Kalum Ost
Kalum Ost
**Rationale to not using the built-in `mutex` implementation already in SCT** To keep the user informed of what their processes are doing (and to make our lives easier if something...
Marking as ready for review to get feedback on the implementation; will add appropriate tests next week (after I finish some other work I have on the docket).
@mguaypaq Requested you simply because you're the one who proposed part of the solution in our last meeting. @joshuacwnewton feel free to jump in instead; I just want some eyes...
> Speaking of deadlocking, since some sct_deepseg tasks require more than one model, and some of the models are shared between tasks, we should probably make sure that we always...
> That said, I think it should still "just work" if we wrap the "[check if model exists, if not, download](https://github.com/spinalcordtoolbox/spinalcordtoolbox/blob/ed1905a548a2100369cbfc20d97d5da6e8e43b42/spinalcordtoolbox/scripts/sct_deepseg.py#L409-L422)" block of code in a mutex? Since the first...
You're running Framework with Python 3.13; I think SCT (and by extension, our installed Numpy version) uses Python 3.10. Can you try using a Framework with Python 3.10 instead? (NOTE:...
If the scripts don't have sensitive information in them, can you share both the (presumably `.sh`) script that you're passing to `sct_run_batch`, and the Python script that it then calls?...
Yup! This is very much a "future plan" issue, rather than something immediately pressing.