Consider including a usecase using nipype/spm completion

[denisri]

We should add usecases, or check that we already have usecases for the following situations:

• completion for a pipeline which includes nipype/SPM processing, which performs its own completion model (prefixing or suffixing input filenames to set output paths)
• completion for several "datasets": several input files (input dataset completion, output dataset completion)
• a pipeline which uses several different, even incompatible requirements, such as a brick using SPM12 and a brick using SPM8
• populse-mia-like attributes deduction: for instance a process which takes as input 2 different subject T1 images, and which outputs data with attributes taken, or built, from both inputs, combined in a custom way. A simple one is an output where subject 2 is registered to subject 1, and another "symmetric" where subject 1 is registered to subject 2. The "subject" attribute of both outputs differ and is taken from different inputs.

(the list here may be modified)