app/voice.py

"""Voice = Simlish babble + emoticon + browser-TTS params. No model, no GPU,
no server cost — the frontend speaks `simlish` via the Web Speech API and shows
`emoticon` in the bubble. Rage register uses a harsher syllable pool + lower,
faster voice."""
from __future__ import annotations
import random
from app.appearance import _face

_WORD = {
    "content": "cheerful", "excited": "excited", "sad": "sad",
    "angry": "angry", "neutral": "calm",
}
_EMOTICON = {
    "content": "(◕‿◕)", "excited": "(★▽★)",
    "sad": "(╥﹏╥)", "angry": "(╬ಠ益ಠ)",
    "neutral": "(・_・)",
}
_SOFT = ["ba", "bo", "doo", "la", "lee", "mee", "na", "no", "wa", "yu", "pi", "to"]
_RAGE = ["grr", "kah", "ztt", "brk", "gnak", "rrg", "tzk", "drr", "skuh", "vrr"]

# region -> (syllable pool, word count, exclaim, pitch, rate)
_VOICE = {
    "content": (_SOFT, 3, False, 1.4, 1.0),
    "excited": (_SOFT, 4, True, 1.7, 1.25),
    "sad": (_SOFT, 2, False, 1.1, 0.8),
    "angry": (_RAGE, 4, True, 0.7, 1.35),
    "neutral": (_SOFT, 3, False, 1.3, 1.0),
}


def _region(mood: list[int]) -> str:
    return _face(mood[0], mood[1])


def mood_word(mood: list[int]) -> str:
    return _WORD[_region(mood)]


def emoticon(mood: list[int]) -> str:
    return _EMOTICON[_region(mood)]


def simlish(mood: list[int], *, rng: random.Random | None = None) -> str:
    pool, nwords, exclaim, _pitch, _rate = _VOICE[_region(mood)]
    rng = rng or random.Random()
    words = ["".join(rng.choice(pool) for _ in range(rng.randint(1, 3)))
             for _ in range(nwords)]
    return " ".join(words) + ("!" if exclaim else "")


def voice_params(mood: list[int]) -> dict:
    _pool, _n, _ex, pitch, rate = _VOICE[_region(mood)]
    return {"pitch": pitch, "rate": rate}