import hashlib
import json
import os
import subprocess
import tempfile
from .. import repo
from ..model import whisper_transcribe

NAME = "ingest.video"


def _yt_dlp_info(url):
    """Returns dict of metadata, or None if yt-dlp could not extract."""
    try:
        out = subprocess.check_output(
            ["yt-dlp", "-J", "--no-warnings", "--no-playlist", url],
            timeout=60
        )
        return json.loads(out)
    except subprocess.CalledProcessError:
        return None


def _yt_dlp_audio(url):
    """Downloads bestaudio to a temp .opus and returns the path."""
    tmp_dir = tempfile.mkdtemp(prefix="void-yt-")
    out_template = os.path.join(tmp_dir, "audio.%(ext)s")
    subprocess.run(
        ["yt-dlp", "-x", "--audio-format", "opus", "-o", out_template,
         "--no-warnings", "--no-playlist", url],
        check=True, timeout=600
    )
    for f in os.listdir(tmp_dir):
        if f.startswith("audio."):
            return os.path.join(tmp_dir, f)
    raise RuntimeError("yt-dlp produced no audio file")


def _idem(space_id, url):
    return hashlib.sha256((space_id + "\x00" + url).encode()).hexdigest()


def _kind(url):
    return "youtube" if ("youtube.com" in url or "youtu.be" in url) else "video"


def handle(job_data: dict) -> dict:
    space_id = job_data["space_id"]
    url = job_data["url"]
    idem = _idem(space_id, url)

    info = _yt_dlp_info(url)
    if info is None:
        return {"skipped": "yt-dlp"}

    audio_path = _yt_dlp_audio(url)
    try:
        transcript = whisper_transcribe(audio_path)
    finally:
        try:
            os.unlink(audio_path)
        except OSError:
            pass

    ref_id = repo.create_ref({
        "space_id": space_id,
        "kind": "video",
        "source_url": url,
        "title": info.get("title") or url,
        "summary": (info.get("description") or "")[:5000],
        "body_text": transcript[:200_000],
        "source_kind": _kind(url),
        "external_id": idem,
        "metadata": {
            "duration_s": info.get("duration"),
            "uploader": info.get("uploader"),
            "thumbnail": info.get("thumbnail"),
            "extract": {"method": "whisper", "chars": len(transcript)},
        }
    })
    return {"ref_id": ref_id, "chars": len(transcript)}