Yes, I have restarted my Mac and the problem persists.
I bought and Installed DEVONthink this morning , so OCR has never worked. This is the first file I try to convert.
Yes, I have restarted my Mac and the problem persists.
I bought and Installed DEVONthink this morning , so OCR has never worked. This is the first file I try to convert.
Ok, so bearing in mind that Jim is already involved and actually knows what he is doing, whilst I am restricted to guessing, what I would do would be to:
Note that the download is approx 600 mb and may take a while.
PS it is normal for PDF Services not to show it has been installed - DT has no feedback from the OS with that regard.
In my experience there will be no harm done from following my steps; however, I am just a regular here - Jim (@BLUEFROG) is from tech support and might well suggest another path if you are inclined to wait for him to come back online (I’m not versed in US time zones, so I’m not sure when that will be; he does generally look in on Sundays, though).
Hey, and I assume you have purchased the Pro edition? You can check at About DEVONthink 3 in the menu; the window which opens will show “Registered Pro edition to” if you are using the Pro version.
When you asked if DEVONthink had full disk access, I checked, and it did not, so I gave it full disk access. Unfortunately, as I previously said, after giving it full disk access, the problems persisted.
But it means that when I installed the OCR add-on DEVONthink did not have proper access, and it failed?
This is what I found inside the abbyy folder in the path you specified
No, I don’t think the installation failed for that reason; but I think I remember having seen a similar problem in the forum a good while back, and I think the conclusion was that the download probably didn’t properly conclude, leaving a defective copy of the OCRHelper.
Have you restarted DEVONthink since doing that?
after deleting the folder and re installing the OCR add-on, the problem persists.
I’ve just searched the forum and here Jim suggests to the user of an m1 Mac that granting FDA to the OCRHelper app may help. I’ve just checked, and on my Intel mac, the OCRHelper app does not have - or need - FDA. And the user sadly didn’t provide information on whether it was that step which solved the problem or not. Worth a try though (but do remove FDA from the OCRHelper app if it doesn’t help).
Presumably you have tried more than one PDF and none works? Is anything logged to the log (Window/Log)?
If this step also does not help, I would suggest waiting the day to see whether Jim provides the solution here, and otherwise opening a support ticket.
I gave the OCRHelper app full-disk-access, and the problem persists
Yes, I have tried different files. I need to convert various screenshots into OCR PDF files (I do this all the time in DEVONthink to Go).
I am indeed using an M1 mac (MacBook air).
Have you rebooted the Mac?
PS: sorry for the tardy reply. I have the flu. I’ll spare you the details
Do you mean restart or shut down the mac?
Yes I have
Ok.
Where did the PDF originate?
From Devonthink.
I have several screenshots that contain text. They are PNG files. I imported the PNG files into a DEVONthink library.
I tried to convert the PNG files into OCR Pdfs directly, and I had this issue. Then I converted the PNG into a regular PDF, that worked.
Then I tried to covert that pdf into an OCR pdf, that did not work
To help track down where the OCR is failing, could you turn on OCR logging, to do this:
The log is reporting a miscellaneous error when trying to start the OCR, could you try the following:
Yup,
That fixed it. I think I know what the problem was.
The DTOCRHelper does not work with M1.
When I double-clicked on the app, I got a message from MacOS indicating that I needed to download and install Rosetta to run the application.
I did not have Rosseta in my Mac because all my applications are M1 compatible, or so I thought…
After installing Rosetta, I launched Devonbthink, and I was able to OCR my pdf file.
My next question is:
is there a way DEVONthink can use the M1 version of the OCR library?
ABBYY are in the process of developing an Apple Silicon version of their OCR library, until that update is released it has to be run under Rosetta. They haven’t announced any release releases dates so I am unable to give you any timeframe for it to be a fully native Apple Silicon application.
Thanks for following this through - that will add to my repertoire of questions to ask users with similar problems
If you haven’t already done so, please remember to remove Full Disk Access from DTOCRHelper, as it probably isn’t required.
I already did
Thank you
Full Disk Access is only usually required when using a ScanSnap scanner with certain configurations of Scansnap Home or ScanSnap Manager.
If you have a brand new MacBook Pro 2021 beast, and it is just installed and haven’t installed any other Intel-only app, DTOCRHelper will fail because lack of Rosetta 2. Once manually run for first time DTOCRHelper and asked for installation of Rossetta, it went flawlessly.