Hi @npeschke, thanks for this!

Would you mind adding a test? Just something small to check that a dask array is being generated with this logic? There's test data in this repo that could be used.

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
0 Code Smells

No Coverage information
No Duplication information

I added a basic type check for the loading from a txt and an equality check of stack comparing both ways of creating it.
The test takes about a second to run on my machine so I am unsure if it should be marked slow.

Pull Request Test Coverage Report for Build 1238474463

• 4 of 4 (100.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.03%) to 86.096%

Totals
Change from base Build 1144092739:	0.03%
Covered Lines:	1096
Relevant Lines:	1273

💛 - Coveralls

Looks great, thanks again!

Eventually all of this I/O and dask support should move to imio, so we can support more data formats with dask, across all software. This general issue has been moved to brainglobe/brainglobe-utils#36, if you feel like tackling some more issues @npeschke!

@npeschke is there a need for this to be on PyPI soon, or can it wait for the next release?

It would be nice to have that in PyPI soon, I plan to deploy cellfinder on a HPC cluster and this would make life easier for the admins that install cellfinder. Thanks for asking!

No worries, I'll aim to have it out by the end of the day.

What kind of cluster is it? If you're using SLURM, I can probably help. If not, would you be willing to (very briefly) document the process, to put in the docs and help out others?

cellfinder-core==0.2.8 is out now.

Thank you very much for the new build! Indeed I use as SLURM cluster and I am planning to wrap it into a snakemake workflow.