"""fetch_hackernews_search — busca en Hacker News via la API Algolia publica.

Funcion impura: hace peticiones HTTP a hn.algolia.com (sin auth ni anti-bot).
Normaliza cada hit a un shape comun de market intelligence.
"""

import requests

_TIMEOUT = 15


def _parse_hits(hits: list, query: str) -> list[dict]:
    """Normaliza la lista hits de la respuesta de Algolia al shape comun."""
    rows = []
    for hit in hits:
        if not isinstance(hit, dict):
            continue
        object_id = str(hit.get("objectID", ""))
        external_url = hit.get("url")
        url = external_url if external_url else (
            f"https://news.ycombinator.com/item?id={object_id}"
        )
        body = hit.get("story_text") or hit.get("comment_text") or ""
        rows.append({
            "source": "hackernews",
            "platform_id": object_id,
            "title": hit.get("title", "") or "",
            "body": body,
            "url": url,
            "author": hit.get("author", "") or "",
            "channel": "hn",
            "created_utc": float(hit.get("created_at_i") or 0.0),
            "platform_score": int(hit.get("points") or 0),
            "query": query,
        })
    return rows


def fetch_hackernews_search(
    query: str,
    limit: int = 50,
    tags: str = "story",
) -> list[dict]:
    """Busca en Hacker News usando la API Algolia publica (sin autenticacion).

    Args:
        query: Termino de busqueda.
        limit: Maximo de resultados (hitsPerPage de Algolia).
        tags: Filtro de tipo de item: "story" (default), "comment",
              "story,comment", "show_hn", "ask_hn", etc.

    Returns:
        Lista de dicts normalizados (puede ser []). Cada dict tiene las claves:
        source, platform_id, title, body, url, author, channel, created_utc,
        platform_score, query.
    """
    url = "https://hn.algolia.com/api/v1/search"
    params = {
        "query": query,
        "tags": tags,
        "hitsPerPage": limit,
    }

    try:
        resp = requests.get(url, params=params, timeout=_TIMEOUT)
        resp.raise_for_status()
        payload = resp.json()
        hits = payload.get("hits", []) if isinstance(payload, dict) else []
        return _parse_hits(hits, query)
    except Exception:
        return []