utils/youtube_transcript.py

from __future__ import annotations

import requests
from bs4 import BeautifulSoup
from langchain_core.tools import tool
from youtube_transcript_api import YouTubeTranscriptApi
from youtube_transcript_api._errors import (
    NoTranscriptFound,
    TranscriptsDisabled,
    VideoUnavailable,
)

# ---- Config -----------------------------------------------------------------

DEFAULT_TIMEOUT = 30
DEFAULT_HEADERS = {
    # Helps avoid consent/anti-bot interstitials on some sites
    "User-Agent": (
        "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 "
        "(KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36"
    )
}


# ---- Small helpers ----------------------------------------------------------


def _fetch_html(url: str, *, timeout: int = DEFAULT_TIMEOUT) -> str:
    """Download raw HTML for a URL or raise on HTTP errors."""
    resp = requests.get(url, headers=DEFAULT_HEADERS, timeout=timeout)
    resp.raise_for_status()
    return resp.text


def parse_title_description(raw: str) -> tuple[str, str]:
    title, desc = raw.split("Description:", 1)
    return title.replace("Title:", "").strip(), desc.strip()


# ---- Pure functions (safe to call directly in Python) -----------------------


def get_youtube_transcript(video_id: str) -> str:
    """
    Return YouTube transcript text for a given video ID.
    One line per chunk; raises a clear error if transcript is unavailable.

    Example of video_id:
    For youtube video: https://www.youtube.com/watch?v=1htKBjuUWec
    The video id is: dQw4w9WgXcQ
    """
    try:
        # Initialize the YouTubeTranscriptApi
        ytt_api = YouTubeTranscriptApi()
        fetched_transcript = ytt_api.fetch(video_id)
        raw_data = fetched_transcript.to_raw_data()
        # raw data is in the form of [{        'text': 'Hey there',        'start': 0.0,        'duration': 1.54    },    {        'text': 'how are you',,        'start': 1.54,        'duration': 4.16    },   ... ] we will return ony the text element as lines
        transcript = "\n".join([item["text"] for item in raw_data])
        return transcript
    except (TranscriptsDisabled, NoTranscriptFound, VideoUnavailable) as e:
        raise RuntimeError(f"Transcript unavailable: {e}") from e


def get_youtube_title_description(video_url: str) -> str:
    """
    get_youtube title and description from youtube url (ex: https://www.youtube.com/watch?v=1htKBjuUWec)

    Extract YouTube title + description from Open Graph meta tags.
    Falls back to standard <meta name="title"/"description"> if needed.
    """
    html = _fetch_html(video_url)
    soup = BeautifulSoup(html, "html.parser")

    title_tag = soup.find("meta", property="og:title") or soup.find(
        "meta", attrs={"name": "title"}
    )
    desc_tag = soup.find("meta", property="og:description") or soup.find(
        "meta", attrs={"name": "description"}
    )

    title = (title_tag.get("content") if title_tag else None) or "No title found"
    description = (
        desc_tag.get("content") if desc_tag else None
    ) or "No description found"

    return f"Title: {title}\nDescription: {description}"


# ---- LangChain tool wrappers (for agents; call with .invoke) ----------------


@tool("get_youtube_transcript")
def get_youtube_transcript_tool(video_id: str) -> str:
    """Tool: return YouTube transcript text for a video ID."""
    return get_youtube_transcript(video_id)


@tool("get_youtube_title_description")
def get_youtube_title_description_tool(video_url: str) -> str:
    """Tool: return YouTube title + description for a video URL."""
    return get_youtube_title_description(video_url)


# ---- Minimal demo -----------------------------------------------------------

if __name__ == "__main__":
    video_id = "1htKBjuUWec"
    url = f"https://www.youtube.com/watch?v={video_id}"

    print(get_youtube_title_description(url))
    try:
        print("\n--- Transcript (first 500 chars) ---")
        tx = get_youtube_transcript(video_id)
        print(tx[:500] + ("..." if len(tx) > 500 else ""))
    except Exception as e:
        print(f"Transcript error: {e}")