#Base tag, only one that will be hardcoded
root:
  - val: "{Word}"
    ipa: ""
    freq: 1

digraphProtection:
  insertChar: "'"
  builtins:
    geminates: True
  sets:
    - [[k], [c]]
    - [[], [a, e, i, í, o, ó, u, ū, y]]
    - [[a, e, i, o, u], [r]]
    - [[r], []]
    - [[s], [p, k, t, n, m, l]]
    - [[sp, sk, st, sn, sm, sl], []]
    - [[t, s, c], [h]]

# Recommended to break syllables with .
# Note: ipa is inserted for every literal in the val. You are free to use this
  # to your advantage. If this is undesirable, then decompose your references
  # such that only one literal is present in any node.

# "freq" values do not need any form of normalization.
  # Only their relative values within a particular array matters.
  # This reference file generally bases them on 1 and prefers integers in most
    # cases for neatness, but could just as easily have phrased them all as
    # percentages.
Word:
  - val: "{WeakSyllable}.{StrSyllable}"
    ipa: "."
    freq: 1
  - val: "{StrSyllable}"
    ipa: ""
    freq: 1
  - val: "{StrSyllable}.{StrSyllable}"
    ipa: "."
    freq: .7
    # The . is inserted for both syllables
  - val: "{StrSyllable}.{WeakSyllable}.{StrSyllable}"
    ipa: "."
    freq: .5
  - val: "{WeakSyllable}.{Word2}"
    ipa: "."
    freq: 0.7
  - val: "{StrSyllable}.{Word2}"
    ipa: "."
    freq: 1.5

# Tail recursive, shouldn't create words that are too long
# after the first few, length-chance is negative-exponential
Word2:
  - val: "{WeakSyllable}.{StrSyllable}"
    ipa: "."
    freq: 2
  - val: "{StrSyllable}"
    ipa: ""
    freq: 1.5
  - val: "{WeakSyllable}.{Word2}"
    ipa: "."
    freq: 0.7
  - val: "{StrSyllable}.{Word2}"
    ipa: "."
    freq: 0.8

# Not used (yet)
Iamb:
  - val: "{WeakSyllable}.{StrSyllable}"
    ipa: "."
    freq: 1
  - val: "{StrSyllable}.{WeakSyllable}"
    ipa: "."
    freq: 1

StrSyllable:
  - val: "{Onset}{FullVowel}"
    ipa: ""
    freq: 4
  - val: "{FullVowel}"
    ipa: ""
    freq: 1.5
  - val: "{FullVowel}{AnyCoda}"
    ipa: ""
    freq: 0.5

StrRhyme:
  - val: "{FullVowel}{AnyCoda}"
    ipa: ""
    freq: 0.5
  - val: "{FullVowelNC}"
    ipa: ""
    freq: 2.0

WeakSyllable:
  - val: "{Onset}{WeakVowel}{AnyCoda}"
    ipa: ""
    freq: 6
  - val: "{WeakVowel}{AnyCoda}"
    ipa: ""
    freq: 2
  - val: "{Onset}{FullVowel}{AnyCoda}"
    ipa: ""
    freq: 1.5

Onset:
  - val: ""
    ipa: ""
    freq: 0
  # u: 1 for unfinalized phones
  - val: "p"
    ipa: "p"
    freq: 1.5
    u: 1
  - val: "b"
    ipa: "b"
    freq: 0.75
    u: 1
  # "c" and "k" together = 2.0
  - val: "c"
    ipa: "k"
    freq: 1.0
  - val: "k"
    ipa: "k"
    freq: 1.0
  - val: "g"
    ipa: "g"
    freq: 0.75
    u: 1
  - val: "t"
    ipa: "t"
    freq: 1.5
  - val: "d"
    ipa: "d"
    freq: 1
  - val: "s"
    ipa: "s"
    freq: 1.5
  - val: "z"
    ipa: "z"
    freq: 0.75
    u: 1
  - val: "s{SOnset}"
    ipa: "s"
    freq: 0.75
  - val: "f"
    #["f", "ɸ"]
    ipa: "ɸ"
    freq: 1.6
  - val: "v"
    #["v", "β"]
    ipa: "β"
    freq: 0.8
  - val: "č"
    ipa: "tʃ"
    freq: 0.7
    u: 1
  - val: "j"
    ipa: "dʒ"
    freq: 0.7
  - val: "n"
    ipa: "n"
    freq: 1.225
  - val: "m"
    ipa: "m"
    freq: 1.225
  - val: "l"
    ipa: "l"
    freq: 1.4
  - val: "h"
    ipa: "h"
    freq: 1.2
  - val: "r"
    ipa: "r"
    freq: 1.4

SOnset:
  - val: "p"
    ipa: "p"
    freq: 1.5
  - val: "k"
    ipa: "k"
    freq: 1.5
  - val: "t"
    ipa: "t"
    freq: 1.5
  - val: "n"
    ipa: "n"
    freq: 1.125
  - val: "m"
    ipa: "m"
    freq: 1.125
  - val: "l"
    ipa: "l"
    freq: 0.25

WeakVowel:
  - val: "u"
    ipa: "ʌ"
    freq: 1.5
  - val: "a"
    ipa: "ə"
    freq: 0.5
  - val: "i"
    ipa: "ɪ"
    freq: 1
  - val: "ur"
    ipa: "ər"
    freq: .5

FullVowel:
  - val: "a"
    ipa: "a"
    freq: 2
  - val: "ó"
    ipa: "o"
    freq: 1.5
  - val: "ū"
    ipa: "ʊ"
    freq: 1.5
  - val: "í"
    ipa: "i"
    freq: 0.75
  - val: "ao"
    ipa: "aʊ"
    freq: 1
  - val: "ai"
    ipa: "aɪ"
    freq: 1.5
  - val: "ay"
    ipa: "e"
    freq: 1
  - val: "{RVowel}"
    ipa: ""
    freq: 3

# Certain vowels are more common in open syllables than closed
FullVowelNC:
  - val: "{FullVowel}"
    ipa: ""
    freq: 5
  - val: "ao"
    ipa: "aʊ"
    freq: 0.66
  - val: "ó"
    ipa: "o"
    freq: 1
  - val: "ai"
    ipa: "aɪ"
    freq: 1

RVowel:
  - val: "ar"
    ipa: "ɑr"
    freq: 1.5
  - val: "or"
    ipa: "ɔr"
    freq: 1.5
  - val: "ur"
    ipa: "ər"
    freq: 2

AnyCoda:
  # "" is allowed, but treated separately
  - val: "p"
    ipa: "p"
    freq: 1
    u: 1
  - val: "b"
    ipa: "b"
    freq: 1
    u: 1
  - val: "k"
    ipa: "k"
    freq: 1.3
  - val: "g"
    ipa: "g"
    freq: 0.7
    u: 1
  - val: "t"
    ipa: "t"
    freq: 1.3
  - val: "d"
    ipa: "d"
    freq: 0.7
    u: 1
  - val: "s"
    ipa: "s"
    freq: 0.2
  - val: "z"
    ipa: "z"
    freq: 0.15
    u: 1
  - val: "s{SCoda}"
    ipa: "s"
    freq: 0.2
  - val: "f"
    #["f", "ɸ"]
    ipa: "ɸ"
    freq: 1.3
  - val: "v"
    #["v", "β"]
    ipa: "β"
    freq: 0.7
    u: 1
  - val: "n"
    ipa: "n"
    freq: 0.8
    u: 1
  - val: "m"
    ipa: "m"
    freq: 0.8
    u: 1
  - val: "l"
    ipa: "l"
    freq: 0.95
  # "r" is part of vowel, not coda

SCoda:
  - val: "p"
    ipa: "p"
    freq: 1
    u: 1
  - val: "k"
    ipa: "k"
    freq: 1
  - val: "t"
    ipa: "t"
    freq: 1

# This data file does not absolutely define the structure of an Old Tongue
  # syllable, because that would be much more difficult. Most of the output of
  # this program will be garbage.