The code runs fairly silently for a while, with only nilearn outputting information. Probably we should log info messages at various milestones. These messages can include that we've started or finished some step, but also useful information such as the sizes or shapes of the intermediate outputs produced. For example, the number of images that are going to be analyzed by nilearn may be interesting to the user because some images may not have survived the process of "handle_missing": "invalidate", etc.
Possibly, some of the logged information is important enough to also output at the save_output() stage.
The code runs fairly silently for a while, with only
nilearnoutputting information. Probably we should log info messages at various milestones. These messages can include that we've started or finished some step, but also useful information such as the sizes or shapes of the intermediate outputs produced. For example, the number of images that are going to be analyzed bynilearnmay be interesting to the user because some images may not have survived the process of"handle_missing": "invalidate", etc.Possibly, some of the logged information is important enough to also output at the
save_output()stage.