In stefanocoretta/phonetisr: A Naive IPA Tokeniser

knitr::opts_chunk$set(echo = TRUE)
knitr::opts_knit$set(root.dir = here::here())
library(tidyverse)

ipa_symbols <- read_csv("./data-raw/ipa_symbols.csv") %>%
  mutate(
    specs_n = lengths(str_split(ipa_name, " ")),
    spec_1 = word(ipa_name, 1),
    type = case_when(
      spec_1 %in% c(
        "voiceless", "voiced", "simultaneous", "epiglottal", "retroflex") &
        specs_n > 1 &
        str_detect(uni_name, pattern = "modifier", negate = TRUE) &
        str_detect(uni_name, pattern = "combining", negate = TRUE) ~ "consonant",
      spec_1 %in% c("open", "close", "open-mid", "close-mid", "lax", "raised", "lowered", "mid-central") & specs_n > 1 ~ "vowel",
      TRUE ~ "diacritic"
    ),
    height_ipa = case_when(
      type == "vowel" & specs_n == 2 ~ word(ipa_name, 1),
      type == "vowel" & specs_n == 3 ~ word(ipa_name, 1),
      type == "vowel" & specs_n == 4 ~ word(ipa_name, 1, 2)
    ),
    height = case_when(
      height_ipa == "open" ~ "low",
      height_ipa == "raised open" ~ "low_raised",
      height_ipa == "lowered" ~ "low_raised",
      height_ipa == "close-mid" ~ "mid-high",
      height_ipa == "mid-central" ~ "mid-central",
      height_ipa == "open-mid" ~ "mid-low",
      height_ipa == "close" ~ "high",
      height_ipa == "lax close" ~ "high_lowered"
    ),
    backness = case_when(
      ipa_name %in% c("lowered schwa", "mid-central schwa") ~ "central",
      type == "vowel" & specs_n == 2 ~ word(ipa_name, 1),
      type == "vowel" & specs_n == 3 ~ word(ipa_name, 2),
      type == "vowel" & specs_n == 4 ~ word(ipa_name, 3)
    ),
    rounding = case_when(
      type == "vowel" ~ ifelse(word(ipa_name, -1) == "schwa", "unrounded", word(ipa_name, -1))
    ),
    voicing = case_when(
      type == "consonant" ~ ifelse(word(ipa_name, 1) %in% c("voiceless", "voiced"), word(ipa_name, 1), "voiceless")
    ),
    place = case_when(
      type == "consonant" & specs_n == 2 ~ word(ipa_name, 1),
      type == "consonant" & specs_n > 2 ~ ifelse(word(ipa_name, 1) == "simultaneous", "postalveolar-velar", word(ipa_name, 2))
    ),
    manner = case_when(
      type == "consonant" & specs_n == 2 ~ word(ipa_name, 2),
      type == "consonant" & specs_n == 3 ~ word(ipa_name, 3),
      type == "consonant" & specs_n == 4 ~ word(ipa_name, 4)
    ),
    lateral = str_detect(ipa_name, "lateral"),
    sonorant = case_when(
      type == "vowel" ~ TRUE,
      type == "consonant" & manner %in% c("approximant", "flap", "nasal", "tap", "trill") ~ TRUE,
      type == "consonant" & manner %in% c("click", "fricative", "implosive", "plosive") ~ FALSE
    )
  ) %>%
  dplyr::select(-specs_n, -spec_1)

usethis::use_data(ipa_symbols, overwrite = TRUE)

stefanocoretta/phonetisr documentation built on June 2, 2025, 2:06 a.m.