As a check against bugs that are consistent within the behavior of the Vietoris–Rips computation, a simple test should compare the result to that of hclust() using method = "single". Here is an example, though note that the tolerance must be specified (and this should be re-checked once {ripserr} converts value_t from float to double):
library(phutil)
library(ripserr)
library(testthat)
d <- dist(UScitiesD)
hc <- hclust(d, method = "single")
hc_ <- as_persistence(hc)
ph <- vietoris_rips(d)
ph_ <- as_persistence(ph)
expect_equivalent(as.data.frame(hc_), as.data.frame(ph_), tolerance = 1e-7)
As a check against bugs that are consistent within the behavior of the Vietoris–Rips computation, a simple test should compare the result to that of
hclust()usingmethod = "single". Here is an example, though note that the tolerance must be specified (and this should be re-checked once {ripserr} convertsvalue_tfrom float to double):