Mathias Goncalves
Mathias Goncalves
Adds support for `FileBasedImage` methods `to_filename` and `instance_to_filename` to support keyword args. Adds keyword arg `validate` to CIFTI's `to_filename` and `instance_to_filename` methods. - If `False` (default), the intent code will...
Hi, I'm packaging niftyreg within a Docker container, and it would be great to avoid having to build the [CBSI release](https://github.com/KCL-BMEIS/niftyreg/releases/tag/CBSI) from source. I see other [releases packaged on sourceforge](https://sourceforge.net/projects/niftyreg/files/).
This won't be relevant for a bit since it is just debuting in 3.11 but just leaving this here. https://realpython.com/python311-exception-groups/#asynchronous-task-groups-in-python-311
3.9 added a [`cancel_futures`](https://docs.python.org/3/library/concurrent.futures.html#concurrent.futures.Executor.shutdown) keyword, but currently running `Future`s are still allowed to complete.
### What happened? Running fmriprep with `--derivatives anat=...` on data processed with a template not already cached in the Docker image results in a few failures. - `/scratch/fmriprep_0_0_wf/sub_01/ds_std_volumes_wf/_in_tuple_MNIPediatricAsymcohort-3.cohort3.resnative/gen_ref` - `/scratch/fmriprep_0_0_wf/sub_01_wf/anat_fit_wf/anat_reports_wf/_in_tuple_MNIPediatricAsymcohort--...
### Summary Nipype still uses pydot for graph rendering: https://github.com/nipy/nipype/blob/bf48f7722f057a1eb2005390eb0a7ebcd35ee8bf/nipype/pipeline/engine/utils.py#L1381 This has lead to problems (https://github.com/nipreps/fmriprep/issues/2870) and is on its way out in networkx (https://github.com/networkx/networkx/issues/5723), and might just easily be...
Popping up in tests ``` /opt/conda/lib/python3.9/site-packages/mriqc/bin/dfcheck.py:68: FutureWarning: Index.ravel returning ndarray is deprecated; in a future version this will return a view on self. feat_names = list(x_df._get_numeric_data().columns.ravel()) /opt/conda/lib/python3.9/site-packages/mriqc/bin/dfcheck.py:68: FutureWarning: Index.ravel returning...
Seeing if this fixes things
Following #319, we now use a few workflows for both T1w/T2w inputs, but some workflow names/fields are hardcoded with the `t1w` prefix.
### What would you like to see added in this software? This has come up in https://github.com/nipreps/fmriprep/pull/3060 To avoid filtering out estimators multiple times, we should incorporate a selective purge...