Error using LoadImages

[ewiger]

The test pipeline through's an error (Precluster step):

Prisca/240215_siRNA_SM_TfRecycling_Simple

PreCluster_with_pipeline: searching for .png or .tif images
PreCluster_with_pipeline: 16634 .png images detected, checking handles.Settings.VariableValues

Starting PreCluster analysis
- Running  module 01: LoadImages (t=0s)
Batch Error: LoadImages Error using LoadImages (line 513)
Image processing was canceled in the LoadImages module because the directory "" does not exist. Be sure that no spaces or unusual characters exist in your typed entry and that the pathname of the directory begins with /.
Error using LoadImages (line 513)
Image processing was canceled in the LoadImages module because the directory ""
does not exist. Be sure that no spaces or unusual characters exist in your
typed entry and that the pathname of the directory begins with /.

Error in LoadImages (line 513)
        error(['Image processing was canceled in the ', ModuleName, ' module
        because the directory "',SpecifiedPathname,'" does not exist. Be sure
        that no spaces or unusual characters exist in your t
Error in PreCluster_with_pipeline (line 137)
            handles = feval(ModuleName,handles);

Is there any obvoius settings missing when LoadImages was configured?

[tstoeger]

It appears that the input directory is empty (or can not be accessed). Besides the obvious explanations, which are likely already checked, one possibility could be the definition of the standard input and output folder (of CP, which also applies to LoadImages).

If left empty, "CPcluster" should set those such that they correspond to the "TIFF" and "BATCH" folder of the project from which the pipeline was started. (Note that in addition to standard CP input folder the standard input for images can be overwritten by the last entry within the LoadImages module - though we usually do not change that parameter)

Possibly the Batch files can elucidate whether the the correct input folders have been set.

Another possibility to test would be to define the standard input and output folder of CP by manually providing the absolute path to the standard input and output folders of CP (e.g.: something corresponding to /BIOL/sonas/biol_uzh_pelkmans_s7/Data/Users/RNAFish/IndividualExperiments/150419-EGFFollowNoMitotracker/TIFF and the corresponding /BATCH )

In addition LoadImages from CPPelkmans has a wrong definition of the function call [ function handles = LoadMoreImages(handles) ] instead of [ function handles = LoadImages(handles) ] , which according to MATLAB can / will give problems in future / (potentially already current?) MATLAB versions

[tstoeger]

the LoadImages module does not have the default value of the last input ( custom set of default import directory for images ) set properly (as the code of the module does not specify the default value correctly). After adding the proper default input (namely '.' instead of empty [] ) the prior error is gone (there are two additional files in channel 1 -> their presence correctly triggers another error)

[liberali]

Ok I do not get it. I am sorry but is the problem the data (I just copied deleted other stuff and ran a new pipeline) or the module? Can someone start maybe adding another dataset to check the problem? I will be in the lab tomorrow so we can check this Prisca

[tstoeger]

The initial problem was the code of the LoadImages module, which did not set the proper default value for the custom image input directory. This resulted in the initially observed error.

Now there is a problem with the amount of images (which indeed gets triggered later in the code). Counting files with "C01." and "C02." and "C03." (the input of LoadImages) yields (also outside of CP) a different amount of files for "C01." and "C02." and "C03."

[ewiger]

Indeed the problem is fixed.

New issue with dataset looks like this https://github.com/pelkmanslab/CellProfilerPelkmans/issues/9