src/tools/search.py

from __future__ import annotations

import json
import os
from pathlib import Path

from src.schemas import SourceCandidate


def fixture_sources(fixtures_path: str | Path = "data/fixtures/demo_sources.json") -> list[SourceCandidate]:
    data = json.loads(Path(fixtures_path).read_text(encoding="utf-8"))
    return [SourceCandidate.from_dict(item) for item in data]


def search_web(query: str, limit: int = 5) -> list[SourceCandidate]:
    mock_mode = os.getenv("MOCK_MODE", "false").lower() == "true"
    if mock_mode:
        return fixture_sources()[:limit]
    if os.getenv("DIRECT_SEARCH_ENABLED", "true").lower() != "true":
        return []
    try:
        from duckduckgo_search import DDGS

        rows = DDGS().text(query, max_results=limit)
        sources = [
            SourceCandidate(
                title=row.get("title") or "Untitled",
                url=row.get("href") or "",
                snippet=row.get("body") or "",
                source_type="web",
                fetch_status="search_only",
            )
            for row in rows
        ]
        return sources
    except Exception:
        return []