Add information to log messages

[HiranChaudhuri]

Hello there.

I came to using ocrmypdf through the paperless-ngx project. Ocrmypdf provides essential functionality, likely not only to paperless. A good job you've been doing here.

Some of the documents that I send to paperless error out, and I am trying to sort out the reasons. Often I read in the logfile lines like

[2025-02-01 10:34:58,866] [INFO] [ocrmypdf._pipeline] page is facing ⇧, confidence 12.11 - rotation appears correct
[2025-02-01 10:35:02,964] [INFO] [ocrmypdf._pipeline] page is facing ⇧, confidence 8.12 - no change
[2025-02-01 10:35:03,267] [INFO] [ocrmypdf._pipeline] page is facing ⇧, confidence 12.03 - rotation appears correct
[2025-02-01 10:35:19,361] [INFO] [ocrmypdf._pipeline] page is facing ⇧, confidence 12.98 - rotation appears correct
[2025-02-01 10:35:26,099] [INFO] [ocrmypdf._exec.tesseract] [tesseract] Too few characters. Skipping this page
[2025-02-01 10:35:26,118] [INFO] [ocrmypdf._exec.tesseract] [tesseract] Too few characters. Skipping this page
[2025-02-01 10:35:26,133] [ERROR] [ocrmypdf._exec.tesseract] [tesseract] Error during processing.
[2025-02-01 10:35:26,146] [INFO] [ocrmypdf._pipeline] page is facing ⇧, confidence 0.00 - no change
[2025-02-01 10:35:27,789] [INFO] [ocrmypdf._pipeline] page is facing ⇧, confidence 15.11 - rotation appears correct
[2025-02-01 10:35:30,031] [INFO] [ocrmypdf._pipeline] page is facing ⇧, confidence 11.39 - no change
[2025-02-01 10:35:43,499] [INFO] [ocrmypdf._exec.tesseract] [tesseract] Too few characters. Skipping this page
[2025-02-01 10:35:43,536] [INFO] [ocrmypdf._exec.tesseract] [tesseract] Too few characters. Skipping this page
[2025-02-01 10:35:43,556] [ERROR] [ocrmypdf._exec.tesseract] [tesseract] Error during processing.
[2025-02-01 10:35:43,578] [INFO] [ocrmypdf._pipeline] page is facing ⇧, confidence 0.00 - no change

Knowing that ocrmypdf processes several pages in parallel I have no clue which of the messages relates to which page of the processed document.

Would it be possible to add such information?

[jbarlow83]

ocrmypdf does track this information. Thread local storage is used to track which page number a thread is assigned to, so all messages sent from that thread contain the assigned page number. Then a logging filter is used to retrieve this information and edit the log message. Most likely paperless-ngx is not setting up the filter when it configures ocrmypdf, since loggers don't automatically do this. See ocrmypdf.api.configure_logging for details; most likely it just needs the equivalent of console.addFilter(PageNumberFilter()) inserted at the appropriate spot in paperless-ngx.