Open nalimilan opened 2 years ago
I second this. It is most likely the reason for the issue I came here to open, namely that only a fraction of the packages I have used contain a Citation.bib file. Below you can see the severity of the issue:
(StatisticalMachineLearning) pkg> st
Status `~/StatisticalMachineLearning_drwatson/Project.toml`
[cbdf2221] AlgebraOfGraphics v0.6.5
[5d742f6a] CSVFiles v1.0.1
[aaaa29a8] Clustering v0.14.2
[a93c6f00] DataFrames v1.3.2
[b4f34e82] Distances v0.10.7
[634d3b9d] DrWatson v2.9.1
[e9467ef8] GLMakie v0.5.5
[6a3955dd] ImageFiltering v0.7.1
[6f286f6a] MultivariateStats v0.9.1
[b8a86587] NearestNeighbors v0.4.9
[6fe1bfb0] OffsetArrays v1.10.8
[50d0efc5] PkgCite v0.1.2
[08abe8d2] PrettyTables v1.3.1
[92933f4c] ProgressMeter v1.7.1
[df47a6cb] RData v0.8.3
[ce6b1742] RDatasets v0.7.7
[90137ffa] StaticArrays v1.4.2
[2913bbd2] StatsBase v0.33.16
[f3b207a7] StatsPlots v0.14.33
[a759f4b9] TimerOutputs v0.5.16
julia> get_tool_citation()
This work was done in \cite[Julia v1.7.1]{Julia-2017} and made use of the following packages: ArrayInterface.jl\cite[ArrayInterface]{DifferentialEquations.jl-2017}, Distributions.jl\cite[Distributions]{JSSv098i16,Distributions.jl-2019}, DrWatson.jl\cite[DrWatson]{DrWatson2020}, FFTW.jl\cite[FFTW]{FFTW.jl-2005}, Makie.jl\cite[Makie]{DanischKrumbiegel2021} and QuadGK.jl\cite[QuadGK]{quadgk}.
[ Info: The above sentence was copied to your clipboard.
┌ Warning: Overwriting julia_citations.bib
└ @ PkgCite ~/.julia/packages/PkgCite/T49Bu/src/citations.jl:69
[ Info: A julia_citations.bib file with the citations for DrWatson, QuadGK, Makie, Distributions, Julia, ArrayInterface and FFTW was generated in the current working directory (/home/legolas/StatisticalMachineLearning_drwatson).
7 out of 20 packages. And several of the most important/relevant ones are missing. I would love if the packages were at least mentioned by name and version, for giving credit and improving reproducibility.
Many packages do not provide CITATION.bib because they are not associated with a paper and/or because their authors are not in academia and don't care about this. I realized recently when writing a paper which cites many packages that it's really painful to cite them for this reason.
For these cases, it would be nice to have a fallback citation generation mechanism, like
citation
does in R. Project.toml contains anauthor
field which could be reused. The result could be something like "$AUTHORS, $DATE, $PACKAGE, Julia package version $VERSION, $URL." Doing this would also help establishing a convention about how to cite Julia packages.It could also be interesting to extract the project description from GitHub but that would require a bit more work I guess.