# Copyright (c) 2025 Patrick Motsch
# All rights reserved.
"""
Voice / Language Catalog — Single Source of Truth.

Every voice-related component (TTS connector, AI tools, REST routes, frontend
language pickers) consumes this catalog. Hard-coded language lists or ad-hoc
ISO→BCP-47 maps elsewhere are forbidden — extend the catalog instead.

Schema per entry:
    bcp47           BCP-47 locale code, e.g. "de-DE", "ru-RU"
    iso             ISO-639-1 short code, e.g. "de", "ru"
    label           Native display label ("Deutsch", "Русский")
    flag            Emoji flag (or empty string for region-neutral codes)
    defaultVoice    Curated Google TTS voice name; None means "let Google
                    pick automatically based on bcp47 + ssml_gender".
"""

from __future__ import annotations

from dataclasses import asdict, dataclass
from typing import Dict, List, Optional


@dataclass(frozen=True)
class VoiceLanguage:
    bcp47: str
    iso: str
    label: str
    flag: str
    defaultVoice: Optional[str]


# Order matters for UI: most common first, then alphabetical groups.
VOICE_LANGUAGES: List[VoiceLanguage] = [
    VoiceLanguage("de-DE", "de", "Deutsch",                 "🇩🇪", "de-DE-Wavenet-A"),
    VoiceLanguage("de-CH", "de", "Deutsch (Schweiz)",       "🇨🇭", "de-DE-Wavenet-A"),
    VoiceLanguage("de-AT", "de", "Deutsch (Österreich)",    "🇦🇹", "de-DE-Wavenet-A"),
    VoiceLanguage("en-US", "en", "English (US)",            "🇺🇸", "en-US-Wavenet-C"),
    VoiceLanguage("en-GB", "en", "English (UK)",            "🇬🇧", "en-GB-Wavenet-A"),
    VoiceLanguage("en-AU", "en", "English (Australia)",     "🇦🇺", "en-AU-Wavenet-A"),
    VoiceLanguage("fr-FR", "fr", "Français",                "🇫🇷", "fr-FR-Wavenet-A"),
    VoiceLanguage("fr-CA", "fr", "Français (Canada)",       "🇨🇦", "fr-CA-Wavenet-A"),
    VoiceLanguage("it-IT", "it", "Italiano",                "🇮🇹", "it-IT-Wavenet-A"),
    VoiceLanguage("es-ES", "es", "Español",                 "🇪🇸", "es-ES-Wavenet-B"),
    VoiceLanguage("es-US", "es", "Español (US)",            "🇺🇸", "es-US-Wavenet-A"),
    VoiceLanguage("pt-BR", "pt", "Português (Brasil)",      "🇧🇷", "pt-BR-Wavenet-A"),
    VoiceLanguage("pt-PT", "pt", "Português (Portugal)",    "🇵🇹", "pt-PT-Wavenet-A"),
    VoiceLanguage("nl-NL", "nl", "Nederlands",              "🇳🇱", "nl-NL-Wavenet-A"),
    VoiceLanguage("pl-PL", "pl", "Polski",                  "🇵🇱", "pl-PL-Wavenet-A"),
    VoiceLanguage("ru-RU", "ru", "Русский",                 "🇷🇺", "ru-RU-Wavenet-A"),
    VoiceLanguage("uk-UA", "uk", "Українська",              "🇺🇦", "uk-UA-Wavenet-A"),
    VoiceLanguage("cs-CZ", "cs", "Čeština",                 "🇨🇿", "cs-CZ-Wavenet-A"),
    VoiceLanguage("sk-SK", "sk", "Slovenčina",              "🇸🇰", "sk-SK-Wavenet-A"),
    VoiceLanguage("hu-HU", "hu", "Magyar",                  "🇭🇺", "hu-HU-Wavenet-A"),
    VoiceLanguage("ro-RO", "ro", "Română",                  "🇷🇴", "ro-RO-Wavenet-A"),
    VoiceLanguage("el-GR", "el", "Ελληνικά",                "🇬🇷", "el-GR-Wavenet-A"),
    VoiceLanguage("sv-SE", "sv", "Svenska",                 "🇸🇪", "sv-SE-Wavenet-A"),
    VoiceLanguage("da-DK", "da", "Dansk",                   "🇩🇰", "da-DK-Wavenet-A"),
    VoiceLanguage("nb-NO", "nb", "Norsk",                   "🇳🇴", "nb-NO-Wavenet-A"),
    VoiceLanguage("fi-FI", "fi", "Suomi",                   "🇫🇮", "fi-FI-Wavenet-A"),
    VoiceLanguage("tr-TR", "tr", "Türkçe",                  "🇹🇷", "tr-TR-Wavenet-A"),
    VoiceLanguage("ar-XA", "ar", "العربية",                 "",   "ar-XA-Wavenet-A"),
    VoiceLanguage("hi-IN", "hi", "हिन्दी",                    "🇮🇳", "hi-IN-Wavenet-A"),
    VoiceLanguage("ja-JP", "ja", "日本語",                  "🇯🇵", "ja-JP-Wavenet-A"),
    VoiceLanguage("ko-KR", "ko", "한국어",                  "🇰🇷", "ko-KR-Wavenet-A"),
    VoiceLanguage("zh-CN", "zh", "中文 (简体)",             "🇨🇳", "cmn-CN-Wavenet-A"),
    VoiceLanguage("vi-VN", "vi", "Tiếng Việt",              "🇻🇳", "vi-VN-Wavenet-A"),
    VoiceLanguage("th-TH", "th", "ไทย",                     "🇹🇭", "th-TH-Standard-A"),
    VoiceLanguage("id-ID", "id", "Bahasa Indonesia",        "🇮🇩", "id-ID-Wavenet-A"),
]


# ---------------------------------------------------------------------------
# Lookup indexes (built once at import).
# ---------------------------------------------------------------------------

_BY_BCP47: Dict[str, VoiceLanguage] = {v.bcp47.lower(): v for v in VOICE_LANGUAGES}
_BY_ISO: Dict[str, VoiceLanguage] = {}
for _v in VOICE_LANGUAGES:
    _BY_ISO.setdefault(_v.iso.lower(), _v)


def listVoiceLanguages() -> List[VoiceLanguage]:
    """Return the canonical, ordered list of supported voice languages."""
    return list(VOICE_LANGUAGES)


def getCatalogPayload() -> List[Dict[str, Optional[str]]]:
    """Return the catalog as plain dicts — ready for JSON serialization."""
    return [asdict(v) for v in VOICE_LANGUAGES]


def getByBcp47(code: Optional[str]) -> Optional[VoiceLanguage]:
    if not code:
        return None
    return _BY_BCP47.get(code.strip().lower())


def getByIso(code: Optional[str]) -> Optional[VoiceLanguage]:
    if not code:
        return None
    return _BY_ISO.get(code.strip().lower())


def getDefaultVoice(bcp47: Optional[str]) -> Optional[str]:
    """Return the curated default Google TTS voice for a BCP-47 code, else None.

    None means: caller must omit `name` in VoiceSelectionParams so Google
    auto-selects a voice for the language code.
    """
    entry = getByBcp47(bcp47)
    return entry.defaultVoice if entry else None


def isoToBcp47(iso: Optional[str]) -> Optional[str]:
    """Map an ISO-639-1 short code to the canonical BCP-47 locale.

    Already-qualified BCP-47 inputs are passed through unchanged (canonicalised
    to the catalog form when known). Unknown ISO codes fall back to
    ``<iso>-<ISO>`` (e.g. "fa" → "fa-FA") so callers always get a parseable
    locale, but unknown codes carry no curated voice.
    """
    if not iso:
        return None
    normalized = iso.strip()
    if not normalized:
        return None
    if "-" in normalized:
        canonical = getByBcp47(normalized)
        return canonical.bcp47 if canonical else normalized
    isoLower = normalized.lower()
    entry = _BY_ISO.get(isoLower)
    if entry:
        return entry.bcp47
    return f"{isoLower}-{isoLower.upper()}"