Make this the official Distributions module? #1

dmbates · 2012-12-06T19:48:41Z

@StefanKarpinski, @johnmyleswhite Should we go ahead and submit a pull request on METADATA.jl to make this the official Distributions module?

johnmyleswhite · 2012-12-06T20:05:34Z

I think I did that a while back.

StefanKarpinski · 2012-12-06T20:06:20Z

Yep, this is already the official version. I deleted the JuliaLang one to avoid confusion.

dmbates · 2012-12-06T20:25:56Z

Sorry for the noise. I was using my own copy of METADATA.jl.

dmbates · 2012-12-06T20:42:51Z

I am really getting confused now. I removed the ~/.julia directory and started from scratch with Pkg.init(). Now the METADATA/Distributions/url file points to this repository but Pkg.add("Distributions") seems to install an old version that doesn't have John's updates from Nov. 20 and earlier. What am I missing?

dmbates · 2012-12-06T21:01:33Z

Okay, so it seems as if the versions/0.0.0/sha1 and the hashes directories in METADATA.jl for this package were not updated after John's changes on Nov. 20. Was that intentional?

Also (and I ask this with some trepidation), do I need to keep multiple copies of the repository locally if I plan to modify the contents of the package? For example, there is a trivial spelling mistake in README.md ("funtions"), which I introduced many versions ago. Suppose that I correct that. Do I need to clone this repository and push from that clone or can I modify the copy in my ~/.julia directory and push from there?

I guess the simpler question is, are there recommended usage patterns for developers yet? I am willing to experiment and to contribute but I am floundering a bit here.

I was going to paraphrase Ronald Reagan and say that "git is not the solution; git is the problem" but that's unfair :-)

StefanKarpinski · 2012-12-06T21:23:25Z

I think the issue here is that pushing to master in the package repo does not publish that version, even when the package is at v0. You also have to push the new SHA1 hash to the appropriate spot in METADATA. @pao has been doing some work to make this much easier and more automatic.

You can definitely just work on your local .julia copy by intentional design. However, you first want to checkout the master branch since managed versions of package are installed with a detached head on a specific commit. I was just starting to work through some use cases here and will produce a document soon...

dmbates · 2012-12-06T21:46:51Z

Thanks @StefanKarpinski . I'll try to be patient and read more.

If I can venture one question, the repository returned by Pkg.origin()

julia> load("pkg")

julia> Pkg.origin()
"[email protected]:dmbates/.julia"

doesn't exist and I find this very confusing. I thought that I got rid of all the stale connections to out-of-date repositories and started from scratch yet somehow I have everything depending on a repository that doesn't exist.

Is it assumed that I should have created such a repository or have I somehow managed to get myself into another git nightmare entanglement?

StefanKarpinski · 2012-12-06T21:50:55Z

That's just a default location it picks for you if you have a GitHub user configuration, assuming that you'll want to keep it there. That's only ever used if you do Pkg.push or Pkg.pull to remotely store/sync your git repo, which is currently kind of broken, so it shouldn't really affect anything.

johnmyleswhite · 2012-12-06T22:01:52Z

On the train, so I'll need to chime in more later tonight, but I wanted to state that I had intentionally not switched over the SHA1 for this module to my latest revision until I had time to run things by with you, Doug.

johnmyleswhite · 2012-12-07T13:32:37Z

Sorry for all the confusion. I was slowly adding things, which is why I wasn't updating the SHA1's. I've now set things up so that METADATA.jl is in sync with this repo. From now on, I'll do that for every change.

We can probably close this issue once you have a working copy again, @dmbates.

pao · 2012-12-07T14:00:10Z

I've now set things up so that METADATA.jl is in sync with this repo.

Would you mind sharing how?

johnmyleswhite · 2012-12-07T14:01:22Z

Sorry for the confusion: it's not automatic. I just meant that I synchronized them manually.

pao · 2012-12-07T15:22:56Z

Ah, darn. I'm hoping to encode as many best practices as possible into Pkg--I'm not sure if that is a best practice yet, but if it turns out to be, I was hoping to steal it!

johnmyleswhite · 2012-12-07T15:26:12Z

Hard to say. I think it's nice to be able to have a bleeding-edge version of GitHub that isn't being used by the package system, but we could require that people use release and dev branches for that purpose.

dmbates · 2012-12-07T17:07:03Z

I'm all better now.

Update utils.jl wrt issue #425

updated multinomial entropy

Faster, more accurate Poisson Binomial PMF

…esCore.RuleConfig` This is what was happening implicitly already: https://github.com/JuliaDiff/ChainRulesCore.jl/blob/9627bd6a949ca88ba867bf4fa177a45bf780a248/src/rules.jl#L137-L138 ``` unning tests: 1 ambiguities found. To get a list, set `broken = false`. Ambiguity JuliaStats#1 frule(::ChainRulesCore.RuleConfig, args...) in ChainRulesCore at /home/runner/.julia/packages/ChainRulesCore/I1EbV/src/rules.jl:64 frule(::Any, ::typeof(Distributions.logpdf), d::Distributions.Uniform, x::Real) in Distributions.DistributionsChainRulesCoreExt at /home/runner/work/Distributions.jl/Distributions.jl/ext/DistributionsChainRulesCoreExt/univariate/continuous/uniform.jl:1 Possible fix, define frule(::ChainRulesCore.RuleConfig, ::typeof(Distributions.logpdf), ::Distributions.Uniform, ::Real) Aqua: Test Failed at /home/runner/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 Expression: iszero(num_ambiguities) Stacktrace: [1] _test_ambiguities(packages::Vector{Base.PkgId}; broken::Bool, kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 [2] _test_ambiguities @ ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:69 [inlined] [3] test_ambiguities(packages::Module; kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:28 [4] test_ambiguities(packages::Module) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:28 [5] macro expansion @ ~/work/Distributions.jl/Distributions.jl/test/aqua.jl:19 [inlined] [6] macro expansion @ /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.6/Test/src/Test.jl:1151 [inlined] [7] top-level scope @ ~/work/Distributions.jl/Distributions.jl/test/aqua.jl:9 ```

…version ``` Running tests: 1 ambiguities found. To get a list, set `broken = false`. Ambiguity JuliaStats#1 frule(::ChainRulesCore.RuleConfig, args...) in ChainRulesCore at /home/runner/.julia/packages/ChainRulesCore/I1EbV/src/rules.jl:64 frule(::Any, ::typeof(Distributions.logpdf), d::Distributions.Uniform, x::Real) in Distributions.DistributionsChainRulesCoreExt at /home/runner/work/Distributions.jl/Distributions.jl/ext/DistributionsChainRulesCoreExt/univariate/continuous/uniform.jl:1 Possible fix, define frule(::ChainRulesCore.RuleConfig, ::typeof(Distributions.logpdf), ::Distributions.Uniform, ::Real) Aqua: Test Failed at /home/runner/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 Expression: iszero(num_ambiguities) Stacktrace: [1] _test_ambiguities(packages::Vector{Base.PkgId}; broken::Bool, kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 [2] _test_ambiguities @ ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:69 [inlined] [3] test_ambiguities(packages::Module; kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:28 [4] test_ambiguities(packages::Module) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:28 [5] macro expansion @ ~/work/Distributions.jl/Distributions.jl/test/aqua.jl:19 [inlined] [6] macro expansion @ /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.6/Test/src/Test.jl:1151 [inlined] [7] top-level scope @ ~/work/Distributions.jl/Distributions.jl/test/aqua.jl:9 ```

I *think* this is a false-positive: ``` Running tests: 1 ambiguities found. To get a list, set `broken = false`. Ambiguity JuliaStats#1 frule(::ChainRulesCore.RuleConfig, args...) in ChainRulesCore at /home/runner/.julia/packages/ChainRulesCore/I1EbV/src/rules.jl:64 frule(::Any, ::typeof(Distributions.logpdf), d::Distributions.Uniform, x::Real) in Distributions.DistributionsChainRulesCoreExt at /home/runner/work/Distributions.jl/Distributions.jl/ext/DistributionsChainRulesCoreExt/univariate/continuous/uniform.jl:1 Possible fix, define frule(::ChainRulesCore.RuleConfig, ::typeof(Distributions.logpdf), ::Distributions.Uniform, ::Real) Aqua: Test Failed at /home/runner/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 Expression: iszero(num_ambiguities) Stacktrace: [1] _test_ambiguities(packages::Vector{Base.PkgId}; broken::Bool, kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 [2] _test_ambiguities @ ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:69 [inlined] [3] test_ambiguities(packages::Module; kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:28 [4] test_ambiguities(packages::Module) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:28 [5] macro expansion @ ~/work/Distributions.jl/Distributions.jl/test/aqua.jl:19 [inlined] [6] macro expansion @ /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.6/Test/src/Test.jl:1151 [inlined] [7] top-level scope @ ~/work/Distributions.jl/Distributions.jl/test/aqua.jl:9 ```

``` Running tests: 1 ambiguities found. To get a list, set `broken = false`. Ambiguity JuliaStats#1 frule(::ChainRulesCore.RuleConfig, args...) in ChainRulesCore at /home/runner/.julia/packages/ChainRulesCore/I1EbV/src/rules.jl:64 frule(::Any, ::typeof(Distributions.logpdf), d::Distributions.Uniform, x::Real) in Distributions.DistributionsChainRulesCoreExt at /home/runner/work/Distributions.jl/Distributions.jl/ext/DistributionsChainRulesCoreExt/univariate/continuous/uniform.jl:1 Possible fix, define frule(::ChainRulesCore.RuleConfig, ::typeof(Distributions.logpdf), ::Distributions.Uniform, ::Real) Aqua: Test Failed at /home/runner/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 Expression: iszero(num_ambiguities) Stacktrace: [1] macro expansion @ /opt/hostedtoolcache/julia/1.8.5/x64/share/julia/stdlib/v1.8/Test/src/Test.jl:464 [inlined] [2] _test_ambiguities(packages::Vector{Base.PkgId}; broken::Bool, kwargs::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 ```

``` Running tests: 1 ambiguities found. To get a list, set `broken = false`. Ambiguity JuliaStats#1 frule(::ChainRulesCore.RuleConfig, args...) in ChainRulesCore at /home/runner/.julia/packages/ChainRulesCore/I1EbV/src/rules.jl:64 frule(::Any, ::typeof(Distributions.logpdf), d::Distributions.Uniform, x::Real) in Distributions.DistributionsChainRulesCoreExt at /home/runner/work/Distributions.jl/Distributions.jl/ext/DistributionsChainRulesCoreExt/univariate/continuous/uniform.jl:1 Possible fix, define frule(::ChainRulesCore.RuleConfig, ::typeof(Distributions.logpdf), ::Distributions.Uniform, ::Real) Aqua: Test Failed at /home/runner/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 Expression: iszero(num_ambiguities) Stacktrace: [1] macro expansion @ /opt/hostedtoolcache/julia/1.8.5/x64/share/julia/stdlib/v1.8/Test/src/Test.jl:464 [inlined] [2] _test_ambiguities(packages::Vector{Base.PkgId}; broken::Bool, kwargs::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}) @ Aqua ~/.julia/packages/Aqua/tHrmY/src/ambiguities.jl:78 ``` ``` pdf/logpdf: Error During Test at /Users/runner/work/Distributions.jl/Distributions.jl/test/multivariate/jointorderstatistics.jl:76 Test threw exception Expression: #= /Users/runner/work/Distributions.jl/Distributions.jl/test/multivariate/jointorderstatistics.jl:76 =# @inferred(logpdf(d, x)) isa T DomainError with -0.31736627: log will only return a complex result if called with a complex argument. Try log(Complex(x)). Stacktrace: [1] throw_complex_domainerror(f::Symbol, x::Float32) @ Base.Math ./math.jl:33 [2] _log(x::Float32, base::Val{:ℯ}, func::Symbol) @ Base.Math ./special/log.jl:336 [3] log @ ./special/log.jl:264 [inlined] [4] log1mexp @ ~/.julia/packages/LogExpFunctions/X95j7/src/basicfuns.jl:240 [inlined] [5] logdiffcdf(d::Uniform{Float32}, x::Float32, y::Float32) @ Distributions ~/work/Distributions.jl/Distributions.jl/src/univariates.jl:372 [6] _marginalize_range @ ~/work/Distributions.jl/Distributions.jl/src/multivariate/jointorderstatistics.jl:125 [inlined] [7] logpdf(d::JointOrderStatistics{Uniform{Float32}, Tuple{Int64, Int64, Int64}}, x::Vector{Float32}) @ Distributions ~/work/Distributions.jl/Distributions.jl/src/multivariate/jointorderstatistics.jl:111 [8] macro expansion @ ~/hostedtoolcache/julia/1.8.5/aarch64/share/julia/stdlib/v1.8/Test/src/Test.jl:464 [inlined] [9] macro expansion @ ~/work/Distributions.jl/Distributions.jl/test/multivariate/jointorderstatistics.jl:76 [inlined] [10] macro expansion @ ~/hostedtoolcache/julia/1.8.5/aarch64/share/julia/stdlib/v1.8/Test/src/Test.jl:1363 [inlined] [11] macro expansion @ ~/work/Distributions.jl/Distributions.jl/test/multivariate/jointorderstatistics.jl:75 [inlined] [12] macro expansion @ ~/hostedtoolcache/julia/1.8.5/aarch64/share/julia/stdlib/v1.8/Test/src/Test.jl:1439 [inlined] [13] macro expansion @ ~/work/Distributions.jl/Distributions.jl/test/multivariate/jointorderstatistics.jl:35 [inlined] [14] macro expansion @ ~/hostedtoolcache/julia/1.8.5/aarch64/share/julia/stdlib/v1.8/Test/src/Test.jl:1363 [inlined] [15] top-level scope @ ~/work/Distributions.jl/Distributions.jl/test/multivariate/jointorderstatistics.jl:4 [16] include @ ./client.jl:476 [inlined] [17] macro expansion @ ~/work/Distributions.jl/Distributions.jl/test/runtests.jl:175 [inlined] [18] macro expansion @ ~/hostedtoolcache/julia/1.8.5/aarch64/share/julia/stdlib/v1.8/Test/src/Test.jl:1439 [inlined] [19] macro expansion @ ~/work/Distributions.jl/Distributions.jl/test/runtests.jl:174 [inlined] [20] macro expansion @ ~/hostedtoolcache/julia/1.8.5/aarch64/share/julia/stdlib/v1.8/Test/src/Test.jl:1363 [inlined] [21] top-level scope @ ~/work/Distributions.jl/Distributions.jl/test/runtests.jl:174 [22] include(fname::String) @ Base.MainInclude ./client.jl:476 [23] top-level scope @ none:6 [24] eval @ ./boot.jl:368 [inlined] [25] exec_options(opts::Base.JLOptions) @ Base ./client.jl:276 [26] _start() @ Base ./client.jl:522 ```

Fix `median` for `Binomial`

dmbates closed this as completed Dec 7, 2012

simonbyrne pushed a commit that referenced this issue Feb 8, 2016

Merge pull request #1 from evanmason/evanmason-patch-1

d477d59

Update utils.jl wrt issue #425

andreasnoack pushed a commit that referenced this issue Apr 22, 2016

Merge pull request #1 from esproff/esproff-patch-1

54daf6f

updated multinomial entropy

atteson mentioned this issue Jul 12, 2019

rand( TDist(Inf) ) gives NaN in v0.20.0 #929

Closed

johnczito pushed a commit that referenced this issue Oct 21, 2020

Merge pull request #1 from bdeonovic/bdeonovic-pb-pmf

951a242

Faster, more accurate Poisson Binomial PMF

devmotion pushed a commit that referenced this issue Dec 14, 2024

Merge pull request #1 from marcusps/marcusps/fix-binomial-median

515c5e2

Fix `median` for `Binomial`

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Make this the official Distributions module? #1

Make this the official Distributions module? #1

dmbates commented Dec 6, 2012

johnmyleswhite commented Dec 6, 2012

StefanKarpinski commented Dec 6, 2012

dmbates commented Dec 6, 2012

dmbates commented Dec 6, 2012

dmbates commented Dec 6, 2012

StefanKarpinski commented Dec 6, 2012

dmbates commented Dec 6, 2012

StefanKarpinski commented Dec 6, 2012

johnmyleswhite commented Dec 6, 2012

johnmyleswhite commented Dec 7, 2012

pao commented Dec 7, 2012

johnmyleswhite commented Dec 7, 2012

pao commented Dec 7, 2012

johnmyleswhite commented Dec 7, 2012

dmbates commented Dec 7, 2012

Make this the official Distributions module? #1

Make this the official Distributions module? #1

Comments

dmbates commented Dec 6, 2012

johnmyleswhite commented Dec 6, 2012

StefanKarpinski commented Dec 6, 2012

dmbates commented Dec 6, 2012

dmbates commented Dec 6, 2012

dmbates commented Dec 6, 2012

StefanKarpinski commented Dec 6, 2012

dmbates commented Dec 6, 2012

StefanKarpinski commented Dec 6, 2012

johnmyleswhite commented Dec 6, 2012

johnmyleswhite commented Dec 7, 2012

pao commented Dec 7, 2012

johnmyleswhite commented Dec 7, 2012

pao commented Dec 7, 2012

johnmyleswhite commented Dec 7, 2012

dmbates commented Dec 7, 2012