# -*- coding: utf-8 -*-

from backend.enrich import news_enricher


def test_classify_news_row_falls_back_to_local_rules(monkeypatch):
    monkeypatch.setattr(news_enricher, "analyze_news_row_with_llm", lambda row: None)
    result = news_enricher.classify_news_row(
        {
            "title": "Apple shares drop after weak guidance",
            "summary": "Investors reacted negatively to softer-than-expected outlook.",
        }
    )
    assert result["analysis_source"] == "local"
    assert result["sentiment"] == "negative"
    assert result["summary"]


def test_classify_news_row_prefers_llm_when_available(monkeypatch):
    monkeypatch.setattr(
        news_enricher,
        "analyze_news_row_with_llm",
        lambda row: {
            "relevance": "high",
            "sentiment": "positive",
            "key_discussion": "Demand resilience",
            "summary": "LLM summary",
            "reason_growth": "Orders remain strong",
            "reason_decrease": "",
            "raw_json": {"provider": "llm"},
        },
    )
    result = news_enricher.classify_news_row(
        {
            "title": "Apple expands AI features",
            "summary": "New devices and software updates were announced.",
        }
    )
    assert result["analysis_source"] == "llm"
    assert result["sentiment"] == "positive"
    assert result["summary"] == "LLM summary"


def test_build_analysis_rows_prefers_batch_llm_and_dedupes(monkeypatch):
    monkeypatch.setattr(news_enricher, "llm_enrichment_enabled", lambda: True)
    monkeypatch.setattr(news_enricher, "get_env_int", lambda key, default=0: 8)
    monkeypatch.setattr(
        news_enricher,
        "analyze_news_rows_with_llm",
        lambda rows: {
            "news-1": {
                "relevance": "high",
                "sentiment": "positive",
                "key_discussion": "Batch result",
                "summary": "Batch summary",
                "reason_growth": "Growth",
                "reason_decrease": "",
                "raw_json": {"provider": "batch"},
            }
        },
    )
    monkeypatch.setattr(news_enricher, "analyze_news_row_with_llm", lambda row: None)
    rows = news_enricher.build_analysis_rows(
        symbol="AAPL",
        news_rows=[
            {"id": "news-1", "trade_date": "2026-03-10", "title": "Same title", "summary": "Same summary"},
            {"id": "news-2", "trade_date": "2026-03-10", "title": "Same title", "summary": "Same summary"},
        ],
        ohlc_rows=[],
    )
    rows, stats = rows
    assert len(rows) == 1
    assert rows[0]["analysis_source"] == "llm"
    assert rows[0]["summary"] == "Batch summary"
    assert stats["deduped_count"] == 1
    assert stats["llm_count"] == 1


def test_enrich_news_for_symbol_skips_existing(monkeypatch):
    class DummyStore:
        def get_news_items(self, symbol, start_date=None, end_date=None, limit=200):
            return [
                {"id": "news-1", "trade_date": "2026-03-10", "title": "One", "summary": "One"},
                {"id": "news-2", "trade_date": "2026-03-11", "title": "Two", "summary": "Two"},
            ]

        def get_analyzed_news_ids(self, symbol, start_date=None, end_date=None):
            return {"news-1"}

        def get_ohlc(self, symbol, start_date, end_date):
            return []

        def upsert_news_analysis(self, symbol, rows, analysis_source="local"):
            self.rows = rows
            return len(rows)

    monkeypatch.setattr(
        news_enricher,
        "build_analysis_rows",
        lambda symbol, news_rows, ohlc_rows: (
            [
                {
                    "news_id": row["id"],
                    "trade_date": row["trade_date"],
                    "summary": row["summary"],
                    "analysis_source": "local",
                }
                for row in news_rows
            ],
            {"deduped_count": 0, "llm_count": 0, "local_count": len(news_rows)},
        ),
    )
    store = DummyStore()
    result = news_enricher.enrich_news_for_symbol(store, "AAPL")
    assert result["news_count"] == 2
    assert result["queued_count"] == 1
    assert result["skipped_existing_count"] == 1
    assert len(store.rows) == 1
    assert store.rows[0]["news_id"] == "news-2"


def test_enrich_news_for_symbol_only_reanalyzes_local(monkeypatch):
    class DummyStore:
        def get_news_items(self, symbol, start_date=None, end_date=None, limit=200):
            return [
                {"id": "news-1", "trade_date": "2026-03-10", "title": "One", "summary": "One"},
                {"id": "news-2", "trade_date": "2026-03-11", "title": "Two", "summary": "Two"},
                {"id": "news-3", "trade_date": "2026-03-12", "title": "Three", "summary": "Three"},
            ]

        def get_analyzed_news_sources(self, symbol, start_date=None, end_date=None):
            return {"news-1": "local", "news-2": "llm"}

        def get_ohlc(self, symbol, start_date, end_date):
            return []

        def upsert_news_analysis(self, symbol, rows, analysis_source="local"):
            self.rows = rows
            return len(rows)

    monkeypatch.setattr(
        news_enricher,
        "build_analysis_rows",
        lambda symbol, news_rows, ohlc_rows: (
            [
                {
                    "news_id": row["id"],
                    "trade_date": row["trade_date"],
                    "summary": row["summary"],
                    "analysis_source": "llm" if row["id"] == "news-1" else "local",
                }
                for row in news_rows
            ],
            {"deduped_count": 0, "llm_count": 1, "local_count": 0},
        ),
    )

    store = DummyStore()
    result = news_enricher.enrich_news_for_symbol(
        store,
        "AAPL",
        only_reanalyze_local=True,
    )

    assert result["news_count"] == 3
    assert result["queued_count"] == 1
    assert result["skipped_existing_count"] == 2
    assert result["only_reanalyze_local"] is True
    assert result["upgraded_local_to_llm_count"] == 1
    assert result["execution_summary"]["upgraded_dates"] == ["2026-03-10"]
    assert result["execution_summary"]["remaining_local_titles"] == []
    assert result["execution_summary"]["skipped_missing_analysis_count"] == 1
    assert result["execution_summary"]["skipped_non_local_count"] == 1
    assert [row["news_id"] for row in store.rows] == ["news-1"]