formatter with JuliaFormatter

[mmikhasenko]

Formatting cleanup using JuliaFormatter.format(".") to ensure consistent code style across the package. This reduces noisy diffs in future changes and improves readability. No functional changes.

Closes #1401

Modifications

• add .formating with scripts
• modified README.md. Added formatting rules
• added .git-blame-ignore-revs to
• applied formatting

Tests

• are passing

Once merged:

Update .git-blame-ignore-revs to point to the merge commit

[devmotion]

I'd actually prefer Runic, it seems to work fine in StatsFuns and PDMats and I've been bitten by JuliaFormatter 2 problems a few times. GLM uses JuliaFormatter though, so there is no agreed on standard in JuliaStats yet.

[mmikhasenko]

I'd actually prefer Runic, it seems to work fine in StatsFuns and PDMats and I've been bitten by JuliaFormatter 2 problems a few times. GLM uses JuliaFormatter though, so there is no agreed on standard in JuliaStats yet.

completely fine with me, I do not have preference

[codecov-commenter]

Codecov Report

❌ Patch coverage is 84.13384% with 294 lines in your changes missing coverage. Please review.
✅ Project coverage is 86.43%. Comparing base (f1ff9e8) to head (56d05d2).
⚠️ Report is 6 commits behind head on master.

Files with missing lines	Patch %	Lines
src/samplers/obsoleted.jl	0.00%	26 Missing ⚠️
src/univariate/continuous/kolmogorov.jl	39.47%	23 Missing ⚠️
src/univariate/continuous/chernoff.jl	52.77%	17 Missing ⚠️
src/multivariate/mvnormal.jl	72.97%	10 Missing ⚠️
src/univariate/continuous/ksdist.jl	75.00%	10 Missing ⚠️
src/univariate/continuous/triweight.jl	56.52%	10 Missing ⚠️
src/univariates.jl	77.27%	10 Missing ⚠️
src/multivariate/dirichlet.jl	71.87%	9 Missing ⚠️
src/multivariate/mvtdist.jl	65.38%	9 Missing ⚠️
src/mixtures/mixturemodel.jl	78.78%	7 Missing ⚠️
... and 73 more

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1987      +/-   ##
==========================================
- Coverage   86.44%   86.43%   -0.01%     
==========================================
  Files         147      147              
  Lines        8838     8861      +23     
==========================================
+ Hits         7640     7659      +19     
- Misses       1198     1202       +4     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[mmikhasenko]

The work is complete. Tests are passing.
Could you please review?

[mmikhasenko]

ping @st-- , the author of original issue #1401, for an opinion.

[devmotion]

No need to change these lines:

Suggested change

	- "min"
	- "1"
	- 'min'
	- '1'

[devmotion]

Maybe let's merge #1905 before changing the formatting.

[ararslan]

I've been bitten by JuliaFormatter 2 problems a few times

You can pin to v1, which the JuliaFormatter README recommends if you run into issues with v2. My employer uses JuliaFormatter v2 with YASGuide extensively and have generally only run into issues in very unusual code, in which case you can surround in #!format off/on. MixedModels uses JuliaFormatter with BlueStyle on CI and that works fine.

The formatting implemented by Runic is largely inconsistent with the prevailing, albeit historically ad hoc, formatting in JuliaStats. I also find that some of its choices actively hinder readability vs. YAS or Blue.

I'd be fine with this PR if we use JuliaFormatter as initially proposed with one of the aforementioned styles, but I do not agree with using Runic.

[mmikhasenko]

Switched back to JuliaFormatting and adopted the MixedModels approach to formatting.
Hope the tests pass, it it can be merged

[devmotion]

Switched back to JuliaFormatting

I'd still prefer Runic. PDMats, StatsFuns and possibly soon StatsBase use Runic. Outside of StatsBase, it is used quite a lot as well - the whole SciML ecosystem recently switched to Runic, Documenter uses Runic, DifferentiationInterface uses Runic.... Generally, there seems to be a significant movement towards Runic

[mmikhasenko]

I have an idea -- you can create a competing PR, and see which one gets merged first.
I can also do it, if you like.

Would be good to close the story, that it does not hang in my open pull requests

[rsenne]

significant movement towards Runic

My unsolicited two cents is that some of the movement toward Runic may be because JuliaFormatter v2 is not considered stable yet, which the package maintainers have acknowledged and is part of why they do not recommend using v2 for testing purposes yet.

Runic is great, but pinning JuliaFormatter to v1.0.62 removes the burden of formatter behavior changing between minor releases. This used to be a bigger issue as this caused conflicts with JET but JuliaSyntax v1 is supported on JET. JuliaFormatter also seems to be more steadily maintained now, so I don’t think people necessarily need to give up on it yet. That said, if people simply prefer Runic’s style, that is totally reasonable too.

Edit: I missed the earlier comment making the same point. Sorry about that. Leaving this up since it adds a bit on pinning JuliaFormatter and the JET/JuliaSyntax compatibility issue.

[devmotion]

There's no consensus but also no pressing need for a change, so this PR or a rebased version of it is not expected to be merge in the near future. Therefor I'm closing the PR as it is outdated anyways.