 2, 'language': 'eng+ita', 'output_type': 'pdfa', 'progress_bar': False, 'skip_text': True, 'clean': True, 'deskew': True, 'rotate_pages': True, 'rotate_pages_threshold': 12.0, 'sidecar': '/tmp/paperless/paperless-sm0keepl/sidecar.txt', 'image_dpi': 72}
[2021-04-02 09:21:06,197] [DEBUG] [paperless.parsing.tesseract] Deleting directory /tmp/paperless/paperless-udaxq7nf
[2021-04-02 09:21:06,201] [ERROR] [paperless.consumer] Error while consuming document 03-20210331-03.jpg: ExifOrientationError: Invalid rotation (0)
[2021-04-02 09:21:06,219] [DEBUG] [paperless.parsing.tesseract] Deleting directory /tmp/paperless/paperless-sm0keepl
[2021-04-02 09:21:06,224] [ERROR] [paperless.consumer] Error while consuming document 03-20210331-04.jpg: ExifOrientationError: Invalid rotation (0)

[jonaswinkler]

paperless-scheduler is the service that does all the processing. You'll need to reload that in order for the config to be read again. consumer is simply responsible for watching the input folder. Also, the pre-consume script is one of the first things that happens, even before "parsing file.pdf...".

[geomule]

Thanks for responding. It makes no difference though, just tries stopping all 3 services and started them again. BTW, the log is making no reference to calling the script. Should it be doing so?

[jonaswinkler]

Actually, no. There's no logging that the consumer executes the scripts. I'll adjust that.

[geomule]

Great! Some debugging will be useful. There is just no way of knowing if Paperless is calling the script. All I can assume at the moment is that its not being called.