jeremygelb / geocmeans Goto Github PK

With the change from raster to terra, the expected input format for raster data is now a list of spaRaster objects and not raster objects. If one gives rasters object, the error produced is not helpful.

geocmeans should test the type of object given and return an error if they are not from the expected class.

I've noticed that the examples provided do not include seed values when utilizing random processes, such as clustering. I was wondering if you expect reproducibility or not. For example, in the following examples:

https://jeremygelb.github.io/geocmeans/articles/introduction.html

While it's understood that in clustering, some variability can be acceptable, setting a seed ensures reproducibility and consistency in the results, especially for examples and tutorials.

Thank you for your hard work on this package, and I look forward to any updates.

This issue is related to openjournals/joss-reviews#5259 .

Hi. Please note that multiprocess is deprecated since future 1.20.0 (2020-11-03) and will soon be defunct:

• https://github.com/JeremyGelb/geocmeans/search?q=multiprocess

Please take a look at the following example. The first block runs successfully, however, the second block raises the following error:

https://jeremygelb.github.io/geocmeans/articles/web_vignettes/advanced_examples.html

> # classical fuzzy cmeans
> result1 <- CMeans(df, 2, 1.5, standardize = TRUE, robust = FALSE,  verbose = FALSE, seed = 176)
Error in colMeans(x, na.rm = TRUE) : 'x' must be numeric

This package depends on (depends, imports or suggests) raster and one or more of the retiring packages rgdal, rgeos or maptools (https://r-spatial.org/r/2022/04/12/evolution.html, https://r-spatial.org/r/2022/12/14/evolution2.html). Since raster 3.6.3, all use of external FOSS library functionality has been transferred to terra, making the retiring packages very likely redundant. It would help greatly if you could remove dependencies on the retiring packages as soon as possible.

I am excited about this package! Thank you for developing it.

If I supply a list with only one raster to any of the clustering functions I get an error

"Error in `colnames<-`(`*tmp*`, value = okname) : 
  attempt to set 'colnames' on an object with less than two dimensions" 

But, supplying a list with two rasters works.

Expected behaviour: One should also be able to make a clustering using just a single variable, correct? e.g. a map of NDVI

> R.Version()
$platform
[1] "x86_64-w64-mingw32"
$arch
[1] "x86_64"
$os
[1] "mingw32"
$system
[1] "x86_64, mingw32"
$version.string
[1] "R version 4.1.2 (2021-11-01)"

Reproducible example

library(geocmeans) # 0.2.0
library(raster) # v 3.5-15

# example rasters r and s
r <- raster(nrows=10, ncols=10)
values(r) <- runif(ncell(r)) 
s <- raster(nrow = 10, ncol = 10)
values(s) <- rnorm(ncell(s))


CMeans(data = list(r), # one raster
               k = 3,
               m = 1.5,
               maxiter = 100,
               seed = 235)
# Error in `colnames<-`(`*tmp*`, value = okname) : 
# attempt to set 'colnames' on an object with less than two dimensions


CMeans(data = list(r,s), #using two rasters works
               k = 3,
               m = 1.5,
               maxiter = 100,
               seed = 235)

# [1] "Standardizing the data (set parameter to FALSE to avoid this step)"
# |=====================                     |  29%[1] "criterion reached"



# Same behaviour for other functions. e.g.


GCMeans(data = list(s), # doesn't work
        k = 3,
        m = 1.5,
        beta = 1,
        maxiter = 100,
        seed = 235)

#[1] "Standardizing the data (set parameter to FALSE to avoid this step)"
#Error in `colnames<-`(`*tmp*`, value = okname) : 
#  attempt to set 'colnames' on an object with less than two dimensions



window.matrix <- matrix(1/9,nrow = 3, ncol = 3)

SFCMeans(data = list(r),
         window = window.matrix,
         k = 3,
         m = 1.5,
         alpha = 1)

# Error in `colnames<-`(`*tmp*`, value = okname) : 
# attempt to set 'colnames' on an object with less than two dimensions

Hi,

First of all, thank you for developing this much needed package.

I ran into an error using mapCluster function.
Code
cmeansMaps<- mapClusters(reseauALL_C_sp,Cmean$Belongings,undecided = 0.45)
Message
Error in fix.by(by.y, y) : 'by' must specify a uniquely valid column

where reseauALL_C_sp is a sp dataframe, the same that was used to create the membership matrix.