services/speaker_profiler.py

"""
Stage 5 — Speaker Profiler (Simplified — Male Voice Forced)
Since we force male voice, this just groups speakers.
"""
import logging
from pathlib import Path
from typing import Dict, List
from collections import defaultdict

logger = logging.getLogger(__name__)


def profile_speakers(
    segments: List[Dict],
    audio_path: Path,
    output_dir: Path
) -> Dict[str, Dict]:
    """Profile speakers — all forced to male."""
    speaker_segments = defaultdict(list)
    for seg in segments:
        speaker_segments[seg.get("speaker", "SPEAKER_00")].append(seg)

    logger.info(f"Found {len(speaker_segments)} speakers — ALL forced to MALE voice")

    profiles = {}
    for speaker_id, segs in speaker_segments.items():
        total_time = sum(s["end"] - s["start"] for s in segs)
        profiles[speaker_id] = {
            "gender": "male",  # FORCED
            "reference_audio": None,
            "total_speaking_time": round(total_time, 2),
            "avg_pitch": 0,
            "segment_count": len(segs),
        }
        logger.info(f"  {speaker_id}: MALE (forced), {len(segs)} segments, {total_time:.1f}s")

    return profiles