Please don't use the qurator namespace

[mikegerber]

trafficstars

Apparently there is still a namespace problem when installing e.g. eynollah and sbb_binarization together and I think it should be resolved by not using "qurator" as part of the namespace (move "eynollah/" one level down and fix all references accordingly)

I've checked sbb_binarization and I don't think it should not conflict anymore but I'll check tomorrow to make sure. (@bertsky reported this in #106)

❗ However IF it still is a problem, then this is a serious bug that needs to be fixed ASAP

[mikegerber]

It's possible that the installation of eynollah does not conflict with sbb_binarization anymore, but with other qurator software (page2tsv etc.). I'll check.

[mikegerber]

Installing eynollah and sbb_binarization (both main/master) into one Python 3.9 venv does not cause immediately visible breakage. I can call both tools' CLIs' --help:

Call	result
eynollah --help	works
ocrd-eynollah-segment --help	works
sbb-binarize --help	works
ocrd-sbb-binarize --help	works

Installing sbb_tools (master again) into the same venv also does not cause breakage with e.g. altotool --help, it seems. So I am somewhat questioning if there is still a problem but as didn't test the released versions this could happen for those.

What I did NOT test: @bertsky mentioned installing in devel mode, and I didn't install it devel mode. This might be relevant in a way I don't see now. (I am also not recommending doing it, but that's not the point.)

What I am STILL saying: Using the qurator namespace should be avoided. Not sharing a namespace makes life easier.

I don't think I have anything else to contribute here.

[mikegerber]

Ah, I see, this is because the revert by @bertsky is in main already. And there's not a problem in sbb_binarization anymore.

So this needs more investigation.

TODO:

• [ ] revert the revert
• [ ] test again with sbb_tools
• [ ] verify that sbb_binarization really has no problem anymore.

[cneud]

Just a brief sidenote fyi - as of Eynollah 0.3.0, the model used by it for binarization is identical to the one of https://github.com/qurator-spk/sbb_binarization. As one of the next steps, binarization will be exposed as an invididual processing step in Eynollah, and eventually we will deprecate the separate https://github.com/qurator-spk/sbb_binarization project.

[bertsky]

The namespace problem has ultimately been fixed by #133, so this should be closed.

(And #151 will bring a functional replacement version of ocrd-sbb-binarize.)

[kba]

Yup, fixed in 84b8442 and sbb_binarization was merged into eynollah as well.