[REF] Merge CLI into workflows #100
Conversation
This could make it easier to update/add workflows. Instead of having to manage two files, we just need to manage one (although we still need to update the parser in step with the workflow and its docstring.
|
We'll need the CI passing before merging, but I don't have a strong feeling on this reorganization either way. I can see the appeal of it, though, and am happy to move forward if it makes things more intuitive ! I imagine it will also make @rmarkello happy, since he thinks we have too many modules 😄 |
Workflows can now take in a filename for data (i.e., a string) in addition to a list of files.
Codecov Report
@@ Coverage Diff @@
## master #100 +/- ##
=========================================
- Coverage 45.81% 45.42% -0.4%
=========================================
Files 29 28 -1
Lines 1602 1585 -17
=========================================
- Hits 734 720 -14
+ Misses 868 865 -3
Continue to review full report at Codecov.
|
| @@ -28,7 +27,7 @@ def compare_nifti(fn, test_dir, res_dir): | |||
| """ | |||
| res_fp = (res_dir/fn).as_posix() | |||
| test_fp = (test_dir/fn).as_posix() | |||
| assert np.allclose(nb.load(res_fp).get_data(), nb.load(test_fp).get_data()) | |||
| assert np.allclose(nib.load(res_fp).get_data(), nib.load(test_fp).get_data()) | |||
There was a problem hiding this comment.
Are we consistently using nib instead of nb, now ? I just want to make sure we don't end up with conflicting imports !
There was a problem hiding this comment.
I can change it back, but my understanding was that nib was the alias used in the nibabel documentation (e.g., here). On the other hand, while they may use that alias in at least some of their documentation, I've never seen anything official about it, so we can use nb if you prefer.
There was a problem hiding this comment.
It looks like that's our only import of it for now, and I know I'm the minority opinion so happy to change :) Just wanted to check !
|
This LGTM ! Ready to merge, @tsalo ? |
|
I think so, unless anyone has any concerns. |
|
I think @rmarkello is most likely to have feedback here (since he was involved in the original organization decisions) ! |
|
Love it! Thrilled to see consolidation of these modules. On a relatively cursory glance it looks good to me, but if you'd like me to do a line-by-line I can. It seems like @emdupre already checked it over though, so I'm happy to see it merged. |
|
💯 Merging, then ! Thanks everyone ! |
This reorganization would make it easier to update workflows, as the argument parser and the workflow function are consolidated in one file for each workflow (instead of spread between
tedana.cliandtedana.workflows.Changes proposed in this pull request:
tedana.clitotedana.workflows. Each workflow submodule now contains a function named[name]_workflow(e.g.,tedana_workflowandt2smap_workflow) as well as private functions such as_mainand_get_parserfor the argument parsing and CLI.setup.pyto call the_mainfunctions in the workflows for the CLI.