weirdness in rpy2

bla

In R:

png("heatmap_before_sorting.png")
image(distance_matrix)
dev.off()
library("gclus")
ordered <- order.single(distance_matrix, clusters=NULL)
sorted_matrix <- distance_matrix[ordered, ordered]

In Python:

import rpy2
# for automatic conversion of numpy matrices
import rpy2.robjects.numpy2ri
rpy2.robjects.numpy2ri.activate()
# import the gclus R package
from rpy2.robjects.packages import importr
gclus = importr('gclus')

import rpy2.robjects as robjects
r = robjects.r 
# make a heatmap out of the distance matrix and plot
grdevices = importr('grDevices')
grdevices.png(file="heatmap_before_sorting.png", width=512, height=512)
# make the heatmap
r["image"](distance_matrix) # notice the weird method call here, standard methods like image() are values in a dictionary (??)
grdevices.dev_off()
# now sort
ordered = gclus.order_single(distance_matrix, clusters=r("NULL")) # packages, on the other hand, have proper methods
                                                                  # also note that R's order.single() became order_single()
                                                                  # also note the weird r("NULL") instead of R's NULL

sorted_matrix = distance_matrix.rx(ordered, ordered) # notice the rx() method - R-like indexing, since ordered is a vector of indices