Leonid Kostrykin

No, I was on the default Ubuntu 22.04 terminal app, I think it's gnome-terminal

Yes, using `delta_frac` to verify the result image is not reliable. I'd suggest to use the `img_diff` comparison method https://docs.galaxyproject.org/en/master/dev/schema.html#tool-tests-test-output, because there are composite images involved (i.e. figures with multiple...

As I see from the test results https://github.com/bgruening/galaxytools/actions/runs/10681249964?pr=1495, the issue is that the difference between the images generated by Cellpose on the one hand, and the expected images on the...

Ok I managed to generate the results using `--update_test_data`… kind of a workaround, but anyways… The issue is that the generated results indeed are very different than the expected results....

> @kostrykin this behaviour is quite strange. Are you using `--biocontainers` in your planemo test command? It takes forever to run with `--biocontainers` like several hours, so I omitted it...

> @kostrykin Ideally the tools needs to be tested with `--biocontainers` as this simulates the exact behaviour of Planemo run on GitHub. > > I updated the test data with...

I just tested the base branch on my computer. Using `--biocontainers` still is horribly slow, but faster than with the new version in this PR (like just 1-2 hours instead...

> @kostrykin I will first migrate the changes of `image_diff` to the base branch and see if all tests pass and then compare the results of CI and locally. For...

Nope, unfortunately, I haven't seen similar issues before. Basically there are two options, I guess: Either a dependency of the [Cellpose package](https://github.com/SaimMomin12/galaxytools/blob/cdd5be10848b9050631fa1110adc29c5591911bf/tools/cellpose/cellpose.xml#L14) is resolved to different versions and this induces...

> What should we do here? Do we assume cellpose knows what they are doing and accept that the outputs have changed so dramatically? @bgruening If I understood [this](https://github.com/bgruening/galaxytools/pull/1495#issuecomment-2360571156) comment...