search

OCR-D / quiver-back-end

The back end of the OCR-D quality dashboard webapp.
MIT License
1 stars 2 forks source link

Add all workflows #43

Closed mweidling closed 1 year ago

mweidling commented 1 year ago

This MR introduces the other workflows proposed on the OCR-D website and adds them to the benchmark creation.

The workflow for slower processors had to be altered, though, because ocrd-skimage-denoise threw some error when using PIL. I substituted it with ocrd-cis-ocropy-denoise.

This MR depends on https://github.com/OCR-D/quiver-back-end/pull/42.

kba commented 1 year ago

This MR depends on #42.

Can we merge this first, so the difference from this PR as opposed to #42 becomes clear?

The workflow for slower processors had to be altered, though, because ocrd-skimage-denoise threw some error when using PIL. I substituted it with ocrd-cis-ocropy-denoise.

This really needs to be investigated. I failed to reproduce the issue with different Pillow/Python versions. Can you open an issue in https://github.com/bertsky/ocrd_wrap so we can find out what's going wrong together with @bertsky?

kba commented 1 year ago

This MR depends on #42.

Can we merge this first, so the difference from this PR as opposed to #42 becomes clear?

TIL that you can specify a commit range for reviews. Neat! https://github.com/OCR-D/quiver-back-end/pull/43/files/b9e4d4d00a53e119fc428eb4feb70b9ba07220b5..4630a1cf7dfe19e4ed67026b180ec88c0ee2c64b

mweidling commented 1 year ago

Can we merge this first, so the difference from this PR as opposed to https://github.com/OCR-D/quiver-back-end/pull/42 becomes clear?

Yes, I thought we tend to #42 first, that's why I made the note.

mweidling commented 1 year ago

@kba Are we ready to go? :)