feat: 日报结构化入库（M10 前后端分离数据层）

- 新增 report_db 包: MySQL 连接/建表/幂等写入 (news_report/news_event, myquant 库)
- 新增 report_import 包: 历史 178 份日报 HTML 解析入库, 表头驱动列映射
- reporter.py 完全切换: generate_report 结构化入库, 不再生成/上传 HTML
- CLI: 新增 report-import 子命令
- 依赖: uv add pymysql; 配置: NEWS_DB_* / REPORT_HISTORY_DIR
- 文档: docs/report_db_design.md(实现逻辑), docs/db_schema.md(表结构供 API/前端)
- 测试: 24 个单测通过 (parser/builder/models/importer)

tests/test_report_parser.py

@@ -0,0 +1,98 @@
+"""历史日报解析器单元测试（基于真实样例 HTML）。"""
+
+from __future__ import annotations
+
+from datetime import date, datetime
+from pathlib import Path
+
+import pytest
+
+from report_import.parser import (
+    ReportParseError,
+    parse_finance_report,
+    parse_intl_report,
+    parse_report,
+)
+
+FIXTURES = Path(__file__).parent / "fixtures"
+FINANCE_HTML = (FIXTURES / "finance_news_daily_20260710_0720.html").read_text(encoding="utf-8")
+INTL_HTML = (FIXTURES / "intl_news_daily_20260711_070304.html").read_text(encoding="utf-8")
+
+
+class TestFinanceParse:
+    def test_metadata(self) -> None:
+        r = parse_finance_report(FINANCE_HTML, "finance_news_daily_20260710_0720.html")
+        assert r.report_date == date(2026, 7, 10)
+        assert r.report_type == "finance"
+        assert r.file_name == "finance_news_daily_20260710_0720.html"
+        assert r.generated_at == datetime(2026, 7, 11, 7, 20, 27)  # header"生成于"优先
+
+    def test_ai_summary_lines(self) -> None:
+        r = parse_finance_report(FINANCE_HTML, "finance_news_daily_20260710_0720.html")
+        assert r.ai_summary is not None
+        assert len(r.ai_summary.splitlines()) >= 5
+        assert "碳达峰" in r.ai_summary
+
+    def test_sections_and_ranks(self) -> None:
+        r = parse_finance_report(FINANCE_HTML, "finance_news_daily_20260710_0720.html")
+        sections = {e.section for e in r.events}
+        assert sections == {"xwlb", "news", "cninfo"}
+        assert sum(1 for e in r.events if e.section == "xwlb") == 16
+        assert sum(1 for e in r.events if e.section == "news") == 20
+        assert sum(1 for e in r.events if e.section == "cninfo") == 20
+
+    def test_event_fields(self) -> None:
+        r = parse_finance_report(FINANCE_HTML, "finance_news_daily_20260710_0720.html")
+        xwlb = next(e for e in r.events if e.section == "xwlb")
+        assert xwlb.importance == 4
+        assert xwlb.event_type == "新闻联播"
+        assert xwlb.sentiment == "neutral"
+        assert "张国清" in xwlb.title
+        news = next(e for e in r.events if e.section == "news" and e.source)
+        assert news.source  # 新闻板块带来源
+
+    def test_stats_keys(self) -> None:
+        r = parse_finance_report(FINANCE_HTML, "finance_news_daily_20260710_0720.html")
+        for key in ("pipeline", "sources", "sentiment", "importance", "event_types"):
+            assert key in r.stats, f"缺少 stats.{key}"
+        assert r.stats["pipeline"]["M1 原始文章"] == 758
+
+
+class TestIntlParse:
+    def test_metadata(self) -> None:
+        r = parse_intl_report(INTL_HTML, "intl_news_daily_20260711_070304.html")
+        assert r.report_date == date(2026, 7, 11)
+        assert r.report_type == "intl"
+        assert r.generated_at == datetime(2026, 7, 11, 7, 3, 30)
+
+    def test_events_and_source_from_small(self) -> None:
+        r = parse_intl_report(INTL_HTML, "intl_news_daily_20260711_070304.html")
+        assert len(r.events) == 19
+        first = r.events[0]
+        assert first.section == "intl"
+        assert first.source == "investinglive.com"  # 从摘要 <small>[来源]</small> 提取
+        assert first.url.startswith("https://investinglive.com/")
+        assert first.importance == 4
+        assert first.event_type == "地缘政治"
+        # 摘要中不应残留 [来源] 标记
+        assert first.summary is not None and "[investinglive.com]" not in first.summary
+
+    def test_stats_keys(self) -> None:
+        r = parse_intl_report(INTL_HTML, "intl_news_daily_20260711_070304.html")
+        for key in ("pipeline", "sentiment", "importance", "event_types", "source_dist"):
+            assert key in r.stats
+        assert r.stats["source_dist"][0]["来源"] == "ForexLive"
+
+
+class TestParseReportDispatch:
+    def test_dispatch_finance(self) -> None:
+        r = parse_report(FINANCE_HTML, "finance_news_daily_20260710_0720.html")
+        assert r.report_type == "finance"
+
+    def test_dispatch_intl(self) -> None:
+        r = parse_report(INTL_HTML, "intl_news_daily_20260711_070304.html")
+        assert r.report_type == "intl"
+
+    def test_bad_filename_raises(self) -> None:
+        with pytest.raises(ReportParseError):
+            parse_report("<html></html>", "not_a_report.html")