When scripts are defining input and output as Scijava parameters, we should write a test of this script. This way it will be easy to detect backward incompatible changes in the future.
Because many of theses scripts are going to be linked to the Wiki, it's important for the authors to quickly be notified of a backward incompatible change and so update his scripts that would automatically update the corresponding page on the Wiki.
I propose the tests to be fairly simple : only test if the script ca run without Error/Exception. I think that is a good start. In case of Dataset as output, we could also check for the excepted number of dimensions I guess.
The issue that need to be solved is the input data. Which source should we use ? IJ Samples ? Dataset hosted in this repo ? The number of input data does not need to be huge, but diverse enough to cover a wide variety of images.
When scripts are defining input and output as Scijava parameters, we should write a test of this script. This way it will be easy to detect backward incompatible changes in the future.
Because many of theses scripts are going to be linked to the Wiki, it's important for the authors to quickly be notified of a backward incompatible change and so update his scripts that would automatically update the corresponding page on the Wiki.
I propose the tests to be fairly simple : only test if the script ca run without
Error
/Exception
. I think that is a good start. In case ofDataset
as output, we could also check for the excepted number of dimensions I guess.The issue that need to be solved is the input data. Which source should we use ? IJ Samples ? Dataset hosted in this repo ? The number of input data does not need to be huge, but diverse enough to cover a wide variety of images.