Submission to CRAN (proposal)

[kokitsuyuzaki]

Hi,

Thanks for the great package. This is really useful when the data structure is complex. Do you have a plan to submit this package to the CRAN repository? I want to use this functionality by importing this package.

Regards,

[const-ae]

Hey,

thank you for the interest in the package. Yes, I actually did plan to submit it to CRAN, but then there were some problems with the CRAN checks. Specifically, the mathjaxr that I used to properly typeset the equations in the documentation was problematic. And then my main research projects, for which I wrote the package moved on, so I simply got busy with other stuff.

I think there might be several ways forward:

1) Figure out why I get the error message and fix it 2) Remove all references to \mjseqn{} from the documentation and replace them with normal \eqn{} references

It actually should be pretty straightforward, so I will try and see what I can do next week.

---

Alternatively, as an immediate solution, you can also use einsum_generator(equation_string, compile_function = FALSE) to generate the source code for a C++ program that you could paste into your R package :)

Best, Constantin

[const-ae]

Of course, you could speed up the process by helping me debug the issue if you want to :)

[kokitsuyuzaki]

It seems that \text in the man/einsum.Rd does not work for some reason. I replaced it to \textrm and R CMD build could be performed. I'm not familiar with mathjaxr though, in R running in terminal, the equations were not converted as shown below, but I wonder if they would be displayed more beautifully in a desktop version of R.

library(einum)
?einsum

Details:

     The following table show, how the Einstein notation abbreviates
     complex summation for arrays/matrices:

       ‘equation_string’                                    Formula
       --------                                       -------------             ------------
       ‘"ij,jk->ik"’                Y_{ik} = sum_{j}{A_{ij} B_{jk}}    Matrix multiplication
       ‘"ij->ji"’`                                        Y = A^{T}                Transpose
       ‘"ii->i"’                               y = \textrm{diag}(A)                 Diagonal
       ‘"ii->ii"’                            Y = \textrm{diag}(A) I  Diagonal times Identity
       ‘"ii->"’               y = \textrm{trace}(A) = sum_i{A_{ii}}                    Trace
       ‘"ijk,mjj->i"’     y_i = sum_{j}sum_{k}sum_{m}A_{ijk}B_{mjj}     Complex 3D operation

[kokitsuyuzaki]

Hi, Constantin

I confirmed that mathjaxr properly worked in the desktop R, although it does not work in the terminal R.

I checked the source code and looks good to register einsum to CRAN.

If there is no problem, could you accept my pull request and register einsum to CRAN?

Koki

[const-ae]

Thanks for your help. I have submitted einsum to CRAN. You can track the process via https://lockedata.github.io/cransays/articles/dashboard.html

[const-ae]

It's on CRAN https://CRAN.R-project.org/package=einsum 🎉 🥳

Thanks again for helping me push this over the finish line :)

[kokitsuyuzaki]

Congrats!!!

[const-ae]

Thanks :)