Languages are not displayed after upgrading Ubuntu to 22.10 or later

[brandones]

~ $  apt list --installed 'tesseract-ocr-*'                             15:12:52
Listing... Done
tesseract-ocr-eng/lunar,lunar,now 1:4.1.0-2 all [installed]
tesseract-ocr-osd/lunar,lunar,now 1:4.1.0-2 all [installed,automatic]
tesseract-ocr-spa/lunar,lunar,now 1:4.1.0-2 all [installed]

[danpla]

Hi,

Could you please show the output of tesseract --list-langs? Also, did you restart dpScreenOCR after installing the languages?

[brandones]

~ $ tesseract --list-langs                                              18:47:55
List of available languages in "/usr/share/tesseract-ocr/5/tessdata/" (3):
eng
osd
spa

And yes, I did. Checked with

~ $ ps aux | grep ocr                                                   18:48:10
brandon   854893  0.0  0.0  14436  2408 pts/1    S+   18:48   0:00 grep --color=auto ocr
~ $                                                                     18:48:22

[danpla]

I tried both v1.3.0 and the current development version on Ubuntu 23.04 (Lunar Lobster), and both work without problems. But I built them manually as they are not available in PPA, so the question now is how dpScreenOCR was installed on your machine?

[brandones]

Yeah I have it installed via apt.

~ $ grep ^ /etc/apt/sources.list /etc/apt/sources.list.d/* | grep dpscreen                                                                     10:50:45
/etc/apt/sources.list.d/daniel_p-ubuntu-dpscreenocr-jammy.list:# deb https://ppa.launchpadcontent.net/daniel.p/dpscreenocr/ubuntu/ kinetic main # disabled on upgrade to kinetic
/etc/apt/sources.list.d/daniel_p-ubuntu-dpscreenocr-jammy.list:# deb-src https://ppa.launchpadcontent.net/daniel.p/dpscreenocr/ubuntu/ jammy main
/etc/apt/sources.list.d/daniel_p-ubuntu-dpscreenocr-jammy.list.distUpgrade:# deb https://ppa.launchpadcontent.net/daniel.p/dpscreenocr/ubuntu/ kinetic main # disabled on upgrade to kinetic
/etc/apt/sources.list.d/daniel_p-ubuntu-dpscreenocr-jammy.list.distUpgrade:# deb-src https://ppa.launchpadcontent.net/daniel.p/dpscreenocr/ubuntu/ jammy main

~ $ apt list dpscreenocr                                                10:47:53
Listing... Done
dpscreenocr/now 1.3.0-1~jammy1 amd64 [installed,local]

I'm on 23.04 (Lunar).

Thanks for your support with this. Other ideas for how to debug?

[danpla]

So the problem appeared after upgrading Ubuntu from 22.04 (Jammy) to 23.04, and the program worked fine on 22.04, right?

I guess I know what happened. Can you please show the output of ldd `which dpscreenocr` | grep tesseract?

[brandones]

Yeah, that's probably the case.

~ $ ldd $(which dpscreenocr) | grep tesseract                                                       17:45:29
        libtesseract.so.4 => /lib/x86_64-linux-gnu/libtesseract.so.4 (0x00007f589be00000)

[danpla]

Ubuntu 23.04 is shipped with Tesseract 5, while 22.04 uses Tesseract 4. During upgrade 22.04 to 23.04, Tesseract 4 was kept as the dependency of dpScreenOCR, but language packages were upgraded as they are not dependencies from the package manager's view. Tesseract 4 still tries to find them in /usr/share/tesseract-ocr/4.00/tessdata/, while they are now in /usr/share/tesseract-ocr/5/tessdata.

Long story short, the PPA now has a build for Lunar. I'm not sure if it will show up as a package upgrade since the package version is the same as in 22.04, so you may need to explicitly reinstall it, e.g.:

sudo apt update
sudo apt remove dpscreenocr
sudo apt install dpscreenocr

[brandones]

Had to re-add the PPA, but that did it! Thanks so much for your help @danpla . It's a great tool and I appreciate your work on it.

[danpla]

You're welcome!