R/classify_bemovi_files.R

In LEEF-UZH/LEEF.analysis: Access Functions, Tests and Basic Analysis of the RRD Data from the LEEF Project

#' Classify algae_traits data.frame
#'
#' @param datadir `character` vector containing the root directory for all files
#' @param bemovi_extract_name `character` vector containing the name of the
#'   bemovi config file including path.
#' @param classifier_constant_name `character` vector of name of the classifier
#'   for temperature treatment **constant** including path
#' @param classifier_increasing_name `character` vector of name of the
#'   classifier for temperature treatment **increasing** including path
#' @param exclude_videos file names of videos to exclude. If \code{NULL}, all will be used.
#'
#' @return `list` containing three objects:
#'       - `mean_density_per_ml`
#'       - `morph_mvt`
#'       - `trajectory_data`
#'
#' @importFrom yaml read_yaml
#' @importFrom utils read.table
#'
#' @export
#'
#' @md
#' @examples
classify_bemovi_files <- function(
    datadir,
    bemovi_extract_name = NULL,
    classifier_constant_name,
    classifier_increasing_name,
    exclude_videos = NULL) {
  p <- yaml::read_yaml(bemovi_extract_name)


  # The classification ------------------------------------------------------

  morph_mvt <- readRDS(file.path(datadir, p$merged.data.folder, p$morph_mvt))
  morph_mvt <- morph_mvt[, grep("_prob|species", names(morph_mvt), invert = TRUE)]

  traj <- readRDS(file.path(datadir, p$merged.data.folder, p$master))

  if (!is.null(exclude_videos)) {
    morph_mvt <- morph_mvt[!(morph_mvt$file %in% exclude_videos), ]
    traj <- traj[!(traj$file %in% exclude_videos), ]
  }
  classified <- LEEF.measurement.bemovi::classify(
    bemovi_extract = bemovi_extract_name,
    morph_mvt = morph_mvt,
    trajectory_data = traj,
    classifiers_constant = readRDS(classifier_constant_name),
    classifiers_increasing = readRDS(classifier_increasing_name),
    video_description_file = as.data.frame(read.table(file.path(datadir, p$video.description.folder, p$video.description.file), sep = "\t", header = TRUE, stringsAsFactors = FALSE)),
    composition = utils::read.csv(file.path(datadir, "compositions.csv"))
  )

# Correction for exclude_videos ----------------------------------------------

  vdf <- as.data.frame(
    read.table(
      file.path(datadir, p$video.description.folder, p$video.description.file),
      sep = "\t",
      header = TRUE
    )
  )

  dens_corr <- vdf |>
    select(bottle, file) |>
    filter(file %in% exclude_videos) |>
    group_by(bottle) |>
    mutate(
      dens_factor = 3 / (3 - n()),
      file = NULL
    )

  classified$mean_density_per_ml <-
    classified$mean_density_per_ml |>
    left_join(
      y = dens_corr,
      by = join_by(bottle)
    ) |>
    mutate(
      dens_factor = ifelse(
        is.na(dens_factor),
        1,
        dens_factor
      )
    ) |>
    mutate(
      density = density * dens_factor
    )

  # Return the classifications --------------------------------------------------


  return(classified)
}