Help with LoadData module

Hi,
I am using the 2.1.0 (rev eecd0b4) on a computing cluster and unfortunately I can’t drag and drop the folder into the input panel.
So I am trying the alternative way which is to use “LoadData” module.

For my setup, image I have the following images in the folder: /home/me/images
filename1_c1.tiff
filename1_c2.tiff
filename2_c1.tiff
filename2_c2.tiff

where I want to call the “_c1” file DNA and “_c2” file FOCI

I have generated a file called 'imagelist.csv" with the following contents:
DNA,FOCI
filename1_c1.tiff,filename1_c2.tiff
filename2_c1.tiff,filename2_c2.tiff

When I do “Press to view CSV file content” button, I see that CP properly loads this file.

I am not able to figure out what the difference is between “Input data file location” and “Base image location”, so in both the cases I have specified the same folder “/home/me/images”

I uncheck the “group images by metadata” .

The next module in my pipeline is “IdentifyPrimaryObjects” in which
in the “Select the input image” I have entered “DNA”

There is a yet another “IdentifyPrimaryObjects” module in which in the “Select the input image” I have entered “FOCI”

However, I get the error (when I hover over “IdentifyPrimaryObject” as “No prior instances of imagegroup were defined”

So I go back to the “LoadData” module and check the 'group images by metadata", however, the “Select metadata fields for grouping” is empty and I am unable to type anything in it.

Can you please tell me what I am doing wrong?
Thanks
Lee

Hi Lee,

[quote]I am using the 2.1.0 (rev eecd0b4) on a computing cluster and unfortunately I can’t drag and drop the folder into the input panel.
So I am trying the alternative way which is to use “LoadData” module.[/quote]

Note that you can alternatively load a folder using the Images module by going to “Edit > Browse for Images”. Perhaps you don’t have any access to the GUI at all, which is understandable.

Also note that CreateBatchFiles is intended to solve this condition! Please read its Help and it may save you some hassle in generating the CSV if you run one cycle locally with CreateBatchFiles at the end of your pipeline.

Another idea is to run CP 2.1 locally and only drag-drop your files/folders into Images, and then File>Export>‘Image set listing’ to generate the CSV for you (with all the headers in place, correctly formatted!). The only issue there is that the paths to your files need to be formatted the same way as how your cluster nodes read their paths (again, CreateBatchFiles solves this issue).

[quote]I have generated a file called 'imagelist.csv" with the following contents:
DNA,FOCI
filename1_c1.tiff,filename1_c2.tiff
filename2_c1.tiff,filename2_c2.tiff[/quote]

Please remember that you need to format your headers for filename channels as “Image_FileName_DNA” and not just “DNA”. The csv file will load just fine, but it won’t be interpreted as images to load.

[quote]I am not able to figure out what the difference is between “Input data file location” and “Base image location”, so in both the cases I have specified the same folder “/home/me/images”
[/quote]

The “?” button next to “Input data file location” states “Select the folder containing the CSV file to be loaded.” That does indeed sound like you have set it correctly. The help for “Base image location” states “The parent (base) folder where images are located. If images are contained in subfolders,…” so if you don’t have any subfolders, then this setting is irrelevant.

I think the problem in your CSV headers missing the “Images_FileName_” prepend.
David

So it sounds like this is now solved?

Thanks,
David