dwiqc.json no longer exists in qsiprep output directory

[dkp]

Summary

The dwiqc.json file was extremely useful, especially when used with the https://www.nipreps.org/dmriprep-viewer/#/
However, as of qsiprep version 0.23.1, this file is no longer produced.

Additional details

Because this file was unique in its ability to allow you to browse multiple subject reports easily, I think it was worth keeping (or reinstating).

Next steps

You may have something else in mind to replace or improve this functionality, in which case, you should ignore my request ; )

[mattcieslak]

Hi @dkp! Quick question for you- do you often use the dwiqc.json files?

When we did The Great Split we were thinking that we could make a built-in workflow that would create these dwiqc.json files. But we don't actually have a good intuition for how often people use these files versus how much space they take up in your derivatives. What do you think would be best?

[dkp]

Hi @mattcieslak,
Thank you for your reply. Sadly, I did use dwiqc.json file, and I taught it to my students (so my last lesson failed spectacularly without it).
So, I'd love to have it back, at least until there is a better alternative. Going through lots of reports for a group is greatly facilitated if this is enabled.

[mattcieslak]

Would you be ok with running a quick

qsirecon \
    /path/to/qsiprep-output \
    /path/for/dwiqc-files \
   participant \
   --recon-spec interactive_qc 

to get a directory full of dwiqc files?

and then a

qsirecon \
    /path/to/qsiprep-output \
    /path/for/dwiqc-files \
   group \
   --recon-spec interactive_qc 

for the group level dwiqc.json?

[dkp]

@mattcieslak, I updated to QSIRecon v1.0.0rc3.dev0+g619341b.d20241121

I ran

export RECON_SPEC=interactive_qc

apptainer run --containall --writable-tmpfs --bind ${MRIS} --bind "${LICENSE}:/license.txt" --bind ${TEMPLATEFLOW_HOME} ${APP} "${MRIS}/derivatives/qsiprep" "${MRIS}/derivatives/qsirecon" participant --recon-spec ${RECON_SPEC} -v -v

QSIrecon cannot find the pipeline, so it does not produce anything:

$ cat slurm-3202798.out 
241209-13:31:19,136 cli INFO:
         Telemetry system to collect crashes and errors is enabled - thanks for your feedback! Use option ``--notrack`` to opt out.
Traceback (most recent call last):
  File "/opt/conda/envs/qsiprep/bin/qsirecon", line 8, in <module>
    sys.exit(main())
  File "/opt/conda/envs/qsiprep/lib/python3.10/site-packages/qsirecon/cli/run.py", line 46, in main
    parse_args()
  File "/opt/conda/envs/qsiprep/lib/python3.10/site-packages/qsirecon/cli/parser.py", line 600, in parse_args
    config.from_dict({})
  File "/opt/conda/envs/qsiprep/lib/python3.10/site-packages/qsirecon/config.py", line 704, in from_dict
    workflow.load(settings, init=initialize("workflow"), ignore=ignore)
  File "/opt/conda/envs/qsiprep/lib/python3.10/site-packages/qsirecon/config.py", line 246, in load
    cls.init()
  File "/opt/conda/envs/qsiprep/lib/python3.10/site-packages/qsirecon/config.py", line 594, in init
    workflow_spec = _load_recon_spec(cls.recon_spec)
  File "/opt/conda/envs/qsiprep/lib/python3.10/site-packages/qsirecon/workflows/base.py", line 389, in _load_recon_spec
    raise Exception(f"{spec_name} is not a file that exists or in {prepackaged}")
Exception: interactive_qc is not a file that exists or in ['amico_noddi', 'mrtrix_multishell_msmt_ACT-fast', 'mrtrix_multishell_msmt_pyafq_tractometry', 'test_workflow', 'dsi_studio_autotrack', 'mrtrix_singleshell_ss3t_noACT', 'mrtrix_singleshell_ss3t_ACT-hsvs', 'hbcd_scalar_maps', 'csdsi_3dshore', 'mrtrix_multishell_msmt_noACT', 'reorient_fslstd', 'mrtrix_singleshell_ss3t_ACT-fast', 'dsi_studio_gqi', 'TORTOISE', 'multishell_scalarfest', 'dipy_dki', 'dipy_mapmri', 'mrtrix_multishell_msmt_ACT-hsvs', 'ss3t_autotrack', 'dipy_3dshore', 'test_scalar_maps', 'pyafq_tractometry']

[mattcieslak]

We haven't implemented it yet, but that code will definitely work once it's in!

[dkp]

Well, that explains my failure ; )

Yes, I'd like that interactive_qc option very much!

I look forward to trying the next iteration.
Can you update this thread when you have a relevant version?

I'm intrigued that there is now a "group" option and not just a participant option. Thanks for all your hard work...the split seems like a REALLY good move and you clearly have some great plans in store.

Thanks for the qsi tool family!