BIDS MRI import integration test (PR 3-bis)

[MaximeBICMTL]

Builds on top of #1389 (diff)

Description

The PR this builds on heavily refactors the BIDS MRI import code, which is obviously a risky move. This PR adds an integration test for the BIDS MRI importer.

Dataset

The bids dataset used is one I made myself using the DCC090_587630_V2 Raisinbread DICOMs, and using the BidsGuess suggestions provided by dcm2niix to arrange the files. This dataset need to be added to the testing imaging bucket for this PR to pass the new test.

The dataset looks as follow:

/data/loris/incoming/DCC090_587630_V2/
├── dataset_description.json
├── derivatives
│   ├── dataset_description.json
│   └── sub-DCC090
│       └── ses-V2
│           ├── eddy
│           │   ├── sub-DCC090_ses-V2_acq-epb5_dir-AP_run-7_dwi.bval
│           │   ├── sub-DCC090_ses-V2_acq-epb5_dir-AP_run-7_dwi.bvec
│           │   ├── sub-DCC090_ses-V2_acq-epb5_dir-AP_run-7_dwi.json
│           │   └── sub-DCC090_ses-V2_acq-epb5_dir-AP_run-7_dwi.nii.gz
│           ├── preprocessing
│           │   ├── sub-DCC090_ses-V2_dir-RL_run-9_dwi.json
│           │   └── sub-DCC090_ses-V2_dir-RL_run-9_dwi.nii.gz
│           └── topup
│               ├── sub-DCC090_ses-V2_acq-epb0_dir-AP_run-6_dwi.json
│               ├── sub-DCC090_ses-V2_acq-epb0_dir-AP_run-6_dwi.nii.gz
│               ├── sub-DCC090_ses-V2_acq-epb0_dir-AP_run-8_dwi.json
│               └── sub-DCC090_ses-V2_acq-epb0_dir-AP_run-8_dwi.nii.gz
├── participants.tsv
└── sub-DCC090
    └── ses-V2
        ├── anat
        │   ├── sub-DCC090_ses-V2_acq-spc3p2_run-4_T2w.json
        │   ├── sub-DCC090_ses-V2_acq-spc3p2_run-4_T2w.nii.gz
        │   ├── sub-DCC090_ses-V2_acq-tfl3p2_run-3_T1w.json
        │   ├── sub-DCC090_ses-V2_acq-tfl3p2_run-3_T1w.nii.gz
        │   ├── sub-DCC090_ses-V2_acq-tse2p2_run-2_echo-1_T2w.json
        │   ├── sub-DCC090_ses-V2_acq-tse2p2_run-2_echo-1_T2w.nii.gz
        │   ├── sub-DCC090_ses-V2_acq-tse2p2_run-2_echo-2_T2w.json
        │   └── sub-DCC090_ses-V2_acq-tse2p2_run-2_echo-2_T2w.nii.gz
        └── dwi
            ├── sub-DCC090_ses-V2_acq-epb0_dir-AP_run-5_dwi.bval
            ├── sub-DCC090_ses-V2_acq-epb0_dir-AP_run-5_dwi.bvec
            ├── sub-DCC090_ses-V2_acq-epb0_dir-AP_run-5_dwi.json
            └── sub-DCC090_ses-V2_acq-epb0_dir-AP_run-5_dwi.nii.gz

Comparison to main

This PR targets the reworked importer instead of the original. It is because the original importer currently does not pass the tests. It seems to suffer from path joins as string concatenations issues (sub-1234ses-456anat kind of issues...). And I honestly don't really want to go fix this kind of issues just to replace the whole code afterwards, especially since the import of bvals and bvecs also seems broken in the main code, as well as the import of MRI derivatives (but well, it does not work here either TBH, at least this PR provides a test dataset to implement it correctly in the future).