convert_by_segment.R

library(tidyverse)
library(phonfieldwork)

files <- list.files(pattern = "TextGrid")
walk(files, function(file){
  textgrid <- textgrid_to_df(file) 
  textgrid |> 
    mutate(content = str_extract(content, "^.*?-"),
           content = str_remove(content, "-"),
           content = str_replace_all(content, "SS", "ss"),
           content = str_replace_all(content, "S", "s"),
           content = str_replace_all(content, "Z", "s"),
           content = str_replace_all(content, "d", "t"),
           content = str_replace_all(content, "g", "k"),
           content = str_replace_all(content, "b", "p"),
           content = str_replace_all(content, "A", "ä")) |> 
    df_to_tier(file)
})

walk(files, function(file){
  textgrid <- textgrid_to_df(file) 
  textgrid |>
    filter(tier == 2) |> 
    select(time_start, time_end, content) |> 
    mutate(content = str_squish(content), 
           content2 = str_split(content, "")) |> 
    unnest_longer(content2) |> 
    group_by(time_start, time_end, content) |> 
    summarise(content2 = str_c(content2, collapse = "-")) |> 
    mutate(n_annotations = str_count(content2, "-")+1,
           content2 = str_split(content2, "-")) |>  
    unnest_longer(content2) |> 
    group_by(time_start, time_end, content) |> 
    mutate(id = 1:n()-1,
           id_rev = n():1-1,
           diff = (time_end - time_start)/n_annotations,
           time_start = time_start+id*diff,
           time_end = time_end-id_rev*diff) |> 
    ungroup()  |> 
    select(time_start, time_end, content2) |> 
    rename(content = content2) ->
    result

  result |> 
    arrange(time_start) |> 
    df_to_tier(file, tier_name = "auSegments")
})