added count_vars_ML.R to check numbers for revised counts

scripts/count_vars_ML.R

@@ -0,0 +1,66 @@
+# count numbers for ML
+
+#source("~/git/LSHTM_analysis/config/alr.R")
+source("~/git/LSHTM_analysis/config/embb.R")
+
+#source("~/git/LSHTM_analysis/config/gid.R")
+#source("~/git/LSHTM_analysis/config/pnca.R")
+
+source("~/git/LSHTM_analysis/scripts/plotting/get_plotting_dfs.R")
+
+nrow(merged_df3)
+##############################################
+#=============
+# mutation_info: revised labels
+#==============
+table(merged_df3$mutation_info)
+sum(table(merged_df3$mutation_info))
+sum(table(merged_df3$mutation_info))
+##############################################
+
+#=============
+# <drug>, dst_mode: revised labels
+#==============
+table(merged_df3$dst) # orig
+sum(table(merged_df3$dst))
+
+table(merged_df3$dst_mode)
+#table(merged_df3[dr_muts_col])
+sum(table(merged_df3$drtype_mode))
+
+##############################################
+#=============
+# drtype: revised labels
+#==============
+table(merged_df3$drtype) #orig
+
+table(merged_df3$drtype_mode)
+# mapping 2.1: numeric
+# drtype_map = {'XDR': 5
+#   , 'Pre-XDR': 4
+#   , 'MDR': 3
+#   , 'Pre-MDR': 2
+#   , 'Other': 1
+#   , 'Sensitive': 0}
+
+# create a labels col that is mapped based on drtype_mode
+merged_df3$drtype_mode_labels = merged_df3$drtype_mode
+merged_df3$drtype_mode_labels = as.factor(merged_df3$drtype_mode)
+
+levels(merged_df3$drtype_mode_labels)
+
+levels(merged_df3$drtype_mode_labels) <- c('Sensitive', 'Other'
+                                    , 'Pre-MDR', 'MDR'
+                                    , 'Pre-XDR', 'XDR')
+levels(merged_df3$drtype_mode_labels)
+
+# check
+#table(merged_df3$drtype)
+table(merged_df3$drtype_mode)
+table(merged_df3$drtype_mode_labels)
+sum(table(merged_df3$drtype_mode_labels))
+##############################################
+# lineage
+table(merged_df3$lineage)
+sum(table(merged_df3$lineage_labels))
+