mcp/src/aileen3_mcp/cli_slides.py

"""Manual slide extraction helper.

Run slide extraction against a local video using the same pipeline the MCP
server uses. Useful for debugging model responses when no slides are returned.
"""

from __future__ import annotations

import argparse
import sys
from pathlib import Path

from .logging_utils import configure_logging
from .media_tools import _build_reference, _extract_slides_flow, _probe_duration, _slides_json_path


def main(argv: list[str] | None = None) -> int:
    configure_logging()
    parser = argparse.ArgumentParser(description="Extract slides from a video using the Gemini pipeline.")
    parser.add_argument("video", type=Path, help="Path to the video file (mp4 recommended).")
    parser.add_argument(
        "--reference",
        help="Reference id to use for cache/output. Defaults to one derived from the filename.",
    )
    parser.add_argument(
        "--duration",
        type=float,
        help="Override duration in seconds (optional). If omitted, ffprobe is used.",
    )

    args = parser.parse_args(argv)

    video_path: Path = args.video.expanduser().resolve()
    if not video_path.exists():
        parser.error(f"Video not found: {video_path}")

    reference = args.reference or _build_reference(None, str(video_path))
    duration = args.duration or _probe_duration(video_path)

    metadata = {
        "reference": reference,
        "download_path": str(video_path),
        "duration": duration,
        "source": str(video_path),
    }

    try:
        result = _extract_slides_flow(metadata)
    except Exception as exc:  # pragma: no cover - CLI convenience
        print(f"[error] slide extraction failed: {exc}", file=sys.stderr)
        return 1

    slides = result.get("slides", [])
    print(f"Extracted {len(slides)} slides for reference '{reference}'.")
    if slides:
        print("First few slides:")
        for slide in slides[:5]:
            start = slide.get("from")
            end = slide.get("to")
            label = slide.get("label") or ""
            print(f"  index={slide.get('index')} from={start:.2f}s to={end:.2f}s label={label}")
    slides_json = _slides_json_path(reference)
    print(f"Slides JSON saved to {slides_json}")
    return 0


if __name__ == "__main__":  # pragma: no cover - CLI execution
    raise SystemExit(main())