Use case: Define and use a Morphologist-like pipeline

[sapetnioc]

This use case contains a full pipeline definition composed of fake processes combined using most important Morphologist pipeline construction features:

• Switch: the pipeline must allow to select one normalization option : (no normalization, normalize with SPM or normalize with Aims).
• Reuse of same process: one process (i.e. split brain) will produce two outputs. Each output will go through the same processing but with different metadata (e.g. side=right and side=left).
• File name completion: the use case code will produce its own faked source database in BIDS organization. Results will be written in an output directory following BrainVISA organization.
• Iterations: Using BIDS organization, all the T1 MRI will be retrieved and processed by the pipeline.
• External software and data. Choosing SPM normalization will require a fake external software as well as a fake template whose default value is located in the software installation directory.

[sapetnioc]

I have a question regarding BIDS organization of derivative. In BIDS, data produced by a pipeline is located in a single directory containing the pipeline name. In this use case, there will be a SPM normalization pipeline that may be reused by Morphologist. If used alone, the SPM normalization result would go in a spm_normalization directory. But if this normalization is used from Morphologist, should the result go in the spm_normalization directory or in the morphologist directory ?

I have the feeling that results should always go in a directory named after the main pipeline. In this case this is morhpologist. But it will make things more difficult to find and reuse this normalization in another pipeline.

[sapetnioc]

The development of TinyMorphologist pipeline is tightly linked to the development of Capsul's new API. Would it be a problem if I move the correponding code in the tests modules of the Capsul project ?

[sapetnioc]

There is no subpipeline in TinyMorphologist now. I think it is an important feature of Capsul that must work quickly in the new API and be tested. I propose to find a way to include subpipelines in TinyMorphologist.