Merge pull request #16 from MoseleyBioinformaticsLab/column_major

Column major

.Rbuildignore

@@ -14,3 +14,5 @@
 ^\.github$
 ^check_long_long_int\.R$
 ^draft_vignettes\.Rmd$
+^.*\.log$
+^other_stuff$

.gitignore

@@ -11,3 +11,5 @@ Meta
 /Meta/
 inst/doc
 *ICIKendallTau_run.log
+/other_stuff/
+/large_test/

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: ICIKendallTau
 Title: Calculates information-content-informed Kendall-tau
-Version: 0.3.22
+Version: 1.0.2
 Date: 2024-01-26
 Authors@R: c(person(given = c("Robert", "M"), family = "Flight", role =
         c("aut", "cre"), email = "[email protected]", comment =
@@ -16,9 +16,9 @@ Encoding: UTF-8
 Roxygen: list(markdown = TRUE)
 RoxygenNote: 7.2.3
 LinkingTo: Rcpp
-Imports: Rcpp, purrr, utils
+Imports: Rcpp, purrr, utils, stringr
 Suggests: furrr, future, testthat (>= 3.0.0), microbenchmark,
-        rmarkdown, knitr, dplyr
+        rmarkdown, knitr, dplyr, logger
 URL: https://moseleybioinformaticslab.github.io/ICIKendallTau
         https://github.com/moseleybioinformaticslab/ICIKendallTau
 Config/testthat/edition: 3

NEWS.md

@@ -1,3 +1,7 @@
+# ICIKendallTau 1.0.0
+
+- Calculates correlation between columns of the matrix, **not** the rows.
+
 # ICIKendallTau 0.3.20
 
 - `kt_fast` now uses the same data.frame format for output as `ici_kendalltau`, but returns a matrix by default. The data.frame is useful when large amounts of comparisons are run.

R/kendalltau.R

@@ -3,7 +3,7 @@
 #' Given a data-matrix, computes the information-content-informed (ICI) Kendall-tau-b between
 #' all samples.
 #' 
-#' @param data_matrix samples are rows, features are columns
+#' @param data_matrix samples are columns, features are rows
 #' @param global_na what values should be treated as missing (NA)?
 #' @param zero_value what is the actual zero value?
 #' @param perspective how to treat missing data in denominator and ties, see details
@@ -25,9 +25,7 @@ ici_kendalltau_ref = function(data_matrix,
                              diag_good = TRUE,
                              progress = FALSE){
 
-  # assume row-wise (because that is what the description states), so need to transpose
-  # because `cor` actually does things columnwise.
-  data_matrix = t(data_matrix)
+
   exclude_loc = matrix(FALSE, nrow = nrow(data_matrix), ncol = ncol(data_matrix))
 
   # Actual NA and Inf values are special cases, so we do
@@ -102,7 +100,7 @@ missing_either = function(in_x, in_y){
 #' Calculates the completeness between any two samples using "or", is an
 #' entry missing in either X "or" Y.
 #' 
-#' @param data_matrix samples are rows, features are columns
+#' @param data_matrix samples are columns, features are rows
 #' @param global_na globally, what should be treated as NA?
 #' @param include_only is there certain comparisons to do?
 #' @param return_matrix should the matrix or data.frame be returned?
@@ -115,7 +113,6 @@ pairwise_completeness = function(data_matrix,
                                 include_only = NULL,
                                 return_matrix = TRUE){
 
-  data_matrix = t(data_matrix)
 
   if (is.null(colnames(data_matrix))) {
     stop("rownames of data_matrix cannot be NULL!")
@@ -248,7 +245,7 @@ pairwise_completeness = function(data_matrix,
 #' Given a data-matrix, computes the information-theoretic Kendall-tau-b between
 #' all samples.
 #' 
-#' @param data_matrix samples are rows, features are columns
+#' @param data_matrix samples are columns, features are rows
 #' @param global_na globally, what should be treated as NA?
 #' @param perspective how to treat missing data in denominator and ties, see details
 #' @param scale_max should everything be scaled compared to the maximum correlation?
@@ -285,7 +282,7 @@ pairwise_completeness = function(data_matrix,
 #' 
 #' matrix_1 = cbind(s1, s2)
 #' 
-#' r_1 = ici_kendalltau(t(matrix_1))
+#' r_1 = ici_kendalltau(matrix_1)
 #' r_1$cor
 #' 
 #' #    s1 s2
@@ -301,16 +298,16 @@ pairwise_completeness = function(data_matrix,
 #' s4[sample(100, 50)] = NA
 #' 
 #' matrix_2 = cbind(s3, s4)
-#' r_2 = ici_kendalltau(t(matrix_2))
+#' r_2 = ici_kendalltau(matrix_2)
 #' r_2$cor
 #' #           s3        s4
 #' # s3 1.0000000 0.9944616
 #' # s4 0.9944616 1.0000000
 #' 
 #' # using include_only
 #' set.seed(1234)
-#' x = matrix(rnorm(5000), nrow = 100, ncol = 50)
-#' rownames(x) = paste0("s", seq(1, nrow(x)))
+#' x = t(matrix(rnorm(5000), nrow = 100, ncol = 50))
+#' colnames(x) = paste0("s", seq(1, nrow(x)))
 #' 
 #' # only calculate correlations of other columns with "s1"
 #' include_s1 = "s1"
@@ -342,11 +339,14 @@ ici_kendalltau = function(data_matrix,
 
   do_log_memory = get("memory", envir = icikt_logger)
   # assume row-wise (because that is what the description states), so need to transpose
-  data_matrix = t(data_matrix)
   exclude_loc = matrix(FALSE, nrow = nrow(data_matrix), ncol = ncol(data_matrix))
 
+  if (is.data.frame(data_matrix)) {
+    data_matrix = as.matrix(data_matrix)
+  }
+
   if (is.null(colnames(data_matrix))) {
-    stop("rownames of data_matrix cannot be NULL!")
+    stop("colnames of data_matrix cannot be NULL!")
   }
 
   log_message("Processing missing values ...\n")

R/reshaping.R

@@ -1,8 +1,10 @@
+utils::globalVariables(c("stack", "ind", "values"))
+
 #' convert matrix to data.frame
 #' 
 #' Given a square correlation matrix, converts it to a long data.frame, with three columns.
 #' 
-#' @param cor_matrix the correlation matrix
+#' @param in_matrix the correlation matrix
 #' 
 #' @details The data.frame contains three columns:
 #' * s1: the first entry of comparison
@@ -32,6 +34,7 @@ cor_matrix_2_long_df = function(in_matrix)
 #' Given a long data.frame, converts it to a possibly square correlation matrix
 #' 
 #' @param long_df the long data.frame
+#' @param is_square should it be a square matrix?
 #' 
 #' 
 #' @export

README.Rmd

@@ -80,7 +80,7 @@ s2 = s1 + 10
 
 matrix_1 = cbind(s1, s2)
 
-r_1 = ici_kendalltau(t(matrix_1))
+r_1 = ici_kendalltau(matrix_1)
 r_1$cor
 ```
 
@@ -95,7 +95,7 @@ s4 = s2
 s4[sample(100, 50)] = NA
 
 matrix_2 = cbind(s3, s4)
-r_2 = ici_kendalltau(t(matrix_2))
+r_2 = ici_kendalltau(matrix_2)
 r_2$cor
 ```
 
@@ -150,7 +150,7 @@ If you have {future} and the {furrr} packages installed, then it is also possibl
 #| eval: false
 library(furrr)
 future::plan(multicore, workers = 4)
-r_3 = ici_kendalltau(t(matrix_2))
+r_3 = ici_kendalltau(matrix_2)
 ```
 
 ## Many Many Comparisons
@@ -163,7 +163,7 @@ To keep the `data.frame` output, add the argument `return_matrix=FALSE` to the c
 
 ```{r}
 #| label: matrix
-r_4 = ici_kendalltau(t(matrix_2), return_matrix = FALSE)
+r_4 = ici_kendalltau(matrix_2, return_matrix = FALSE)
 r_4
 ```
 

README.md

@@ -83,7 +83,7 @@ s2 = s1 + 10
 
 matrix_1 = cbind(s1, s2)
 
-r_1 = ici_kendalltau(t(matrix_1))
+r_1 = ici_kendalltau(matrix_1)
 r_1$cor
 #>    s1 s2
 #> s1  1  1
@@ -101,7 +101,7 @@ s4 = s2
 s4[sample(100, 50)] = NA
 
 matrix_2 = cbind(s3, s4)
-r_2 = ici_kendalltau(t(matrix_2))
+r_2 = ici_kendalltau(matrix_2)
 r_2$cor
 #>           s3        s4
 #> s3 1.0000000 0.9944616
@@ -136,10 +136,10 @@ microbenchmark(
   times = 5
 )
 #> Unit: microseconds
-#>                           expr       min        lq       mean    median        uq       max
-#>  cor(x, y, method = "kendall") 13379.183 13401.532 13493.0316 13401.829 13595.070 13687.544
-#>         ici_kt(x, y, "global")   271.586   279.007   369.9088   314.779   357.086   627.086
-#>       ici_kt(x2, y2, "global") 15983.902 16484.666 16952.9788 16667.885 17124.932 18503.509
+#>                           expr      min        lq      mean    median        uq       max
+#>  cor(x, y, method = "kendall") 13568.01 13818.243 14017.885 14090.436 14174.800 14437.933
+#>         ici_kt(x, y, "global")   261.84   272.733   304.271   287.571   335.669   363.542
+#>       ici_kt(x2, y2, "global") 14631.30 15614.509 16419.494 16340.620 17646.365 17864.673
 #>  neval
 #>      5
 #>      5
@@ -177,7 +177,7 @@ resources for any multiprocessing engine registered with {future}.
 ``` r
 library(furrr)
 future::plan(multicore, workers = 4)
-r_3 = ici_kendalltau(t(matrix_2))
+r_3 = ici_kendalltau(matrix_2)
 ```
 
 ## Many Many Comparisons
@@ -193,7 +193,7 @@ are converted to a matrix form if requested (the default).s To keep the
 of `ici_kendalltau`.
 
 ``` r
-r_4 = ici_kendalltau(t(matrix_2), return_matrix = FALSE)
+r_4 = ici_kendalltau(matrix_2, return_matrix = FALSE)
 r_4
 #> $cor
 #>   s1 s2 core       raw pvalue   taumax       cor
@@ -202,7 +202,7 @@ r_4
 #> 3 s4 s4    0 1.0000000      0 1.000000 1.0000000
 #> 
 #> $run_time
-#> [1] 0.02034521
+#> [1] 0.0196352
 ```
 
 ## Code of Conduct