Add reports module

[JulioAPeraza]

Closes #773.

Changes proposed in this pull request:

• Create HTML files with a summary and figures from a MetaResult object.

[codecov]

Codecov Report

Patch coverage: 98.27% and project coverage change: +0.50 🎉

Comparison is base (c1e8c31) 87.82% compared to head (0c881cd) 88.33%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #802      +/-   ##
==========================================
+ Coverage   87.82%   88.33%   +0.50%     
==========================================
  Files          43       46       +3     
  Lines        5643     5931     +288     
==========================================
+ Hits         4956     5239     +283     
- Misses        687      692       +5     

Impacted Files	Coverage Δ
nimare/__init__.py	100.00% <ø> (ø)
nimare/diagnostics.py	100.00% <ø> (ø)
nimare/reports/figures.py	97.95% <97.95%> (ø)
nimare/reports/base.py	98.36% <98.36%> (ø)
nimare/reports/__init__.py	100.00% <100.00%> (ø)
nimare/workflows/cbma.py	100.00% <100.00%> (ø)

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[JulioAPeraza]

It will be difficult to combine view_connectome and the heatmap in the same report.html file. Both plot use plotly, but view_connectome use a fixed plotly.js that is quite old. When I give priority to the latest version view_connectome doesn't render properly, and when I give priority to the old version view_connectome looks fine but the heatmap looks totally wrong (it showed a scatter plot instead of the heatmap).
One solution could be to embed the view_connectome plot to report.html using iframe. Or add a button to open the view_connectome in a separate window.

[JulioAPeraza]

I also noticed that when the heatmap is rendered before view_connectome, both plot looks fine.
This is the error a get when view_connectome is rendered first:

Uncaught TypeError: Plotly.plot is not a function

given that Plotly.plot was deprecated.

[jdkent]

what does the actual html file look like? I'm curious if different versions of the javascript library can be loaded at different sections of the html file.

[JulioAPeraza]

I think this is now ready for review.

[jdkent]

Thanks @JulioAPeraza, this is looking awesome! Just want to make sure we aren't working with a combinatorial explosion of parameters when specifying default.yml.

[jdkent]

just so I understand correctly, FDR, 'indep` are just defaults, there is no particular reason why these defaults were chosen, correct?

[JulioAPeraza]

We chose this combination for the documentation because the FWE montecarlo takes a long time to run with 1000 iterations. The reports function is going to show whatever it finds in the MetaResult object.

[jdkent]

do you need to enumerate every potential combination of parameters? what would need to be added here if we were to add a new corrector/diagnostic?

[JulioAPeraza]

do you need to enumerate every potential combination of parameters?

Yeah, that's very inefficient, but I couldn't figure out a good way to perform the combinations, like we do with the OS distributions and Python versions during testing.

what would need to be added here if we were to add a new corrector/diagnostic?

We would need to add new bids variables with all their possible combinations.

[jdkent]

I think we should mention that we re-order the columns and rows so they form clusters in the heatmap

[jdkent]

These changes look good to me! Great work, default.yml looks way more maintainable, we will tackle any incompatibilities when we come around to supporting some pairwise estimators and image based meta-analyses.