Cannot directly skip `rearrange` in `autoplot`

[jmbarbone]

The problem

I have a correlation matrix with a lot of missing values. This causes the backend rearranging to fail when it tries to change the order. However, there's direct way of preventing the reordering attempt, which I would like to do. Using method = "Identity" still checks if the object is valid for seriation.

Reproducible example

library(corrr)
x <- matrix(1:20, nrow = 5)
x[seq(1, 20, 2)] <- NA
x <- as.data.frame(x)
x <- correlate(x)
#> Correlation computed with
#> • Method: 'pearson'
#> • Missing treated using: 'pairwise.complete.obs'
x
#> # A tibble: 4 × 5
#>   term     V1    V2    V3    V4
#>   <chr> <dbl> <dbl> <dbl> <dbl>
#> 1 V1       NA    NA     1    NA
#> 2 V2       NA    NA    NA     1
#> 3 V3        1    NA    NA    NA
#> 4 V4       NA     1    NA    NA
try(autoplot(x))
#> Error in svd(x, nu = 0, nv = k) : infinite or missing values in 'x'
try(autoplot(x, method = "Identity"))
#> Error in seriate.dist(dist(m), method = method) : 
#>   NAs not allowed in distance matrix x!
try(rearrange(x))
#> Error in svd(x, nu = 0, nv = k) : infinite or missing values in 'x'
try(rearrange(x, method = "Identity"))
#> Error in seriate.dist(dist(m), method = method) : 
#>   NAs not allowed in distance matrix x!

class(x) <- c("foo", class(x))

rearrange.foo <- function(x, method = NA, absolute = TRUE) {
  if (is.na(method)) {
    return(x)
  }
  
  NextMethod()
}

try(rearrange(x, method = NA))
#> # A tibble: 4 × 5
#>   term     V1    V2    V3    V4
#>   <chr> <dbl> <dbl> <dbl> <dbl>
#> 1 V1       NA    NA     1    NA
#> 2 V2       NA    NA    NA     1
#> 3 V3        1    NA    NA    NA
#> 4 V4       NA     1    NA    NA
try(autoplot(x, method = NA))

^{Created on 2022-12-13 with reprex v2.0.2}