r-tda

TDA is the general-purpose R package for the field, and it carries statistical machinery the Python stack tends not to — confidence bands, bootstrap inference on diagrams. It matters here because a great deal of bioinformatics is written in R, and a Python-only corpus would quietly exclude it. See r-tda-environment.

It is on CRAN and on no conda channel — not conda-forge, not bioconda, not the r channel — so this recipe is what makes the fixture installable at all.

This one has never been built, and unlike the others that is not merely an oversight. The package is NeedsCompilation: yes, linking against BH, Rcpp, and RcppEigen with gmp and make as system dependencies, and the host dependency set here was modelled on the r-designit recipe plus CRAN metadata rather than derived from a real build. The recipe says so itself. A rattler-build run is the only thing that will confirm it, and it is more likely to find something than the pure-Python entries are. r-tdastats-recipe, the compiled R sibling, does build green, which is mild evidence the general shape is right.