A simple regex exercise in R whereby I build pinyin syllables from base consonant and vowel pairs and then pare down to valid syllables.

pinyin.R

# The purpose of this script is to generate a random Mandarin syllable pairing for use in tone
# practice drills. The random syllable pair is used to populate a grid that does not have tones.
# The tone_table has the tones of the two syllables. The student then tries to say the random
# pairing with the tones they see. The partner then tries to guess which tones were meant.
# 
# The script generates the syllables from scratch, starting with all possible vowel and consonant
# pairings. Then it uses a very complicated string of regex to exclude out invalid syllables.
# My goal with this script was to make the tone tables, but it was also meant to learn some
# useful regex, which is why I indulged in it rather than just typing out all possible syllables
# (which would have been much faster at my typing speed than finding all the exclusions I needed).


# Pinyin initials include many of the same consonants as in the English alphabet except for V 
# and with the addition of sh, ch and zh; we build a vector of pinyin initials here
initials <- c(letters[1:26][-c(1,5,9,15,21,22)], "sh", "zh", "ch", "")

# Pinyin final sounds include the standard English vowels, dipthongs, and several others
finals <- c(letters[c(1,5,9,15,21)], "ai", "ei", "ou", "ao", "an", "en", "eng", "ang", "ong", "er", "ia", "iao", "ie", "iu", "ian", "in", "iang", "ing", "iong", "ua", "uo", "uai", "ui", "uan", "un", "uang", "uen", "ueng", "ü", "üe", "üan", "ün")

# Combine intials and finals
  # Bring together in a list
initials_finals <- expand.grid(initials, finals)
  # Flatten the list
concate_initials_finals <- do.call(paste, initials_finals)
  # Remove spaces and cheat by appending known syllables that don't naturally fall into the initials/finals combinations
all_possible_syllables <- c(gsub(" ", "", concate_initials_finals), "xue", "jue", "que", "yue")

# A ridiculously long, probably ill-advised regex to exclude impossible syllables
# The entire point of this exercise was to learn regex, so we'll allow it this time
exclusions <- !grepl("\\<r[ao]\\>|ruang|ruen|r[u]ai|^[uiü]|^ong|yai|yen.|yi[ae]|wao|wi|[xjqy]u[ae]ng|\\<[xjqy]ua[i]\\>|\\<[xjqy]u[aoi]\\>|[xjq][aoe]|[szcrw]i.|[scry]ei|chei|[szc]hi.|\\<shong\\>|\\<[szc]ho\\>|\\<[tly]en\\>|\\<[dtnlszcy]uang\\>|[dtnlgkhszcqjxy]uen|\\<[dtnlsczr]ua\\>|[dtnlscz]ua[i]|\\<[dt]in\\>|\\<[dtngkhscz]o\\>|^hi|[fgk]i|fa[io]|\\<[bpfxjqw]e\\>|.er|[bxjqw]o.|[pmf]ong|\\<[bpmftngkhszcrwy]ia\\>|[bpftgkhszcrwy]iu|[bpmfdtgkhszcrwy]i[ao]ng|[nls]iong|[bpmfw]u.+|[nl]ui|[bpmfdtgkhszcrxhqwyjq]ü|[nl]üan|[nl]ün", all_possible_syllables)

# Apply exclusions
allowed_syllables <- all_possible_syllables[exclusions]

# Pull together two random syllables
random_syllables <- paste(sample(allowed_syllables, 2), collapse = "")

# Create pinyin table
pinyin_table <- matrix(rep.int(random_syllables,9), ncol = 3, byrow = TRUE)

# Create tones key
tone_table <- matrix(replicate(9, paste(sample.int(4, size = 2, replace = FALSE), collapse = "")), ncol = 3, byrow = TRUE)