"""Heuristiska emocionālā konteksta noteikšana."""

from __future__ import annotations

from dataclasses import dataclass

KEYWORD_MATCH_SCORE = 0.22
TECHNICAL_CONTEXT_SCORE = 0.08
EXCLAMATION_BONUS = 0.08
QUESTION_BONUS = 0.04
MAX_PUNCTUATION_BONUS = 0.2


@dataclass(frozen=True, slots=True)
class EmotionalContext:
    emotion: str
    confidence: float
    response_style: str
    guidance: str
    description: str


_EMOTION_KEYWORDS: dict[str, tuple[str, ...]] = {
    "distressed": (
        "palīdzi",
        "help me",
        "panic",
        "panika",
        "esmu nobijies",
        "bail",
        "anxiety",
        "trauksme",
        "overwhelmed",
        "nespēju",
        "hopeless",
        "izmisums",
    ),
    "frustrated": (
        "nestrādā",
        "neiet",
        "salūza",
        "broken",
        "besī",
        "kaitina",
        "frustrated",
        "stuck",
    ),
    "urgent": ("steidzami", "urgent", "asap", "immediately", "tagad", "right now"),
    "excited": (
        "super",
        "wow",
        "lieliski",
        "amazing",
        "awesome",
        "can't wait",
        "nevaru sagaidīt",
        "let's go",
        "fantastic",
    ),
    "curious": (
        "kā",
        "why",
        "paskaidro",
        "explain",
        "pastāsti",
        "gribu saprast",
        "analyze",
        "izanalizē",
        "salīdzini",
        "compare",
    ),
}

_TECHNICAL_ISSUE_KEYWORDS = ("bug", "error", "kļūda", "issue", "problem", "problēma")

_STYLE_GUIDANCE: dict[str, tuple[str, str]] = {
    "neutral": (
        "clear_grounded",
        "Saglabā mierīgu, profesionālu un skaidru toni.",
    ),
    "curious": (
        "exploratory_explanatory",
        "Esi izzinošs, strukturē atbildi un paskaidro domāšanas gaitu skaidri.",
    ),
    "excited": (
        "energized_positive",
        "Atbildi ar pozitīvu enerģiju, bet saglabā saturu konkrētu un praktisku.",
    ),
    "frustrated": (
        "calm_reassuring_step_by_step",
        "Atzīsti grūtību, nomierini toni un ved lietotāju pa skaidriem soļiem.",
    ),
    "distressed": (
        "empathetic_supportive",
        "Atbildi empātiski, mierinoši un ar drošu, vienkāršu nākamo soli.",
    ),
    "urgent": (
        "decisive_action_focused",
        "Īsi prioritizē, uzsver svarīgāko un dod tūlītēji izpildāmus nākamos soļus.",
    ),
}

_STYLE_DESCRIPTIONS: dict[str, str] = {
    "neutral": "mierīgs un līdzsvarots",
    "curious": "izzinošs un ieinteresēts",
    "excited": "pozitīvi uzlādēts",
    "frustrated": "saspringts un neapmierināts",
    "distressed": "noraizējies un pārņemts",
    "urgent": "steidzams un fokusēts",
}


def analyze_emotional_context(text: str) -> EmotionalContext:
    """Atgriež labāko emocionālo interpretāciju lietotāja tekstam."""
    normalized = " ".join(text.lower().split())
    if not normalized:
        return _context_for("neutral", 0.35)

    scores: dict[str, float] = {emotion: 0.0 for emotion in _STYLE_GUIDANCE}
    punctuation_bonus = min(
        text.count("!") * EXCLAMATION_BONUS + text.count("?") * QUESTION_BONUS,
        MAX_PUNCTUATION_BONUS,
    )

    for emotion, keywords in _EMOTION_KEYWORDS.items():
        for keyword in keywords:
            if keyword in normalized:
                scores[emotion] += KEYWORD_MATCH_SCORE

    if any(keyword in normalized for keyword in _TECHNICAL_ISSUE_KEYWORDS):
        scores["frustrated"] += TECHNICAL_CONTEXT_SCORE

    if "?" in text and scores["curious"] == 0:
        scores["curious"] += 0.16
    if text.count("!") >= 2:
        scores["excited"] += 0.14
    if any(token.isupper() and len(token) > 2 for token in text.split()):
        scores["urgent"] += 0.1

    strongest_emotion = max(
        (emotion for emotion in scores if emotion != "neutral"),
        key=lambda emotion: scores[emotion],
        default="neutral",
    )
    strongest_score = scores[strongest_emotion]

    if strongest_score <= 0:
        return _context_for("neutral", min(0.45 + punctuation_bonus, 0.55))

    confidence = min(0.5 + strongest_score + punctuation_bonus, 0.93)
    return _context_for(strongest_emotion, confidence)


def _context_for(emotion: str, confidence: float) -> EmotionalContext:
    response_style, guidance = _STYLE_GUIDANCE[emotion]
    return EmotionalContext(
        emotion=emotion,
        confidence=round(confidence, 2),
        response_style=response_style,
        guidance=guidance,
        description=_STYLE_DESCRIPTIONS[emotion],
    )