NF: Add array parsing capabilities to the CLIs

[areeshatariq]

Description

Add array parsing capabilities to the CLIs (e.g. that would allow to get coherent segmentations with the dipy_median_otsu command) . Once this is fixed, the screenshots in the following PRs would need to be generated and included in #2275 https://github.com/dipy/dipy_data/pull/29 https://github.com/dipy/dipy_data/pull/30 https://github.com/dipy/dipy_data/pull/31 https://github.com/dipy/dipy_data/pull/32 https://github.com/dipy/dipy_data/pull/33

[skoudoro]

Hi @areeshatariq, Hi @jhlegarreta,

I do not remember what needs to be done here. I do not really understand the issue. Can one of you clarify the problem here?

Thanks!

[jhlegarreta]

If I remember correctly, array-like arguments are not parsed correctly by the CLIs. For the CLI that raised this issue, the vol_idx argument should accept None or array-like arguments, and it looked not to be parsing those correctly. Should be reproducible by calling the CLI at issue and checking if they are parsed correctly. Sorry for not being more specific.

I guess the problem is common to all CLIs, so processing arguments of this type at a single place would be best.

If I remember correctly, at the time you already had implemented a tentative fix locally that needed more testing.

[skoudoro]

I tried this: dipy_median_otsu ~/.dipy/cfin_multib/__DTI_AX_ep2d_2_5_iso_33d_20141015095334_4.nii --vol_idx 0 2 4 6

and it works as expected. not sure what was the issue, I can not reproduce it.

[skoudoro]

ok, I found a little issue but not related.

it seems we have an issue with median_otsu and stanford_hardi data. The result looks bad. I tried

dipy_median_otsu ~/.dipy/stanford_hardi/HARDI150.nii.gz --vol_idx 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 --out_dir "stanford_hardi_mask" --force --numpass 1 --dilate 2 --median_radius 3 --autocrop

I think we should rename this issue.

EDIT: dipy_horizon ~/.dipy/stanford_hardi/HARDI150.nii.gz: The data look a bit strange so the mask seems normal finally

[drombas]

Hi @skoudoro,

In case it helps. This is what I get:

• A: using your command
• B: using the command in the reconstruction tutorial:

Don't both look incomplete/strange ? especially the B, which is taking non-brain regions in. It might be worth checking how that affects the entire tutorial.

[skoudoro]

closing in favor of #3176

Tutorials need to be checked. Other PR's exists for this task