X-Financial/server/tests/test_knowledge_normalizer.py

from __future__ import annotations

from sqlalchemy import create_engine
from sqlalchemy.orm import Session, sessionmaker
from sqlalchemy.pool import StaticPool

from app.db.base import Base
from app.services.knowledge_normalizer import KnowledgeNormalizationService


def build_session_factory() -> sessionmaker[Session]:
    engine = create_engine(
        "sqlite+pysqlite:///:memory:",
        connect_args={"check_same_thread": False},
        poolclass=StaticPool,
    )
    Base.metadata.create_all(bind=engine)
    return sessionmaker(bind=engine, autoflush=False, autocommit=False)


def test_knowledge_normalizer_appends_structured_table(monkeypatch) -> None:
    session_factory = build_session_factory()
    raw_text = (
        "表3 出差补贴标准\n\n"
        "单位：人民币元/天\n"
        "补助类型 项目 港澳台 直辖市/特区/西藏 其他地区 国外\n"
        "餐补 自行解决餐食 75 65 55 140\n"
        "基本出差补贴 35 35 35 35\n"
        "合计 110 100 90 175\n"
    )
    with session_factory() as db:
        service = KnowledgeNormalizationService(db)
        monkeypatch.setattr(
            service.runtime_chat_service,
            "complete",
            lambda *args, **kwargs: (
                "| 补助类型 | 港澳台 | 直辖市/特区/西藏 | 其他地区 | 国外 |\n"
                "|---|---:|---:|---:|---:|\n"
                "| 餐补 | 75 | 65 | 55 | 140 |\n"
                "| 基本出差补贴 | 35 | 35 | 35 | 35 |\n"
                "| 合计 | 110 | 100 | 90 | 175 |"
            ),
        )

        enriched = service.build_enriched_text(raw_text)

    assert enriched.startswith(raw_text.strip())
    assert "| 餐补 | 75 | 65 | 55 | 140 |" in enriched
    assert enriched.endswith("| 合计 | 110 | 100 | 90 | 175 |")


def test_knowledge_normalizer_keeps_only_markdown_table_body() -> None:
    cleaned = KnowledgeNormalizationService._sanitize_answer(
        "## 表3 出差补贴标准\n\n"
        "| 补助类型 | 港澳台 | 直辖市/特区/西藏 |\n"
        "|---|---:|---:|\n"
        "| 餐补 | 75 | 65 |\n\n"
        "注：主办方统一安排餐食时，不再报销餐补。"
    )

    assert cleaned == (
        "| 补助类型 | 港澳台 | 直辖市/特区/西藏 |\n"
        "|---|---:|---:|\n"
        "| 餐补 | 75 | 65 |"
    )


def test_knowledge_normalizer_builds_section_navigation_without_table() -> None:
    session_factory = build_session_factory()
    raw_text = (
        "第一章 总则\n"
        "本制度适用于员工差旅报销和审批管理。\n\n"
        "第二章 住宿费标准\n"
        "住宿费按照出差城市档位和职级标准执行。\n\n"
        "第三章 交通费标准\n"
        "交通费应结合出差工具、舱位和审批要求报销。\n"
    )
    with session_factory() as db:
        service = KnowledgeNormalizationService(db)
        enriched = service.build_enriched_text(raw_text)

    assert enriched.startswith(raw_text.strip())
    assert "- 第一章 总则" in enriched
    assert "## 第二章 住宿费标准" in enriched
    assert "# 问答线索补充" in enriched
    assert "- 第二章 住宿费标准：住宿费按照出差城市档位和职级标准执行" in enriched
    assert "# 章节导航" in enriched


def test_knowledge_normalizer_builds_answer_clues_from_lists_and_kv_lines() -> None:
    session_factory = build_session_factory()
    raw_text = (
        "第一章 报销要求\n"
        "报销时限：费用发生后 30 日内提交申请。\n"
        "- 超过 30 日需补充审批说明。\n"
        "第十条 发票遗失的，应先提交遗失说明。\n"
    )

    with session_factory() as db:
        service = KnowledgeNormalizationService(db)
        enriched = service.build_enriched_text(raw_text)

    assert "# 问答线索补充" in enriched
    assert "- 第一章 报销要求：报销时限：费用发生后 30 日内提交申请" in enriched
    assert "- 第一章 报销要求：超过 30 日需补充审批说明" in enriched
    assert "- 第一章 报销要求：第十条 发票遗失的，应先提交遗失说明" in enriched


def test_knowledge_normalizer_builds_answer_clues_without_section_headings() -> None:
    session_factory = build_session_factory()
    raw_text = (
        "报销时限：费用发生后 30 日内提交申请。\n"
        "超过 30 日需补充审批说明。\n"
        "审批材料包括发票、行程单和付款凭证。\n"
    )

    with session_factory() as db:
        service = KnowledgeNormalizationService(db)
        enriched = service.build_enriched_text(raw_text)

    assert "# 问答线索补充" in enriched
    assert "- 正文：报销时限：费用发生后 30 日内提交申请" in enriched
    assert "- 正文：超过 30 日需补充审批说明" in enriched