Robert Sachunsky

Moreover, since @kba [says `local_filename` will now officially be `: Path`](https://github.com/OCR-D/core/pull/1167), we need yet another fix here.

Has been implemented in https://github.com/qurator-spk/eynollah/commit/6b52da227c042c52d79ebbfc711e1a0f3b093e89 and https://github.com/qurator-spk/eynollah/commit/38a2d60fa2766aac3dc8f0412bb60315fa38ffdf on master – many thanks!

Or should we keep the issue open until marginalia and non-light branches are extended for confidence aggregation? (Or even other segments like TextLine, ImageRegion, SeparatorRegion, Border??)

> Just for your information, 'marginalia' cannot have a confidence value since it is detected using heuristic methods. IIUC the base textline mask from the pixel classifier is also used...

Regarding the command line client, IMO it should be as consistent with the existing CLIs as possible. And I would prefer names of operations instead of HTTP mnemonics (POST/GET/PUT), for...

> Worth noting that the `list` option can get big really fast if the listing is not done just on user level when there are many users. Yes, but adding...

@MehmedGIT we currently have the `ocrd network client processing processor NAME --address ADDR --agent-type=worker` (via `publish_to_queue` asynchronously) and `--agent-type=server` (via `push_to_processor_server` synchronously). The latter also uses the Processing Server –...

I believe this can be closed.

To elaborate: - [ ] add an option `--jobs` to `ocrd process` which would split the workspace into per-page pipelines synchronised via METS server and managed by Python's builtin `multiprocessing`...

This has been superseded by the v3.0 API changes: > ocrd process --jobs 4 --timeout 2m --on-error=empty … became … OCRD_MAX_PARALLEL_PAGES=4 OCRD_PROCESSING_PAGE_TIMEOUT=120 OCRD_MISSING_OUTPUT=COPY ocrd process ...