HTML report in 0.23.0 missing sections?

[michellewang]

Summary

Hi, I have previously run QSIPrep version 0.21.4 on a participant, then tried the latest version 0.23.0 on the same participant. It looks like the HTML report generated by 0.23.0 only contains the "About", "Methods" and "Errors" sections (and "About" is empty):

Whereas I expected the report to contain more information like in 0.21.4:

I can confirm that in 0.23.0 the figures are still present in sub-<LABEL>/figures, which also contains HTML files such as sub-<LABEL>_about.html and sub-<LABEL>_summary.html, which as far as I can tell are not found in the 0.21.4 output. So I just wanted to check if this was expected behaviour or not.

Thank you!

Additional details

• QSIPrep version: 0.23.0 from DockerHub (note: the HTML report says 0.23.1.dev0+g634483f.d20240830)
• Apptainer version: 1.1.8

[tsalo]

Can you share the full QSIPrep call you used?

[tsalo]

Also, any chance you could try out pennbbl/qsiprep:unstable? Just to make sure that it's not fixed in the current version.

[michellewang]

Ok, will try running it with the unstable version.

Here is the command that was used (paths/IDs redacted):

/opt/conda/envs/qsiprep/bin/qsiprep \
    <PATH_BIDS> \
    <PATH_OUTPUT> \
    participant \
    --skip_bids_validation \
    --participant-label <LABEL> \
    --bids-database-dir <PATH_BIDS_DATABASE> \
    --omp-nthreads 4 \
    --anat-modality T1w \
    --b0-threshold 100 \
    --dwi_denoise_window auto \
    --denoise-method dwidenoise \
    --b1-biascorrect-stage final \
    --distortion-group-merge none \
    --anatomical-template MNI152NLin2009cAsym \
    --output-resolution 1.0 \
    --b0-to-t1w-transform Rigid \
    --intramodal-template-transform BSplineSyN \
    --fs-license-file <PATH_FS_LICENSE> \
    --b0-motion-corr-to iterative \
    --hmc-transform Affine \
    --hmc_model eddy \
    --shoreline_iters 2 \
    --pepolar-method TOPUP \
    --fmap-no-demean \
    -w <PATH_WORK_DIR> \
    --stop-on-first-crash

[michellewang]

I can confirm that the unstable version produces a report with more information, so it seems this has been fixed, thanks!

One thing is that the latest HTML reports seem to have dropped a few sections (including: "T1 to MNI registration", "DWI Bias correction", "b=0 Reference Image", "DWI Sampling Scheme") -- is this expected?

Is there an expected date for the next release? I am planning to run QSIPrep for a project and for reproducibility purposes it would be good to use a stable version container.

Thanks again!

[mattcieslak]

Hi @michellewang we're about to do a 1.0.0 release, which will have these issues ironed out

[michellewang]

That's great to hear @mattcieslak! I see that 1.0.0rc1 is already out. Do you have an estimate for when the final 1.0.0 release will come out? E.g. will it be in the order of days or weeks? Thanks!

[michellewang]

Hi @mattcieslak, I just wanted to follow up on the ETA for the 1.0.0 release. 1.0.0rc1 worked for me on test participants, so I am considering just going with it unless 1.0.0 will be released soon. What do you think?