nvihrs14 / tcherry Goto Github PK

This issue is part of the JOSS review for tcherry.

I had some issues installing the package and building the vignettes due to the dependencies.

Bioconductor packages: I would note in your install instructions that the graph and RBGL packages come from Bioconductor, and that you might need to install them through BiocManager. At first, I tried to install them with install.packages and the error message was not helpful.

Dependencies in the vignette: I was unable to build the vignette due to missing packages gRain and Rgraphviz. Either add these to the dependencies, or better yet, remove them from the vignette if they are not essential.

Hello! It might be nice to follow the R naming conventions for function names, though certainly not required - I think this would just require changing "thinning_edges" to "thin_edges" and "BIC_junction_tree" to something like "compute_BIC_junction_tree".

This issue is part of the JOSS review for tcherry.

Statement of need: I feel that the paper is unclear on A) what kind of problem the software solves (e.g. finding relationships between variables in some certain kind of data) and B) who the audience for this package is (e.g. biologists?). Please clarify these points.

References: not all citations have a corresponding DOI. Please add the DOI to every reference that has has one.

Proofread: once you have revised the content, please check the grammar for verb tense. I am happy to help with this if you'd like!

This issue is part of the JOSS review for tcherry.

One criteria is that the GitHub version matches the package version. Currently, the repository shows no release version for the software. Please tag the package release once you are done with the other revisions.

I successfully install the package and all its dependence, however, when I run the example code, this error emerged, is there any idea about it?

Hi, What R version do I need to install the "tcherry" package? Thank you

Check the readme.md file. The argument for increase_order2 is not called 'cliques' (but tch_cliq) and therefore the example code does not run.

tcherry::tcherry_step() gives adjacency matrix via adj_matrix, whereas tcherry::tcherry_CL() gives it via adj_tcherry. The interface would be simpler to program with (e.g. learner function as argument to a function) if they (and possibly others) were called the same.

If possible, please add a contributing.md file so that others know how to contribute. Details can be found here and an example here.