misalignment of graphic layer about ocrmypdf HOT 4 CLOSED

Thanks for the report. The image is visually centred in the page, but the page in this PDF is not located at (0, 0) in PDF coordinates, which is allowed but unusual.

I should have a fix working for the next release.

The warning is for an intermediate single page PDF produced by ocrmypdf using PyPDF2, not your input PDF.

A temporary workaround which fixes both issues is to use --force-ocr. The drawback of using this is that the image will be transcoded, but with a high quality scan and --oversample set you will probably not notice any quality loss.

from ocrmypdf.

Thanks!

Could you please provide the command on the main page how to update/uninstall ocrmypdf HEAD revision from sources?

Because of my weak memory i always

• search for the right command,
• perform it (sudo pip3 uninstall ocrmypdf), and then
• follow your installation instructions.

(This is my new memo from now on.)

from ocrmypdf.

Remember I haven't released the change yet, but in one shot you can upgrade without uninstall/reinstall with the following (provided pip is reasonably new, all ducks are sufficiently aligned, enough gods have been sacrificed too):

[sudo] pip3 install --upgrade git+https://github.com/jbarlow83/OCRmyPDF.git

That is buried in the main page, but I will update that to make it clear.

from ocrmypdf.

Fixed in v4.0.
Hopefully.

from ocrmypdf.