services/whisper/utils/clipper.py

from moviepy.editor import VideoFileClip
import os
import logging

logger = logging.getLogger(__name__)


# =====================================================
# SEGMENT NORMALIZER (CRITICAL V8 FIX)
# =====================================================

def normalize_segments(segments):
    """
    Accepts:
      - dict segments: {"start": x, "end": y}
      - list segments: [{"start":x,"end":y}, ...]
      - tuple/list segments: [(start,end), ...]

    Returns:
      - clean list of dicts
    """

    if not segments:
        return []

    normalized = []

    # Case 1: single dict
    if isinstance(segments, dict):
        segments = [segments]

    for s in segments:

        # dict format (preferred)
        if isinstance(s, dict):
            if "start" in s and "end" in s:
                normalized.append({
                    "start": float(s["start"]),
                    "end": float(s["end"])
                })
            continue

        # list/tuple format
        if isinstance(s, (list, tuple)) and len(s) >= 2:
            try:
                normalized.append({
                    "start": float(s[0]),
                    "end": float(s[1])
                })
            except Exception:
                continue

    return normalized


# =====================================================
# CORE CLIP GENERATOR (SAFE VERSION)
# =====================================================

def create_clip(video_path, start, end, index):
    """
    Creates a single clip safely with validation
    """

    try:
        start = float(start)
        end = float(end)

        if end <= start:
            logger.warning(f"Invalid segment skipped: {start}-{end}")
            return None

        clip = VideoFileClip(video_path).subclip(start, end)

        output = video_path.replace(
            ".mp4",
            f"_clip_{index}.mp4"
        )

        clip.write_videofile(
            output,
            codec="libx264",
            audio_codec="aac",
            preset="ultrafast",
            threads=2,
            logger=None  # prevents HF log spam
        )

        return output

    except Exception as e:
        logger.error(f"Clip creation failed: {str(e)}")
        return None


# =====================================================
# BATCH CLIP ENGINE (V8 AUTOCUT CORE FIX)
# =====================================================

def create_clips(video_path, segments):
    """
    Main entry used by V8 Highlights / AutoClip engine
    """

    segments = normalize_segments(segments)

    if not segments:
        logger.warning("No valid segments found")
        return []

    outputs = []

    for i, seg in enumerate(segments):

        try:
            out = create_clip(
                video_path,
                seg["start"],
                seg["end"],
                i
            )

            if out:
                outputs.append(out)

        except Exception as e:
            logger.error(f"Segment {i} failed: {str(e)}")
            continue

    return outputs