R/read_acq_text.R

In SDAMcIntyre/countenance: Read, process and analyse facial EMG data

#' Read a text file exported from BIOPAC AcqKnowledge.
#'
#' Wrapper around readr::read_delim(). Additionally processes header lines with readr::read_lines(), and adds a time variable for convenient plotting.
#'
#' @param file Name of data file to read.
#' @param delim Single character used to separate fields within a record.
#' @param keep_channels A vector with the names of the channels in the data file to keep. All others are discarded.
#'
#' @return A tibble.
#' @export
#'
#' @examples
#' read_acq_text(
#'     countenance_example("femg_raw_exported_sample.txt"),
#'     delim = ",",
#'     keep_channels = c("Zyg Processed","Corr Processed","Marker")
#'     )
read_acq_text <- function(file, delim, keep_channels) {
  print(paste("Reading file", file))
  # extract numbers from 2nd row to get sample duration in ms and convert to sampling rate
  samp_rate_hz <- file %>%
    readr::read_lines(skip = 1, n_max = 1) %>%
    stringr::str_extract("[0-9]+") %>%
    as.numeric() %>%
    ( function(x) (1 / (x / 1000)) )

  # extract numbers from 3rd row to get number of channels in file
  raw_n_channels <- readr::read_lines(
    file,
    skip = 2,
    n_max = 1
    ) %>%
    stringr::str_extract("[0-9]+") %>%
    as.numeric()

  # read channel names
  # just get the odd ones for the names (even rows are measurement units)
  raw_channel_names <- readr::read_lines(
    file,
    skip = 3,
    n_max = 2*raw_n_channels
    )[seq(1,raw_n_channels*2, 2)]

  # get the column numbers for the data and stimulus channels by matching the names
  columns_to_keep <- raw_channel_names %>%
    stringr::str_which(paste(keep_channels, collapse = "|"))

  # read in the data, keep only the wanted channels
  raw_acq_data <- readr::read_delim(
    file,
    delim = delim,
    skip = raw_n_channels*2+5,
    col_names = FALSE,
    col_types = readr::cols()
    )[,columns_to_keep]
  # put the names back
  names(raw_acq_data) <- raw_channel_names[columns_to_keep]

  # add time variable
  raw_acq_data <- raw_acq_data %>%
    dplyr::mutate(time_sec = seq(0, dplyr::n()-1) / samp_rate_hz)

  return(raw_acq_data)
}