In FlukeAndFeather/breath: Analyze whale breathing behavior

library(tidyverse)

drive_path <- "/Volumes/GoogleDrive/Shared drives/CATS/"
find_nc <- function(id, root) {
  dep_dir <- fs::dir_ls(
    fs::path(drive_path, "tag_data"),
    regexp = sprintf("%s", id)
  )
  if (length(dep_dir) != 1) {
    return(NA)
  }

  result <- fs::dir_ls(dep_dir, regexp = sprintf(".*%s.*nc$", id))
  if (length(result) != 1) {
    return(NA)
  } 

  result
}
bb_prh <- readxl::read_excel("prh_guide.xlsx", sheet = "bb") %>% 
  mutate(
    tz = sprintf("Etc/GMT%+d", -utc),
    tag_on_utc = lubridate::force_tzs(tag_on, tzones = tz),
    tag_off_utc = lubridate::force_tzs(tag_off, tzones = tz),
    prh_path = map_chr(id, find_nc, root = drive_path)
  )

22/27 deployments with NetCDF

# Deployment metadata
this_row <- filter(bb_prh, !is.na(prh_path)) %>% slice(1)
this_id <- this_row$id
this_path <- this_row$prh_path
this_tz <- this_row$tz

# Read data and truncate to tagon/tagoff
this_deployment <- read_deployment(this_path, this_tz) %>%
  smooth_p(1) %>%
  decimate(1)
this_deployment$data <- filter(
  this_deployment$data, 
  between(dt, this_row$tag_on_utc, this_row$tag_off_utc)
)

# Find dives
this_dives <- find_breaths_dives(
  this_deployment,
  interval =  10,
  surface = 2,
  ibi_thr = 300
)

summarize_dives(this_dives)

plot_dive_event(this_dives, 8, 60)

this_dives$data %>% 
  mutate(hour = floor(as.numeric(dt - min(dt), unit = "hours"))) %>% 
  group_by(id = this_id,
           hour) %>% 
  summarize(n_breaths = sum(is_breath)) %>% 
  ungroup() %>% 
  head(-1)

Run it for everyone

all_dives <- bb_prh %>% 
  filter(!is.na(prh_path)) %>% 
  group_by_all() %>% 
  group_modify(function(data, keys) {
    deployment <- read_deployment(keys$prh_path, keys$tz) %>%
      smooth_p(1) %>%
      decimate(1)
    deployment$data <- filter(
      deployment$data, 
      between(dt, keys$tag_on_utc, keys$tag_off_utc)
    )
    find_breaths_dives(
      deployment,
      interval =  10,
      surface = 2,
      ibi_thr = 300
    )$data
  }) %>% 
  ungroup()

breath_rates <- all_dives %>% 
  group_by(id) %>% 
  mutate(hour = floor(as.numeric(dt - min(dt), unit = "hours"))) %>% 
  group_by(id, hour) %>% 
  summarize(n_breaths = sum(is_breath)) %>% 
  ungroup() %>% 
  group_by(id) %>% 
  filter(hour < max(hour)) %>% 
  ungroup() 

med_rates <- breath_rates %>% 
  group_by(id) %>% 
  summarize(med_rate = median(n_breaths)) %>% 
  ungroup() %>% 
  arrange(med_rate)
breath_rates %>% 
  mutate(id = factor(id, levels = med_rates$id)) %>% 
  ggplot(aes(id, n_breaths)) +
  geom_boxplot() +
  coord_flip()