2020-09-29 21:18:48 +02:00
|
|
|
source("renv/activate.R")
|
|
|
|
|
library("dplyr")
|
2020-09-29 21:35:44 +02:00
|
|
|
library("tidyr")
|
2020-09-29 21:18:48 +02:00
|
|
|
|
|
|
|
|
# Using mostly indeixng instead of tidyr because is faster
|
|
|
|
|
resampled_episodes <- read.csv(snakemake@input[[1]])
|
|
|
|
|
resampled_episodes["n_resamples"] <- 1 + (resampled_episodes["end_timestamp"] - resampled_episodes["start_timestamp"]) %/% 60001
|
|
|
|
|
resampled_episodes <- resampled_episodes %>% uncount(n_resamples, .id = "nrow")
|
|
|
|
|
|
|
|
|
|
resampled_episodes["nrow"] <- (resampled_episodes["nrow"] - 1) * 60000
|
|
|
|
|
resampled_episodes["start_timestamp"] <- resampled_episodes["start_timestamp"] + resampled_episodes["nrow"]
|
|
|
|
|
# Use +59999 because each resampled minute should not overlap with each other
|
|
|
|
|
resampled_episodes["end_timestamp"] <- pmin(resampled_episodes["start_timestamp"] + 59999, resampled_episodes["end_timestamp"])
|
|
|
|
|
resampled_episodes <- resampled_episodes %>% select(-nrow)
|
|
|
|
|
resampled_episodes <- resampled_episodes %>% uncount(2, .id = "end_flag")
|
|
|
|
|
|
|
|
|
|
resampled_episodes["timestamp"] = NA_real_
|
|
|
|
|
resampled_episodes[resampled_episodes$end_flag ==1, "timestamp"] = resampled_episodes[resampled_episodes$end_flag ==1, "start_timestamp"]
|
|
|
|
|
resampled_episodes[resampled_episodes$end_flag ==2, "timestamp"] = resampled_episodes[resampled_episodes$end_flag ==2, "end_timestamp"]
|
|
|
|
|
resampled_episodes <- resampled_episodes %>% select(-end_flag)
|
|
|
|
|
|
|
|
|
|
write.csv(resampled_episodes, snakemake@output[[1]], row.names = FALSE)
|
