R/video_scores.R

In transforEmotion: Sentiment Analysis for Text, Image and Video using Transformer Models

#' Run FER on YouTube video
#'
#' This function retrieves FER scores a specific number of frames extracted from YouTube video. It uses Python libraries for facial recognition and emotion detection in text, images, and videos.
#'
#' @param video The URL of the YouTube video to analyze.
#' @param classes A character vector specifying the classes to analyze.
#' @param nframes The number of frames to analyze in the video. Default is 100.
#' @param face_selection The method for selecting faces in the video. Options are "largest", "left", or "right". Default is "largest".
#' @param start The start time of the video range to analyze. Default is 0.
#' @param end The end time of the video range to analyze. Default is -1 and this means that video won't be cut. If end is a positive number greater than start, the video will be cut from start to end.
#' @param uniform Logical indicating whether to uniformly sample frames from the video. Default is FALSE.
#' @param ffreq The frame frequency for sampling frames from the video. Default is 15.
#' @param save_video Logical indicating whether to save the analyzed video. Default is FALSE.
#' @param save_frames Logical indicating whether to save the analyzed frames. Default is FALSE.
#' @param save_dir The directory to save the analyzed frames. Default is "temp/".
#' @param video_name The name of the analyzed video. Default is "temp".
#'
#' @return A result object containing the analyzed video scores.
#'
#' @author Aleksandar Tomasevic <atomashevic@gmail.com>
#' @import reticulate
#'
#' @export
#

video_scores <- function(video, classes, nframes=100,
                         face_selection = "largest", start = 0, end = -1, uniform = FALSE, ffreq = 15, save_video = FALSE, save_frames = FALSE, save_dir = "temp/", video_name = "temp"){
  if (!conda_check()){
      stop("Python environment 'transforEmotion' is not available. Please run setup_miniconda() to install it.")
    
  }
  else {
    reticulate::use_condaenv("transforEmotion", required = FALSE)
  }
  
  reticulate::source_python(system.file("python", "image.py", package = "transforEmotion"))

  reticulate::source_python(system.file("python", "video.py", package = "transforEmotion"))

  # check if classes is a character vector
  if(!is.character(classes)){
    stop("Classes must be a character vector.")
  }

  # check if we have at least 2 classes
  if(length(classes) < 2){
    stop("Classes must have at least 2 elements.")
  }

  # check if face_selection is valid
  if(!face_selection %in% c("largest", "left", "right")){
    stop("Argument face_selection must be one of: largest, left, right")
  }

  # create save_dir
  if(!dir.exists(save_dir)){
      dir.create(save_dir)
  }

  if (!grepl("^http", video)) {
    if(!file.exists(video)){
      stop("Video file does not exist. If path is an URL make sure it includes http:// or https://")
    }}
  else { if(!grepl("youtu", video)){
    stop("You need to provide a YouTube video URL.")
  }
  }
  
  result = reticulate::py$yt_analyze(url = video, nframes = nframes, labels = classes,
             side= face_selection, start = start, end = end, uniform = uniform, ff = ffreq, frame_dir = save_dir, video_name = video_name)

  if (!save_video & grepl("youtu", video)){
    file.remove(paste0(save_dir, video_name, ".mp4"))
  }
  if(!save_frames){
     file.remove(paste0(save_dir, list.files(save_dir, pattern = ".jpg")))
  }
  return(result)
}