Merge branch 'master' into 36-cell-level-metadata

Signed-off-by: German Beldorati Stark <[email protected]>

pipeline-runner/R/gem2s-7-upload_to_aws.R

@@ -256,7 +256,9 @@ get_metadata_cell_ids <- function(scdata_list, valid_metadata_name, value) {
 extract_subset_user_metadata <- function(subset_cellsets) {
   # metadata keys are the <track_name>-<value>, and name are the values alone
   metadata <- unique(subset_cellsets[type == "metadata"], by = "key")
-  metadata[, metadata_track := gsub(paste0("-", name, "$"), "", key), by = "key"]
+  # escape regex characters to avoid errors when creating cellsets
+  metadata[, escaped_name := gsub("([.\\\\+*?\\[^\\]$(){}=!<>|:-])", "\\\\\\1", name, perl = TRUE)]
+  metadata[, metadata_track := gsub(paste0("-", escaped_name, "$"), "", key), by = "key"]
 
   metadata <- metadata[, c("metadata_track", "name")]
   data.table::setnames(metadata, "name", "metadata_value")

pipeline-runner/R/qc-7-embed_and_cluster.R

@@ -68,6 +68,7 @@ format_cluster_cellsets <- function(cell_sets,
                                     name = paste0(clustering_method, " clusters")) {
   message("Formatting cluster cellsets.")
 
+  # careful with capital l on type for the key.
   cell_sets_object <-
     list(
       key = clustering_method,
@@ -76,7 +77,7 @@ format_cluster_cellsets <- function(cell_sets,
       type = "cellSets",
       children = list()
     )
-  for (cluster in sort(unique(cell_sets$cluster))) {
+  for (cluster in sort_cluster_names(unique(cell_sets$cluster))) {
     cells <- cell_sets[cell_sets$cluster == cluster, "cell_ids"]
     is.num <- !is.na(as.numeric(cluster))
     set_name <- ifelse(is.num, paste("Cluster", cluster), cluster)
@@ -97,6 +98,7 @@ format_cluster_cellsets <- function(cell_sets,
 }
 
 
+
 #' replace cell class by key
 #'
 #' Replaces an entire cell class (group of cellsets) by key. Used to replace
@@ -431,3 +433,22 @@ format_cl_metadata_cellsets <-
 
     return(cl_metadata_cellsets)
   }
+
+#' Sort cluster names
+#'
+#' Sorts cluster names naturally, i.e. Cluster 1, Cluster 2, Cluster 10
+#'
+#' @param strings cluster names
+#'
+#' @return sorted vector
+#' @export
+#'
+sort_cluster_names <- function(strings) {
+  # extract letters and digits
+  char <- gsub("\\d", "", strings)
+  nums <- gsub("\\D", "", strings)
+
+  sorted_indices <- order(char, as.integer(nums))
+
+  return(strings[sorted_indices])
+}