from __future__ import annotations

import re
from typing import Optional

import requests
from bs4 import BeautifulSoup
from ddgs import DDGS


USER_AGENT = "Mozilla/5.0 (compatible; HF-Benchmark-Agent/1.0)"


def web_search_first(query: str, max_results: int = 5) -> list[dict]:
    results: list[dict] = []
    try:
        with DDGS() as ddgs:
            for r in ddgs.text(query, max_results=max_results):
                results.append(r)
    except Exception:
        return []
    return results


def fetch_url_text(url: str, max_chars: int = 12000) -> str:
    try:
        resp = requests.get(url, headers={"User-Agent": USER_AGENT}, timeout=20)
        resp.raise_for_status()
    except Exception:
        return ""

    html = resp.text
    soup = BeautifulSoup(html, "html.parser")

    for tag in soup(["script", "style", "noscript"]):
        tag.decompose()

    text = soup.get_text("\n")
    text = re.sub(r"\n{2,}", "\n", text)
    return text[:max_chars].strip()


def search_and_fetch(query: str, max_results: int = 3, max_chars: int = 12000) -> str:
    results = web_search_first(query, max_results=max_results)
    chunks = []

    for r in results[:max_results]:
        title = r.get("title", "")
        href = r.get("href", "")
        body = r.get("body", "")
        page_text = fetch_url_text(href, max_chars=max_chars // max(1, max_results)) if href else ""
        chunks.append(
            f"[TITLE]\n{title}\n[URL]\n{href}\n[SNIPPET]\n{body}\n[PAGE TEXT]\n{page_text}"
        )

    return "\n\n".join(chunks).strip()