[WIP][NF] Add dipy_gtable cli

[skoudoro]

This PR adds a new CLI named dipy_gtable. The main goal of this cli is to merged bvec and bval into one file named *.gtab.csv. This new file will be required for the other CLI 's in a close future.

to test it, just run one this command line:

• dipy_gtable --force dipy/data/files/small_64D.bval dipy/data/files/small_64D.bvec
• dipy_gtable --force dipy/data/files/small_64D.bval dipy/data/files/small_64D.bvec --btens "LTE"
• dipy_gtable --force dipy/data/files/small_64D.bval dipy/data/files/small_64D.bvec --small_delta 0.01 --big_delta 0.03 --btens "LTE"

then display the new file: dipy_info gt_condensed.gtab.csv

[pep8speaks]

Hello @skoudoro, Thank you for updating !

Cheers ! There are no PEP8 issues in this Pull Request. 🍻

Comment last updated at 2024-03-21 15:30:42 UTC

[codecov]

Codecov Report

Attention: Patch coverage is 82.56881% with 19 lines in your changes are missing coverage. Please review.

Project coverage is 81.92%. Comparing base (d75eedd) to head (85f6218).
Report is 30 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3143      +/-   ##
==========================================
+ Coverage   81.78%   81.92%   +0.14%     
==========================================
  Files         150      150              
  Lines       21277    21365      +88     
  Branches     3412     3424      +12     
==========================================
+ Hits        17401    17504     +103     
+ Misses       3037     3022      -15     
  Partials      839      839              

Files	Coverage Δ
dipy/workflows/cli.py	42.85% <ø> (ø)
dipy/io/gradients.py	88.75% <91.89%> (+2.08%)	⬆️
dipy/workflows/io.py	75.00% <77.77%> (+0.21%)	⬆️
dipy/core/gradients.py	85.58% <77.77%> (-1.26%)	⬇️

... and 6 files with indirect coverage changes

[arokem]

Could we try to conform the parts that we can with https://mrtrix.readthedocs.io/en/latest/concepts/dw_scheme.html?

[skoudoro]

Could we try to conform the parts that we can with https://mrtrix.readthedocs.io/en/latest/concepts/dw_scheme.html?

I was not aware, I will look into it and let you know.

[skoudoro]

ok, I just look into it @arokem,

Both are ASCII files (csv, txt, dat, etc...). So you mean:

• space separated instead of comma separated
• change the order (bvec, bval) instead of (bval bvec)
• not enforce the extension instead of enforcing the extension
• do not insert header instead of inserting header ?

Also, we add more information like btensor, small_delta, big_delta, if available.

Can you tell me which one is important to respect. I personally would like to keep the header which help non-technical user to open the file in excel and understand it. The other point should be ok to update

[arokem]

I just mean that we should spend some time thinking before we invent yet another file format.

In terms of matching the .b files, I think that we get closer by using tab separated (rather than comma-separated; also more BIDS-friendly, if I am not mistaken), and inverting the order of b-values and b-vectors.

One way to ask this is to see whether a file that is generated here could immediately be used by mrtrix, without any change. I am not sure what mrtrix does on i/o, so the added information (i.e., b-tensor) might throw this off, but it's worth trying.

[skoudoro]

ok, thank you for the feedback!

ok for tab separated and inverting b-values and b-vectors. I will update

Otherwise, I will play with other software to check what they do when the file contain extra information like btensor or b0_threshold.

[Garyfallidis]

This is not the same purpose @arokem. Our version contains gradients, b-tensor encodings, small and big deltas etc. We will explain. And also provide a converter. Also we do not introduce a new file format. This is just a csv file. We already have pandas as a dependency.

[Garyfallidis]

Please see updated picture above (in description).

[Garyfallidis]

In addition, it makes sense for people to see b-values as the first column because this is what people look first. It is also part of our API we first provide b-values and then b-vectors.