Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.09%. Comparing base (4ff4d61) to head (c9bedc4).

@@            Coverage Diff             @@
##           master     #128      +/-   ##
==========================================
+ Coverage   95.72%   96.09%   +0.36%     
==========================================
  Files          24       25       +1     
  Lines        2385     2585     +200     
==========================================
+ Hits         2283     2484     +201     
+ Misses        102      101       -1     

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

Alright, tests pass, coverage passes. The only thing failing is invalidations which as far as I can tell has nothing to do with my PR, it didn't add invalidations I think there is some version incompatibility in the action which causes an error?

@heltonmc What's the PR process for this package? I would be happy for a review or suggestions.

Also @juliohm and @oscardssmith since they were part of the chat on Julia Discourse.

Things I intend to leave for the future for now:

1. Dedicated real BigFloat branch rather than using the slower Complex{BigFloat} logic.
2. Export loggammas and gammas? Rebrand? Make Gamma.jl elsewhere? Not sure which if any of this is desirable.

Thank you for leading this @TSGut ! @oscardssmith is more qualified to help with a review.

Hope this initiative will get the attention it deserves from package authors and maintainers in JuliaMath 🙏🏽

Benchmarking the implementations with the strategy suggested by the readme I get (compared to ArbNumerics.jl):

And execution time is roughly the same as using SpecialFunctions.jl:

@oscardssmith: I think this covers everything in your code review -- anything else left to do before this can be merged?

I leave it to you whether you choose to export these functions but the sooner this can be tagged the sooner I can get started on making the hypergeometric functions be Julia native. 😄 I have had a chat with the maintainer of HypergeometricFunctions.jl and he's happy for me to basically copy and adjust his code as functions within MeijerG.jl, so MeijerG.jl will just contain its own native hypergeometric function logic and probably also export a Julia native pFq, with its only dependency becoming Bessels.jl.

The code looks good to me, but I do think or should probably live in a new Gamma.jl. there are a ton more people that likely will want a good gamma function (e.g. Primes.jl) that probably don't want to pull in all the bessel functions as well.

I am happy to do that, can you make a Gamma.jl in JuliaMath (we should check that no such named package is currently registered)? While we're at it could I be added to JuliaMath - would make this easier going forward.

I'll move this code into a PR for the new Gamma.jl then.

Also, if this will create a new Julia native Special functions ecosystem then we probably want to have references pointing at the other relevant packages in that ecosystem from the readmes or docs.

https://github.com/JuliaMath/Gamma.jl. If you make the PR, I'll work on getting this to be an actual Julia repo.

Great, I will close this PR since we are moving development there.