Error using pixel classification with 2D h5 files

[hfahad]

issue #1749 solved that open in ilastik but i found the bug in lazyflow

during 2d Data in pixelClassificationDataExportGui.py file show error that 'c' is not found that why i put a condition if 'c' is not found then add a channel during hdf5 reader and also 'c' in axistags and in the end execute function we already change the dimension and axis info of the data. but when we read again data from orignal hdf5 file i face some error thats why i comeback to original shape to execute the data

[k-dominik]

hmm, yes, could you please rebase this branch onto master and force push it again?

# on lazyflow master branch
# assuming upstream is configured to be ilastik/lazyflow
git fetch upstream master
git rebase upstream/master

# then switch to your feature branch, rebase it onto the current master and force push the result
git checkout pixel-clasification-2d-image-error
git rebase master
git push --force

just in order to include recent fixes, thanks!

[hfahad]

done it .

[k-dominik]

There seem to be some legitimate test failures in testOpFormattedDataExport.TestOpFormattedDataExport.testBasic and . have you run the tests yourself?

you can do that by

# assuming you are in ilastik-meta/lazyflow
nosetests --where=./tests

you should see then, that two of the tests fail.

I think this has something to do with the additional channel, that is added to data without a channel.

Please also note that due some recent changes you have to update the fastfilters package in your conda environment:

# assuming you have activated the environment
conda update -c ilastik-forge fastfilters 

[k-dominik]

fixes https://github.com/ilastik/ilastik/issues/1845, https://github.com/ilastik/ilastik/issues/1749

[k-dominik]

on hold for now, as this breaks some things in ilastik, mostly tests