Standalone Processor Server (#884 + #974)

[MehmedGIT]

This draft PR is supposed to get important aspects from #884 and adapt to #974 so it can extend the ocrd_network package to provide the standalone Processor Server feature.

I am creating this a bit early so it's easy to follow what is going on with this. Still broken and fixes are needed. A lot more refactoring is needed to follow afterward. EDIT: Refactored and on the surface seems to work now.

Feel free to provide feedback.

[bertsky]

Looks good already.

[MehmedGIT]

@bertsky

Looks good already.

Not yet. There is still duplication of code. Mainly in the deployment phase.

This is still far away from a review, so please restrain to do so yet.

The reason I said that was to keep #1030 small because most of the things you mentioned in your review would have probably changed in the coming commits. For me going through your review for the current time being and answering everything separately in detail is more time-consuming than actually fixing things. Especially, when things are still not optimized and duplication of code is still out there (due to a clash of implementation ideas of 3 different people which I am trying to combine in #1030). Also, the logging in ocrd_network will change completely after the logging change in the core.

[MehmedGIT]

@kba, the blocking API for the Processor Servers and non-blocking requests to the Processor Servers from Processing Server have been implemented. This PR should now be good to be merged.

[kba]

AFAICT the only open discussion is the one around timeouts - shall I move this to a separate issue?

Otherwise, I'll merge after our call later, many thanks!

[bertsky]

Now or never!