twitch-highlight-detector/two_game_extractor.py

#!/usr/bin/env python3
"""
TWO-GAME HIGHLIGHT EXTRACTOR
Extrae múltiples highlights de los 2 juegos: Diana y Mundo
"""

import json
import re


def extract_game_highlights():
    """Extrae highlights de Diana y Mundo por separado."""

    print("=" * 60)
    print("TWO-GAME HIGHLIGHT EXTRACTOR")
    print("=" * 60)

    with open("transcripcion_rage.json", "r") as f:
        trans = json.load(f)

    # Identificar segmentos por campeón
    diana_segments = []
    mundo_segments = []

    for seg in trans["segments"]:
        text = seg["text"].lower()

        if "diana" in text:
            diana_segments.append(seg)
        elif "mundo" in text or "warwick" in text:
            mundo_segments.append(seg)

    print(f"Segmentos mencionando Diana: {len(diana_segments)}")
    print(f"Segmentos mencionando Mundo/Warwick: {len(mundo_segments)}")

    # Encontrar rangos de tiempo
    if diana_segments:
        diana_start = min(s["start"] for s in diana_segments)
        diana_end = max(s["end"] for s in diana_segments)
        print(f"\nJuego Diana: {diana_start / 60:.0f}m - {diana_end / 60:.0f}m")
    else:
        diana_start, diana_end = 0, 0

    if mundo_segments:
        mundo_start = min(s["start"] for s in mundo_segments)
        mundo_end = max(s["end"] for s in mundo_segments)
        print(f"Juego Mundo: {mundo_start / 60:.0f}m - {mundo_end / 60:.0f}m")
    else:
        mundo_start, mundo_end = 0, 0

    # Buscar momentos épicos en cada juego
    def find_moments_in_range(segments, game_name, start_time, end_time, min_score=6):
        """Busca momentos épicos en un rango específico."""

        moments = []

        rage_patterns = [
            (r"\bputa\w*", 10, "EXTREME"),
            (r"\bme mataron\b", 12, "DEATH"),
            (r"\bme mori\b", 12, "DEATH"),
            (r"\bmierda\b", 8, "RAGE"),
            (r"\bjoder\b", 8, "RAGE"),
            (r"\bretrasad\w*", 9, "INSULT"),
            (r"\bimbecil\b", 9, "INSULT"),
            (r"\bla cague\b", 8, "FAIL"),
            (r"\bnooo+\b", 6, "FRUSTRATION"),
        ]

        for seg in segments:
            if seg["start"] < start_time or seg["end"] > end_time:
                continue

            text = seg["text"].lower()
            score = 0
            reasons = []

            for pattern, points, reason in rage_patterns:
                if re.search(pattern, text, re.IGNORECASE):
                    score += points
                    if reason not in reasons:
                        reasons.append(reason)

            if score >= min_score:
                moments.append(
                    {
                        "start": seg["start"],
                        "end": seg["end"],
                        "score": score,
                        "text": seg["text"][:70],
                        "reasons": reasons,
                        "game": game_name,
                    }
                )

        return moments

    # Buscar en Diana
    print(f"\n=== JUEGO DIANA ===")
    diana_moments = find_moments_in_range(
        trans["segments"],
        "Diana",
        max(455, diana_start - 300),  # 5 min antes de primera mención
        diana_end + 300,  # 5 min después
        min_score=5,
    )
    print(f"Momentos encontrados: {len(diana_moments)}")

    # Buscar en Mundo
    print(f"\n=== JUEGO MUNDO ===")
    mundo_moments = find_moments_in_range(
        trans["segments"],
        "Mundo",
        max(455, mundo_start - 300),
        mundo_end + 300,
        min_score=5,
    )
    print(f"Momentos encontrados: {len(mundo_moments)}")

    # Ordenar por score
    diana_moments.sort(key=lambda x: -x["score"])
    mundo_moments.sort(key=lambda x: -x["score"])

    # Tomar top 6 de cada juego
    best_diana = diana_moments[:6]
    best_mundo = mundo_moments[:6]

    print(f"\nMejores momentos Diana: {len(best_diana)}")
    for i, m in enumerate(best_diana, 1):
        mins = int(m["start"]) // 60
        secs = int(m["start"]) % 60
        print(
            f"  {i}. {mins:02d}:{secs:02d} [Score: {m['score']}] {'/'.join(m['reasons'])}"
        )

    print(f"\nMejores momentos Mundo: {len(best_mundo)}")
    for i, m in enumerate(best_mundo, 1):
        mins = int(m["start"]) // 60
        secs = int(m["start"]) % 60
        print(
            f"  {i}. {mins:02d}:{secs:02d} [Score: {m['score']}] {'/'.join(m['reasons'])}"
        )

    # Combinar y crear clips
    all_moments = best_diana + best_mundo

    # Crear clips extendidos
    clips = []
    for m in all_moments:
        start = max(455, int(m["start"]) - 10)
        end = min(8237, int(m["end"]) + 15)
        if end - start >= 15:
            clips.append(
                {
                    "start": start,
                    "end": end,
                    "score": m["score"],
                    "reasons": m["reasons"],
                    "game": m["game"],
                }
            )

    # Eliminar solapamientos
    clips.sort(key=lambda x: x["start"])
    filtered = []
    for clip in clips:
        if not filtered:
            filtered.append(clip)
        else:
            last = filtered[-1]
            if clip["start"] <= last["end"] + 5:
                # Fusionar
                last["end"] = max(last["end"], clip["end"])
                last["score"] = max(last["score"], clip["score"])
                last["reasons"] = list(set(last["reasons"] + clip["reasons"]))
                if clip["game"] not in last["game"]:
                    last["game"] += "/" + clip["game"]
            else:
                filtered.append(clip)

    # Ordenar por tiempo
    filtered.sort(key=lambda x: x["start"])

    print(f"\n{'=' * 60}")
    print(f"TOTAL: {len(filtered)} clips")
    total_dur = sum(c["end"] - c["start"] for c in filtered)
    print(f"Duración: {total_dur}s ({total_dur // 60}m {total_dur % 60}s)")
    print(f"{'=' * 60}")

    print(f"\nTimeline final:")
    for i, c in enumerate(filtered, 1):
        mins, secs = divmod(c["start"], 60)
        dur = c["end"] - c["start"]
        print(
            f"{i:2d}. {mins:02d}:{secs:02d} - {dur}s [{c['game']}] {'/'.join(c['reasons'])}"
        )

    return filtered


if __name__ == "__main__":
    clips = extract_game_highlights()

    # Guardar
    highlights = [[c["start"], c["end"]] for c in clips]
    with open("highlights_two_games.json", "w") as f:
        json.dump(highlights, f)

    print(f"\nGuardado en highlights_two_games.json")