Parse nextflow execution_trace file

parse_execution_trace.R

library(ggplot2)

execution_trace <- '/home/chase/Documents/pipeline_info/execution_trace_2021-08-14_09-52-45.txt' 

execution_trace <- read.delim(execution_trace, sep = "\t")
process_names <- sapply(strsplit(execution_trace$name, " "), function(x) x[[1]])


split_names <- strsplit(process_names, ":")
max_len <- max(lengths(split_names))
split_names <- lapply(split_names,
                      function(x){
                        x[1:max_len]
                      })

split_names <- do.call(rbind, split_names)
colnames(split_names) <- paste0("group_", 1:ncol(split_names))
execution_trace <- cbind.data.frame(execution_trace, split_names)


temp <- strsplit(execution_trace$duration, " ")
duration <- lapply(temp, 
                   function(x){
                     time_unit <- sub("[^a-z]+", "", x)
                     sum(as.vector(c(ms = 1, s = 1e3, m = 1e3*60)[time_unit]) * as.numeric(sub("[a-z]+$", "", x)))
                   })
execution_trace$duration_seconds <- unlist(duration)

temp <- strsplit(execution_trace$realtime, " ")
realtime <- lapply(temp, 
                   function(x){
                     time_unit <- sub("[^a-z]+", "", x)
                     sum(as.vector(c(ms = 1, s = 1e3, m = 1e3*60)[time_unit]) * as.numeric(sub("[a-z]+$", "", x)))
                   })
execution_trace$realtime_seconds <- unlist(realtime)

ggplot(execution_trace) +
  geom_boxplot(aes( group_2, realtime_seconds)) +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))

https://www.nextflow.io/docs/latest/tracing.html#trace-report