Robert Sachunsky

> Can you run it with `--raise-on-error`? That should give you a proper stack trace and me something to work with. Oh, wow! That looks really pretty: ``` ╭─────────────────────────────── Traceback...

Note: it also happens with coordinates that do have more than 3 points **syntactically** but will have Shapely collapse duplicates: ```xml ```

Clearly, this is related to https://github.com/qurator-spk/eynollah/pull/86/commits/c606391c312eceab9aa3ebff071bdf12a30b45cc (`dir_in` mode), but it's a huge changeset, and I don't understand it (also, the diff is suboptimal). BTW, what was the reason for `dir_in`...

@mikegerber for a recent version usable in the OCR-D ecosystem, you need https://github.com/qurator-spk/eynollah/tree/706433c5049c63c6e16fee5f71d81a7e507abe06 which is part of https://github.com/qurator-spk/eynollah/pull/108

> Thanks but - I think - I don't currently have this problem - no conflicting qurator packages here. (And my recommendation to the team was to move away from...

> @bertsky Is this particular problem even connected to the namespace problem? It seems unrelated. @mikegerber no, but as as argued in #108 and above: you need the namespace problem...

I also don't see the point of urging this right now. We had a bit of a hustle when all this surfaced during the last ocrd_all release/repair sprint, but as...

Should have been closed via #129.

Ok, then (besides reformulation of the description) I highly recommend **renaming** that option, e.g. `apply_binarization`: after all, it's not the **input** that must/can be binary, but the internal **step** that...

Integrating sbb_binarization / experimenting with external tools: the OCR-D way would be to just use whatever derived images with `binarized` in `@comments` can be found, i.e. whatever binarization has been...