Closed melissalinkert closed 8 months ago
Unsurprisingly, this caused a number of build failures:
samples
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/34972/consolehamamatsu
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35019/consoleafi
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35023/consolesvs
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35025/consolezeiss-czi
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/34996/consoleventana
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35043/consoleome-tiff
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35003/consoleimaris
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35059/consoleleica-scn
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35067/consolecellsens
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35063/consolepyramid
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35089/consolejpeg2000
: https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-test-folder/35098/consoleExcluding for the moment, can re-include once a configuration PR is ready.
On the positive side, it looks like tests have been failing if and only if the format is multi-resolution which was the expectation.
As discussed today with @melissalinkert and @joshmoore, I agree the latest state of this PR puts us in a nice compromise between:
Next step here is to update the configuration files for all pyramidal file formats and re-include this PR in the nightly CI builds
In most cases, the
Image
names stored in aMetadataStore
will not depend onhasFlattenedResolutions()
. For some pyramid formats (e.g. SVS, CZI), theImage
names are noticeably different.This updates the config file generation to add an
Alternate_Name
representing the unflattenedImage
name, if the unflattenedImage
name is different from the flattenedImage
name for a particular series/resolution. The newtestUnflattenedImageNames
checks this value, and errors if the name provided by an unflattened reader does not match either configured name.I would expect this to cause test failures for SVS, CZI, and potentially others. I am happy for this to be excluded once at least test run has completed (so we know the scope of config changes required).