fix: 日报摘要可靠性(去模型兜底+重试) 与取数逻辑优化

- llm/client: 移除内置默认模型兜底(deepseek-chat/qwen-plus), 模型必须显式配置否则报错
- reporter._llm_call: 指数退避重试(LLM_RETRY_TIMES 默认3 / LLM_RETRY_BACKOFF_SEC 默认2s)
- pipeline report: report_date 改为当天(原昨天+回溯3天)
- reporter._collect_news_events: 读当天+前一天目录, publish_time 30h 回溯(NEWS_LOOKBACK_HOURS=30), 统一时区
- reporter._collect_xwlb: 固定取 day_str 前一日(已播出联播), source_date 同步
- 公告/调研/互动保持近15日设置(CNINFO_DAYS_BACK), 不受 30h 影响
- 测试: 新增 30h回溯/带时区/重试/模型缺失/xwlb 前一日 用例

tests/test_llm.py

@@ -322,21 +322,24 @@ async def test_extract_event_async_retries(fake_config: LLMConfig) -> None:
 def test_load_llm_config_deepseek_from_env(monkeypatch: pytest.MonkeyPatch) -> None:
     monkeypatch.setenv("LLM_PROVIDER", "deepseek")
     monkeypatch.setenv("DEEPSEEK_API_KEY", "sk-test-deepseek")
+    monkeypatch.setenv("DEEPSEEK_MODEL", "deepseek-v4-flash")
     monkeypatch.delenv("LLM_MODEL", raising=False)
     cfg = load_llm_config()
     assert cfg.provider == "deepseek"
     assert cfg.api_key == "sk-test-deepseek"
-    assert cfg.model.startswith("deepseek")
+    assert cfg.model == "deepseek-v4-flash"
     assert "deepseek" in cfg.base_url
 
 
 def test_load_llm_config_qwen_from_env(monkeypatch: pytest.MonkeyPatch) -> None:
     monkeypatch.setenv("LLM_PROVIDER", "qwen")
     monkeypatch.setenv("DASHSCOPE_API_KEY", "sk-test-qwen")
+    monkeypatch.setenv("QWEN_MODEL", "qwen-plus")
     monkeypatch.delenv("LLM_MODEL", raising=False)
     cfg = load_llm_config()
     assert cfg.provider == "qwen"
     assert cfg.api_key == "sk-test-qwen"
+    assert cfg.model == "qwen-plus"
     assert "dashscope" in cfg.base_url or "aliyuncs" in cfg.base_url
 
 
@@ -347,5 +350,15 @@ def test_load_llm_config_unknown_provider_raises(monkeypatch: pytest.MonkeyPatch
 
 def test_load_llm_config_missing_key_raises(monkeypatch: pytest.MonkeyPatch) -> None:
     monkeypatch.delenv("DEEPSEEK_API_KEY", raising=False)
-    with pytest.raises(ValueError):
+    monkeypatch.setenv("DEEPSEEK_MODEL", "deepseek-v4-flash")
+    with pytest.raises(ValueError, match="API key"):
+        load_llm_config(provider="deepseek")
+
+
+def test_load_llm_config_missing_model_raises(monkeypatch: pytest.MonkeyPatch) -> None:
+    """去掉内置默认模型后：未显式配置模型必须报错（不再回退 deepseek-chat）。"""
+    monkeypatch.setenv("DEEPSEEK_API_KEY", "sk-test")
+    monkeypatch.delenv("DEEPSEEK_MODEL", raising=False)
+    monkeypatch.delenv("LLM_MODEL", raising=False)
+    with pytest.raises(ValueError, match="模型"):
         load_llm_config(provider="deepseek")

tests/test_report_builder.py

@@ -5,6 +5,8 @@ from __future__ import annotations
 import json
 from datetime import date
 
+import pytest
+
 from scheduler.reporter import _build_report_data
 
 
@@ -85,3 +87,139 @@ class TestBuildReportData:
         r = _build_report_data(news, cninfo, {}, None, "20260710")
         assert r.events == []
         assert r.ai_summary is None
+
+
+class TestLlmCallRetry:
+    """_llm_call 重试逻辑（纯逻辑，mock client）。"""
+
+    @staticmethod
+    def _fake_client(failures: int):
+        """构造 mock client：前 failures 次抛 ConnectionError，之后成功。"""
+        from types import SimpleNamespace
+
+        n = {"count": 0}
+
+        class Completions:
+            def create(self, **kwargs):
+                n["count"] += 1
+                if n["count"] <= failures:
+                    raise ConnectionError("transient")
+                return SimpleNamespace(
+                    choices=[SimpleNamespace(
+                        message=SimpleNamespace(content="今日要点摘要"),
+                        finish_reason="stop",
+                    )]
+                )
+
+        return SimpleNamespace(chat=SimpleNamespace(completions=Completions())), n
+
+    def test_success_first_try(self) -> None:
+        from scheduler.reporter import _llm_call
+        client, n = self._fake_client(0)
+        out = _llm_call(client, "deepseek-v4-flash", "p")
+        assert out == "今日要点摘要"
+        assert n["count"] == 1
+
+    def test_retry_then_success(self, monkeypatch) -> None:
+        import scheduler.reporter as rep
+        monkeypatch.setattr(rep, "_LLM_RETRY_TIMES", 3)
+        monkeypatch.setattr(rep, "_LLM_RETRY_BACKOFF_SEC", 0.01)
+        client, n = self._fake_client(2)  # 前 2 次失败，第 3 次成功
+        out = rep._llm_call(client, "deepseek-v4-flash", "p")
+        assert out == "今日要点摘要"
+        assert n["count"] == 3
+
+    def test_exhausts_retries_raises(self, monkeypatch) -> None:
+        import scheduler.reporter as rep
+        monkeypatch.setattr(rep, "_LLM_RETRY_TIMES", 2)
+        monkeypatch.setattr(rep, "_LLM_RETRY_BACKOFF_SEC", 0.01)
+        client, n = self._fake_client(99)  # 一直失败
+        with pytest.raises(ConnectionError):
+            rep._llm_call(client, "deepseek-v4-flash", "p")
+        assert n["count"] == 2  # 重试 2 次后放弃
+
+
+class TestCollectXwlb:
+    """_collect_xwlb 取数逻辑：应查询日报前一日(已播出的联播)，并跳过内容提要。"""
+
+    def test_queries_previous_day_and_skips_toc(self, monkeypatch) -> None:
+        import json as _json
+        import urllib.request
+
+        captured: dict[str, str] = {}
+
+        def fake_urlopen(req, timeout=15):  # noqa: ARG001
+            captured["url"] = req.full_url
+
+            class Resp:
+                def __enter__(self):
+                    return self
+
+                def __exit__(self, *args):
+                    return False
+
+                def read(self):
+                    return _json.dumps({"data": {"news": [
+                        {"daily_sub_id": 1, "news_title": "内容提要", "news_days": "2026-08-04", "news_improve": "开场白"},
+                        {"daily_sub_id": 2, "news_title": "联播要闻A", "news_days": "2026-08-04", "news_improve": "正文A"},
+                    ]}}).encode("utf-8")
+
+            return Resp()
+
+        monkeypatch.setattr(urllib.request, "urlopen", fake_urlopen)
+        from scheduler.reporter import _collect_xwlb
+
+        result = _collect_xwlb("20260805")
+        # 查询的是前一日(20260804)而非当日
+        assert "start_date=20260804" in captured["url"]
+        assert "end_date=20260804" in captured["url"]
+        # 跳过第 1 条内容提要
+        assert len(result["items"]) == 1
+        assert result["items"][0]["title"] == "联播要闻A"
+        assert result["source_date"] == "20260804"
+        assert result["date"] == "08月04日"
+
+
+class TestCollectNewsEventsLookback:
+    """_collect_news_events 30 小时回溯逻辑。"""
+
+    def test_filters_30h_and_excludes_cninfo(self, monkeypatch) -> None:
+        from datetime import datetime, timedelta
+
+        import scheduler.reporter as rep
+
+        now = datetime.now().astimezone()
+
+        def fake_load(day_str: str) -> list[dict]:  # noqa: ARG001
+            def ev(title: str, hours_ago: float | None, source: str = "cls",
+                   importance: int = 5, aware: bool = False) -> dict:
+                pt = None
+                if hours_ago is not None:
+                    t = now - timedelta(hours=hours_ago)
+                    pt = t.isoformat() if not aware else t.astimezone().isoformat()
+                return {
+                    "title": title, "url": "u", "source_id": source,
+                    "publish_time": pt,
+                    "event": {"importance": importance, "sentiment": "neutral",
+                              "event_type": "其他", "summary": "s"},
+                }
+
+            return [
+                ev("窗口内新闻", 10),
+                ev("窗口内新闻带时区", 12, aware=True),
+                ev("窗口外旧闻", 40),
+                ev("无时间戳", None),
+                ev("公告排除", 5, source="cninfo", importance=2),
+            ]
+
+        monkeypatch.setattr(rep, "_load_events_from_dir", fake_load)
+        result = rep._collect_news_events("20260805")
+
+        # 两个日期目录各返回 5 条(共 10): 旧闻×2、公告×2 被滤, 保留 6 条
+        assert result["total"] == 6
+        titles = {e["title"] for e in result["high"]}
+        assert "窗口内新闻" in titles
+        assert "窗口内新闻带时区" in titles
+        assert "无时间戳" in titles
+        assert "窗口外旧闻" not in titles
+        assert "公告排除" not in titles