Description
Interface to 'Geomstats'.
Description
Provides an interface to the Python package 'Geomstats' authored by Miolane et al. (2020) <arXiv:2004.04667>.
README.md
rgeomstats
The goal of rgeomstats is to provide accessibility to the Geomstats Python library for the community of R
users through an R interface that mimics as closely as possible the carefully designed Python API.
Installation
You can install the development version of rgeomstats via:
# install.packages("remotes")
remotes::install_github("LMJL-Alea/rgeomstats")
Example
You can instantiate the space $\mathrm{SO}(3)$ of 3D rotations and sample random points in this space as follows:
library(rgeomstats)
so3 <- SpecialOrthogonal(n = 3)
spl <- so3$random_point(n_samples = 5)
dim(spl)
#> [1] 5 3 3
All Geomstats-like computations are stored in arrays. In particular, sample IDs are always stored along the first dimension. Hence, it is always possible to convert a sample into a list via:
purrr::array_tree(spl, margin = 1)
#> [[1]]
#> [,1] [,2] [,3]
#> [1,] -0.03532813 -0.9154660 0.4008416
#> [2,] 0.68066355 0.2716386 0.6803746
#> [3,] -0.73174384 0.2968746 0.6135278
#>
#> [[2]]
#> [,1] [,2] [,3]
#> [1,] -0.3117327 -0.8280555376 0.4659901
#> [2,] 0.8305219 0.0007836864 0.5569855
#> [3,] -0.4615801 0.5606455319 0.6874739
#>
#> [[3]]
#> [,1] [,2] [,3]
#> [1,] -0.7025064 0.66878384 -0.2433370
#> [2,] -0.3513459 -0.02856766 0.9358098
#> [3,] 0.6189029 0.74290785 0.2550434
#>
#> [[4]]
#> [,1] [,2] [,3]
#> [1,] -0.8970871 -0.2258339 0.3797811
#> [2,] -0.1157827 -0.7093377 -0.6952945
#> [3,] 0.4264141 -0.6677119 0.6101900
#>
#> [[5]]
#> [,1] [,2] [,3]
#> [1,] -0.3877247 -0.3760100 -0.8415973
#> [2,] 0.4484487 -0.8746292 0.1841673
#> [3,] -0.8053343 -0.3060069 0.5077366