In Lightbridge-KS/labChartHRV: LabChart's HRV Data Import and Manipulation

knitr::opts_knit$set(root.dir = rprojroot::find_rstudio_root_file()) # Set WD to Root
here::i_am("dev/hrv.Rmd")
library(here)

Plan

Function

read_HRV_reports: (USER) a wrapper

Input: path-to-hrv-folder
Return: a data frame of HRV reports combined

parse_HRV_reports (USER)

Input: character vector of HRV reports
Return: a data frame of HRV reports combined

new_extract_HRV_field (Internal): a constructor of extract_HRV_field

extract_HRV_field (Internal ?)

Input: character vector of HRV reports
Return: character vector of particular fields

Example

extract_HRV_field$Name()

Lookup

regex_hrv() (internal): Function that return look up list for HRV regex.

HRV_vars_desc: variable and description of HRV report.

HRV_vars_domain: list containing variable names of time and frequency domain.

usethis::use_r("HRV_reports")
rstudioapi::navigateToFile("tests/testthat/test-HRV_reports.R")
# usethis::use_test("HRV_reports")

Read HRV reports

read_HRV_reports <- function(file,
                             format_cols = TRUE,
                             encoding = "UTF-16LE",
                             ...
                             ){
  # Read Raw
  df_raw <- readtext::readtext(file = file, 
                               encoding = encoding,
                               ...
                               )
  # Parse Character to HRV tibble
  df_hrv <- parse_HRV_reports(df_raw[["text"]], 
                              format_cols = format_cols)
  # Bind `doc_id` to HRV tibble
  df_out <- dplyr::bind_cols(doc_id = df_raw[["doc_id"]], df_hrv)
  df_out
}

read_HRV_reports(sipsANS_example("HRV"))

Howto: Read HRV

Testing bind columns

t_df <- tibble(a = 1:2, b = 3:4)
c <- c(5:6)

dplyr::bind_cols(t_df, data.frame(c = c))

cbind(t_df, c = c)

readtext::encoding(sipsANS_example("HRV"))

readtext::readtext(sipsANS_example("HRV"), encoding = "UTF-16LE")$text %>% 
  parse_HRV_reports()

Parse HRV reports

parse_HRV_reports <- function(x, format_cols = TRUE) {

  # Regex: Extract after equal sign
  after_equal_regex <- c("(?<== )([:digit:]|[:punct:])+")

  nm_self <-  stats::setNames(names(regex_hrv()), names(regex_hrv())) 

  report_df <- nm_self %>%
    # Step 1: Extract HRV fields into DF
    purrr::map_dfc(
      ~ extract_HRV_field[[.x]](x)
    ) %>%
    # Step 2: Some vars have to extract number after equal sign
    dplyr::mutate(
      dplyr::across(
        c("VLF", "LF", "HF", "LF_HF"),
        ~ stringr::str_extract(.x, after_equal_regex)
      )
    )

  ## Not format column
  if( !format_cols){ return(report_df) }

  ## Format column 
  report_df %>%   
    dplyr::mutate(dplyr::across(c(Channel, Gender), factor),
         dplyr::across(Age, as.integer),
         dplyr::across(c(Beats_tot,Rec_length,
                         Discontinuities:Power_tot,
                  VLF,LF, LF_nu, HF, HF_nu, LF_HF), as.numeric)
         ) %>% 
    # Add pNN50
    dplyr::mutate(NN50_percent = 100 * (NN50_count/Normals_count),
                  .after = NN50_count) 

}

parse_HRV_reports(hrv_sim_raw$text)

Howto: Parse HRV

hrv_sim_raw$text %>% 
  extract_HRV_field[["File_LabChart"]]()

2 steps extract

VLF, LF, HF, LF_HF

after_equal <-  c("(?<== )([:digit:]|[:punct:])+")

names(extract_HRV_field) %>% 
  setNames(names(extract_HRV_field)) %>% 
  purrr::map_dfc(
    ~extract_HRV_field[[.x]](hrv_sim_raw$text)
  ) %>% 
  mutate(across(c("VLF", "LF", "HF", "LF_HF"), 
                ~stringr::str_extract(.x, after_equal))
         )

Extract Field

usethis::use_r("HRV_field")
usethis::use_test("HRV_field")

Constructor

new_extract_HRV_field <- function() {

  extract_factory <- function(nm) {
    function(x) {
      trimws(stringr::str_extract(x, regex_hrv()[[nm]]))
    }
  }

  nm_self <-  stats::setNames(names(regex_hrv()), names(regex_hrv())) 
  f <- purrr::map(nm_self, extract_factory)
  f
}

trimws(c(" a ", " b "))

Extract Fields

extract_HRV_field <- new_extract_HRV_field()
names(extract_HRV_field)

extract_HRV_field$Date(c("Date: 122", "Date: 123"))

extract_HRV_field$Channel(hrv1_chr_raw)

#extract_HRV_field$after_equal("a = 12.3")

extract_HRV_field$VLF("VLF (DC-0.04Hz) = 114.5 ms²")
extract_HRV_field$VLF_freq("VLF (DC-0.04Hz) = 114.5 ms²")

Lookup Regexes

## Extract after equal sign
after_equal = c("(?<== )([:digit:]|[:punct:])+")

regex_hrv <- function() {

  list(
    # Line 1
    File_LabChart = c("(?<=File: \").+(?=\")"),
    Channel = c("(?<=Channel: ).+(?=Date)"),
    Date = c("(?<=Date: ).+"),
    # Line 2
    Start_time = c("(?<=Start time: ).+(?=End time)"),
    End_time = c("(?<=End time: ).+"),
    # Line 3
    Name = c("(?<=Name: )[:graph:]+"),
    Gender = c("(?<=Gender: )[:alpha:]+"),
    Age = c("(?<=Age: )([:digit:]|[:punct:])+"),
    # Line 4
    Beats_tot = c("(?<=Total number of beats = )([:digit:]|[:punct:])+"),
    Rec_length = c("(?<=Length of recording = )([:digit:]|[:punct:])+"),
    # Line 5
    Class_bound = c("(?<=Classification boundaries: ).+"),
    # Line 6
    Discontinuities = c("(?<=Discontinuities = )([:digit:]|[:punct:])+"),
    Beats_inserted = c("(?<=Manually inserted beats = )([:digit:]|[:punct:])+"),
    Beats_deleted = c("(?<=Manually deleted beats = )([:digit:]|[:punct:])+"),
    # Line 7
    NN_max = c("(?<=Maximum NN = )([:digit:]|[:punct:])+"),
    NN_min = c("(?<=Minimum NN = )([:digit:]|[:punct:])+"),
    NN_range = c("(?<=Range = )([:digit:]|[:punct:])+"),
    # Line 8
    NN_mean = c("(?<=Mean NN = )([:digit:]|[:punct:])+"),
    NN_median = c("(?<=Median NN = )([:digit:]|[:punct:])+"),
    HR_avg = c("(?<=Average heart rate = )([:digit:]|[:punct:])+"),
    # Line 9
    SDNN = c("(?<=SDNN = )([:digit:]|[:punct:])+"),
    SD_del_NN = c("(?<=SD of delta NN = )([:digit:]|[:punct:])+"),
    RMSSD = c("(?<=RMSSD = )([:digit:]|[:punct:])+"),
    # Line 10
    Normals_count = c("(?<=Normals = )([:digit:]|[:punct:])+"),
    Ectopics_count = c("(?<=Ectopics = )([:digit:]|[:punct:])+"),
    Artifacts_count = c("(?<=Artifacts = )([:digit:]|[:punct:])+"),
    NN50_count = c("(?<=NN50 = )([:digit:]|[:punct:])+"),
    # Line 11
    Spec_intv = c("(?<=Spectrum intervals = )([:digit:]|[:punct:])+"),
    Spec_mean_NN = c("(?<=Mean spectrum NN = )([:digit:]|[:punct:])+"),

    # Line 12
    Power_tot = c("(?<=Total power = )([:digit:]|[:punct:])+"),
    VLF_freq = c("(?<=VLF \\().+(?=\\))"),
    VLF = c("VLF \\(.+\\) = ([:digit:]|[:punct:])+"),
    # Last Line
    LF_freq = c("(?<=((?<!V)(LF) \\())[^\\)]+"),
    LF = c("(?<!V)(LF) \\(.+\\) = ([:digit:]|[:punct:])+"),
    LF_nu = c("(\\.|[:digit:])+(?=( nu\\)[:blank:]+HF))"),

    HF_freq = c("(?<=((HF) \\())[^\\)]+"),
    HF = c("(HF) \\(.+\\) = ([:digit:]|[:punct:])+"),
    HF_nu = c("(\\.|[:digit:])+(?=( nu\\)[:blank:]+LF/HF))"),

    LF_HF = c("(LF/HF) = ([:digit:]|[:punct:])+")

  )

}