Quentin Blampey

Here is some code to do a similar approach to yours but slightly more automated, since it will directly use the 5th and 95th percentiles to scale the values: ```python...

Hello @lopollar, do you have any updates regarding this notebook?

Hi @PietroD, could you please show me the logs of the rule that failed? According to your Snakemake logs, the rule logs should be inside `/beegfs/scratch/prj/Spatial/code/sopa/workflow/.snakemake/slurm_logs/rule_resolve_comseg/5435796.log`

Hello @PietroD, Are you sure this is the right log? If so, it seems `snakemake` executed the wrong rule. It should run `resolve_comseg`, but it submitted a job with rule...

Did you run Sopa on many samples in parallel? > Could it be related to how snakemake and slurm interface each other when jobs are pending for resources? Yes, it...

Do you mean several Sopa config files with different segmentation methods? If so, yes, it could be the reason for the issue, although snakemake is supposed to handle that... >...

Actually I had some time this afternoon so I did it quickly! So, if you want, you can try the Nextflow version of Sopa, see the documentation [here](https://nf-co.re/sopa/dev/). Note that,...

Yes good idea @stergioc, you're right it will likely be much cleaner 😊 I think we could even implement the `__getitem__` dunder method to extract a mini batch, so that...

Hi, Unfortunately, it seems snakemake doesn't support uv yet, see [here](https://github.com/snakemake/snakemake/issues/3251) It says you could directly run snakemake via `uv` (e.g., `uv run snakemake ...`), but I assume it wouldn't...

Thanks @LucaMarconato, your change looks good!