ENH: Add parametric.py module, add plot_regresion_profiles to plot.py, and update plot_als_comparison.py example

[earoy]

Added a new module parametric.py that performs OLS and LME modeling over the nodes of a given tract to allow for group comparisons, while controlling for potential covariates. Additionally, added another plotting function to plot.py that allows the user to visualize the "tract profiles" generated by the linear models and updated the plot_als_comparison.py example to demonstrate how to use these functions.

[arokem]

@earoy : I just added you as a collaborator to this repo, so hopefully you won't have to wait for me to approve running the CI on each commit.

[arokem]

Quick Q for now, but no time for a full review yet: Is it possible to tell from the formula whether this is an lme or not?

[arokem]

Meaning, is this argument redundant with information already provided in the formula?

[earoy]

I've some digging through the statsmodels documentation and it looks like for mixed-effects models, the LME function takes an R-style formula for the fixed effects (which is identical to the OLS formula) and a groups parameter to specify the random effects, so I don't think we can tell from the formula....One thought though is to have the user pass in their model formula which we can then parse to determine whether it's OLS or MLE and then populate the call to the statsmodels LME function with the corresponding random effects structure if needed

[arokem]

I'd avoid creating our own "domain specific language" and addressing all the possible corner cases may become pretty hairy, so maybe asking the user to be explicit about it (as you already do here) is best.

[coveralls]

Pull Request Test Coverage Report for Build 18de8b4d33257e1b712c59bc64d8a00d0d4a6dc2-PR-148

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 0.0%

---

Totals
Change from base Build cbf13cf8f2660feceb8f57d5d1eb4b2c3e98b073:	0.0%
Covered Lines:	0
Relevant Lines:	0

---

💛 - Coveralls

[arokem]

Looks like the CI error is unrelated to your changes: https://github.com/yeatmanlab/AFQ-Insight/actions/runs/9519578215/job/26243067587?pr=148#step:5:146, but instead comes from this line:

AFQ-Insight/afqinsight/plot.py

Line 323 in a43faac

for legobj in leg.legendHandles:

, that you didn't even touch. I'll try to fix this here by pinning the version of Matplotlib as a stop-gap, so that we can make progress on this PR. Stand by.

[arokem]

OK - I pushed this straight into your branch, so before you keep working here, please pull the recent commit from this branch to your local copy. Let's see if pinning Matplotlib fixes the error we saw in the CI before, and then we can proceed to review and merge this PR.

[coveralls]

Pull Request Test Coverage Report for Build ea525697945ac011e2cb21bfcc37087f0686adb0-PR-148

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 0.0%

---

Totals
Change from base Build cbf13cf8f2660feceb8f57d5d1eb4b2c3e98b073:	0.0%
Covered Lines:	0
Relevant Lines:	0

---

💛 - Coveralls

[arokem]

This will be fixed once this PR in copt is merged: openopt/copt#111 and we cut a new release of copt.

[arokem]

Hey @earoy : this took a while to fix and now we have a new "upstream" fork in https://github.com/tractometry/AFQ-Insight. Do you have an interest in trying to resume this work with a PR there? If you are, I can give some guidance on how you would redirect this PR there.

[earoy]

Hi @arokem, I'd love to keep working on getting this bundled into the new fork! Any guidance on how to properly do that would be most welcome!