if batch = TRUE in training, fgraph of umap_transform is transposed if n_epoch is set to 0

[PedroMilanezAlmeida]

on my system, this is a reproducible example:

iris_train <- iris[1:100,1:4]
iris_test <- iris[101:150,1:4]

iris_train_umap <- uwot::umap(iris_train, 
                              ret_model = TRUE,
                              batch = TRUE)

dim(uwot::umap_transform(iris_test, 
                         iris_train_umap,
                         n_epochs = NULL,
                         ret_extra = "fgraph")$fgraph)
[1] 100  50

dim(uwot::umap_transform(iris_test, 
                         iris_train_umap,
                         n_epochs = 0,
                         ret_extra = "fgraph")$fgraph)
[1]  50 100

R version 4.3.2 (2023-10-31)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)

Matrix products: default
BLAS/LAPACK: FlexiBLAS OPENBLAS;  LAPACK version 3.11.0

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8     LC_MONETARY=en_US.UTF-8   
 [6] LC_MESSAGES=en_US.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                  LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

time zone: America/New_York
tzcode source: system (glibc)

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
 [1] microbenchmark_1.4.10 compiler_4.3.2        RcppAnnoy_0.0.21      Matrix_1.6-4          tools_4.3.2           rstudioapi_0.15.0    
 [7] Rcpp_1.0.11           codetools_0.2-19      grid_4.3.2            snedata_0.0.0.9000    lattice_0.22-5        uwot_0.1.16

[PedroMilanezAlmeida]

fgraph of umap_transform also transposed if batch = TRUE and n_epochs = NULL

iris_train_umap <- uwot::umap(iris_train, 
                              ret_model = TRUE,
                              batch = FALSE)

iris_test_umap <- uwot::umap_transform(iris_test, 
                                       iris_train_umap,
                                       n_epochs = NULL,
                                       ret_extra = "fgraph")
dim(iris_test_umap$fgraph)
[1]  50 100

iris_train_umap <- uwot::umap(iris_train, 
                              ret_model = TRUE,
                              batch = TRUE)

iris_test_umap <- uwot::umap_transform(iris_test, 
                                       iris_train_umap,
                                       n_epochs = NULL,
                                       ret_extra = "fgraph")
dim(iris_test_umap$fgraph)
[1] 100  50

[jlmelville]

Thank you for the report! This was due to trying to avoid doing unnecessary processing when the UMAP optimization won't occur (but some of it actually is necessary either way). Should be fixed in the dev version now.

[jlmelville]

Will be fixed in the next CRAN release

[PedroMilanezAlmeida]

amazing, thank you!