X-Financial/server/tests/test_user_agent_service.py

from __future__ import annotations

from datetime import UTC, datetime, timedelta

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

from app.db.base import Base
from app.schemas.ontology import OntologyParseRequest
from app.schemas.user_agent import UserAgentRequest
from app.services.ontology import SemanticOntologyService
from app.services.user_agent import UserAgentService


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_user_agent_query_returns_readable_answer_and_actions() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="张三 4 月差旅报销金额是多少",
                user_id="pytest",
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="张三 4 月差旅报销金额是多少",
                ontology=ontology,
                tool_payload={"record_count": 2, "total_amount": 8800.0},
            )
        )

        assert "8800.00" in response.answer
        assert len(response.suggested_actions) >= 1


def test_user_agent_returns_readable_query_answer_when_runtime_model_is_skipped(monkeypatch) -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="张三 4 月差旅报销金额是多少",
                user_id="pytest",
            )
        )
        service = UserAgentService(db)
        monkeypatch.setattr(service, "_generate_answer_with_model", lambda *args, **kwargs: "这是模型回答")

        response = service.respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="张三 4 月差旅报销金额是多少",
                ontology=ontology,
                tool_payload={"record_count": 2, "total_amount": 8800.0},
            )
        )

        assert "共 2 笔" in response.answer
        assert "8800.00" in response.answer


def test_user_agent_sanitizes_model_thinking_blocks() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        service = UserAgentService(db)

        assert (
            service._sanitize_model_answer("<think>内部推理</think>\n最终答复")
            == "最终答复"
        )


def test_user_agent_guides_generic_expense_request() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我要报销",
                user_id="pytest",
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我要报销",
                ontology=ontology,
                tool_payload={"record_count": 9, "total_amount": 12345.0},
            )
        )

        assert response.review_payload is not None
        assert response.answer == response.review_payload.body_message
        assert response.review_payload.can_proceed is False
        assert response.review_payload.missing_slots == [
            "报销类型",
            "发生时间",
            "金额",
            "事由说明",
            "票据附件",
        ]
        assert [item.action_type for item in response.review_payload.confirmation_actions] == [
            "cancel_review",
            "edit_review",
            "save_draft",
        ]


def test_user_agent_guides_implicit_expense_draft_request() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        today = datetime.now(UTC).date().isoformat()
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我今天去客户现场，招待了客户，花销了1000元",
                user_id="pytest",
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我今天去客户现场，招待了客户，花销了1000元",
                ontology=ontology,
                tool_payload={"draft_only": True},
            )
        )

        assert response.review_payload is not None
        assert response.answer == response.review_payload.body_message
        assert response.review_payload.intent_summary.startswith("识别到您希望报销一笔“业务招待费”费用。")
        assert response.review_payload.missing_slots == ["客户名称", "参与人员", "票据附件"]
        assert [item.action_type for item in response.review_payload.confirmation_actions] == [
            "cancel_review",
            "edit_review",
            "save_draft",
        ]

        slot_map = {item.key: item for item in response.review_payload.slot_cards}
        assert slot_map["expense_type"].value == "业务招待费"
        assert slot_map["time_range"].value == today
        assert slot_map["time_range"].raw_value == "今天"
        assert slot_map["location"].value == "客户现场"
        assert slot_map["amount"].value == "1000.00元"


def test_user_agent_guides_narrative_with_day_before_yesterday() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我前天请客户吃饭花了200元",
                user_id="pytest",
                context_json={
                    "client_now_iso": "2026-05-12T16:30:00.000Z",
                    "client_timezone_offset_minutes": -480,
                },
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我前天请客户吃饭花了200元",
                ontology=ontology,
                tool_payload={"draft_only": True},
            )
        )

        assert response.review_payload is not None
        slot_map = {item.key: item for item in response.review_payload.slot_cards}
        assert slot_map["time_range"].raw_value == "前天"
        assert slot_map["time_range"].value == "2026-05-11"
        assert "时间为 2026-05-11" in response.review_payload.intent_summary


def test_user_agent_attachment_only_upload_uses_generic_scene_reason_without_fabrication() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我上传了 1 份票据，请结合附件名称给出报销建议并尽量生成草稿。",
                user_id="pytest",
                context_json={
                    "attachment_names": ["didi-trip.png"],
                    "attachment_count": 1,
                    "ocr_documents": [
                        {
                            "filename": "didi-trip.png",
                            "summary": "滴滴出行 订单金额 32 元",
                            "text": "滴滴出行 订单金额 32 元",
                            "document_type": "taxi_receipt",
                            "scene_code": "transport",
                        }
                    ],
                    "user_input_text": "",
                },
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我上传了 1 份票据，请结合附件名称给出报销建议并尽量生成草稿。\n附件名称：didi-trip.png",
                ontology=ontology,
                context_json={
                    "attachment_names": ["didi-trip.png"],
                    "attachment_count": 1,
                    "ocr_documents": [
                        {
                            "filename": "didi-trip.png",
                            "summary": "滴滴出行 订单金额 32 元",
                            "text": "滴滴出行 订单金额 32 元",
                            "document_type": "taxi_receipt",
                            "scene_code": "transport",
                        }
                    ],
                    "user_input_text": "",
                },
                tool_payload={"draft_only": True},
            )
        )

        assert response.review_payload is not None
        slot_map = {item.key: item for item in response.review_payload.slot_cards}
        assert slot_map["reason"].value == "交通出行"
        assert slot_map["reason"].status == "inferred"


def test_user_agent_transport_flow_infers_reason_and_does_not_require_location_or_merchant() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我上传了交通票据，帮我生成报销草稿",
                user_id="pytest",
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我上传了交通票据，帮我生成报销草稿",
                ontology=ontology,
                context_json={
                    "attachment_names": ["didi-trip.png"],
                    "attachment_count": 1,
                    "ocr_documents": [
                        {
                            "filename": "didi-trip.png",
                            "summary": "滴滴出行 支付金额 32 元",
                            "text": "滴滴出行 支付金额 32 元",
                            "document_type": "taxi_receipt",
                            "scene_code": "transport",
                            "scene_label": "交通票据",
                        }
                    ],
                },
                tool_payload={"draft_only": True},
            )
        )

        assert response.review_payload is not None
        slot_map = {item.key: item for item in response.review_payload.slot_cards}
        assert slot_map["reason"].value == "交通出行"
        assert slot_map["reason"].status == "inferred"
        assert "酒店/商户" not in response.review_payload.missing_slots
        assert "地点" not in response.review_payload.missing_slots
        assert "事由说明" not in response.review_payload.missing_slots


def test_user_agent_risk_response_includes_rule_citations() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="检查重复报销风险",
                user_id="pytest",
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="检查重复报销风险",
                ontology=ontology,
                tool_payload={"risk_flags": ["duplicate_expense"]},
            )
        )

        assert response.risk_flags == ["duplicate_expense"]
        assert any(item.source_type == "rule" for item in response.citations)
        assert "duplicate_expense" in response.answer


def test_user_agent_draft_returns_structured_payload() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="帮我生成张三4月差旅报销草稿",
                user_id="pytest",
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="帮我生成张三4月差旅报销草稿",
                ontology=ontology,
                tool_payload={"draft_only": True},
            )
        )

        assert response.draft_payload is not None
        assert response.draft_payload.confirmation_required is True
        assert response.review_payload is not None
        assert response.review_payload.can_proceed is False
        assert response.review_payload.missing_slots == ["金额", "事由说明", "票据附件"]
        assert [item.action_type for item in response.review_payload.confirmation_actions] == [
            "cancel_review",
            "edit_review",
            "save_draft",
        ]
        assert response.answer == response.review_payload.body_message


def test_user_agent_returns_draft_limit_message_when_save_is_blocked() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="请按当前识别信息保存报销草稿",
                user_id="pytest",
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="请按当前识别信息保存报销草稿",
                ontology=ontology,
                context_json={"review_action": "save_draft"},
                tool_payload={
                    "draft_limit_reached": True,
                    "message": "你当前已保存 3 个草稿，请先完成已保存的草稿，才能再次新建草稿。",
                    "status": "blocked",
                },
            )
        )

        assert (
            response.answer
            == "你当前已保存 3 个草稿，请先完成已保存的草稿，才能再次新建草稿。"
        )


def test_user_agent_builds_review_payload_for_multi_document_expense_flow() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        yesterday = (datetime.now(UTC).date() - timedelta(days=1)).isoformat()
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我昨天去上海出差，还请客户A吃饭，帮我生成报销草稿",
                user_id="pytest",
                context_json={
                    "attachment_names": ["机票行程单.png", "餐饮发票.jpg"],
                    "attachment_count": 2,
                    "ocr_documents": [
                        {
                            "filename": "机票行程单.png",
                            "summary": "机票行程单 上海-北京 金额 680 元",
                            "text": "机票行程单 上海-北京 金额 680 元",
                            "avg_score": 0.93,
                            "warnings": [],
                        },
                        {
                            "filename": "餐饮发票.jpg",
                            "summary": "餐饮发票 客户招待 金额 320 元",
                            "text": "餐饮发票 客户招待 金额 320 元",
                            "avg_score": 0.91,
                            "warnings": [],
                        },
                    ],
                },
            )
        )
        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我昨天去上海出差，还请客户A吃饭，帮我生成报销草稿",
                ontology=ontology,
                context_json={
                    "name": "张三",
                    "attachment_names": ["机票行程单.png", "餐饮发票.jpg"],
                    "attachment_count": 2,
                    "ocr_documents": [
                        {
                            "filename": "机票行程单.png",
                            "summary": "机票行程单 上海-北京 金额 680 元",
                            "text": "机票行程单 上海-北京 金额 680 元",
                            "avg_score": 0.93,
                            "warnings": [],
                        },
                        {
                            "filename": "餐饮发票.jpg",
                            "summary": "餐饮发票 客户招待 金额 320 元",
                            "text": "餐饮发票 客户招待 金额 320 元",
                            "avg_score": 0.91,
                            "warnings": [],
                        },
                    ],
                },
                tool_payload={"draft_only": True, "claim_no": "EXP-202605-009", "status": "draft"},
            )
        )

        assert response.review_payload is not None
        assert len(response.review_payload.document_cards) == 2
        assert len(response.review_payload.claim_groups) == 2
        assert response.review_payload.missing_slots == ["参与人员"]
        assert [item.action_type for item in response.review_payload.confirmation_actions] == [
            "cancel_review",
            "edit_review",
            "save_draft",
        ]
        assert any(item.scene_label == "业务招待费" for item in response.review_payload.document_cards)
        assert f"时间为 {yesterday}" in response.review_payload.intent_summary
        slot_map = {item.key: item for item in response.review_payload.slot_cards}
        assert slot_map["time_range"].value == yesterday
        assert slot_map["time_range"].raw_value == "昨天"


def test_user_agent_sums_multi_document_amounts_from_synonym_fields() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我上传了两张交通票据，帮我生成报销草稿",
                user_id="pytest",
                context_json={
                    "attachment_names": ["滴滴行程单.png", "停车票.jpg"],
                    "attachment_count": 2,
                    "ocr_documents": [
                        {
                            "filename": "滴滴行程单.png",
                            "summary": "滴滴出行电子行程单",
                            "text": "滴滴出行 订单金额 ￥32.50",
                            "avg_score": 0.94,
                            "document_fields": [
                                {"key": "amount", "label": "支付金额", "value": "32.50"},
                            ],
                            "warnings": [],
                        },
                        {
                            "filename": "停车票.jpg",
                            "summary": "停车票",
                            "text": "停车费 合计 18 元",
                            "avg_score": 0.92,
                            "document_fields": [
                                {"key": "total_amount", "label": "合计金额", "value": "18"},
                            ],
                            "warnings": [],
                        },
                    ],
                },
            )
        )

        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我上传了两张交通票据，帮我生成报销草稿",
                ontology=ontology,
                context_json={
                    "attachment_names": ["滴滴行程单.png", "停车票.jpg"],
                    "attachment_count": 2,
                    "ocr_documents": [
                        {
                            "filename": "滴滴行程单.png",
                            "summary": "滴滴出行电子行程单",
                            "text": "滴滴出行 订单金额 ￥32.50",
                            "avg_score": 0.94,
                            "document_fields": [
                                {"key": "amount", "label": "支付金额", "value": "32.50"},
                            ],
                            "warnings": [],
                        },
                        {
                            "filename": "停车票.jpg",
                            "summary": "停车票",
                            "text": "停车费 合计 18 元",
                            "avg_score": 0.92,
                            "document_fields": [
                                {"key": "total_amount", "label": "合计金额", "value": "18"},
                            ],
                            "warnings": [],
                        },
                    ],
                },
                tool_payload={"draft_only": True},
            )
        )

        assert response.review_payload is not None
        slot_map = {item.key: item for item in response.review_payload.slot_cards}
        assert slot_map["amount"].value == "50.50元"
        document_field_labels = [
            field.label
            for card in response.review_payload.document_cards
            for field in card.fields
        ]
        assert "金额" in document_field_labels


def test_user_agent_prefers_larger_decimal_amount_from_ocr_text_candidates() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我上传了打车票据，帮我生成报销草稿",
                user_id="pytest",
                context_json={
                    "attachment_names": ["滴滴行程单.png"],
                    "attachment_count": 1,
                    "ocr_documents": [
                        {
                            "filename": "滴滴行程单.png",
                            "summary": "滴滴出行电子行程单",
                            "text": "滴滴出行 支付金额 1 元，实付 13.4 元，订单号 12345678",
                            "avg_score": 0.94,
                            "warnings": [],
                        },
                    ],
                },
            )
        )

        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我上传了打车票据，帮我生成报销草稿",
                ontology=ontology,
                context_json={
                    "attachment_names": ["滴滴行程单.png"],
                    "attachment_count": 1,
                    "ocr_documents": [
                        {
                            "filename": "滴滴行程单.png",
                            "summary": "滴滴出行电子行程单",
                            "text": "滴滴出行 支付金额 1 元，实付 13.4 元，订单号 12345678",
                            "avg_score": 0.94,
                            "warnings": [],
                        },
                    ],
                },
                tool_payload={"draft_only": True},
            )
        )

        assert response.review_payload is not None
        slot_map = {item.key: item for item in response.review_payload.slot_cards}
        assert slot_map["amount"].value == "13.40元"


def test_user_agent_review_payload_keeps_document_preview_data() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我上传了打车票据，帮我生成报销草稿",
                user_id="pytest",
                context_json={
                    "attachment_names": ["滴滴行程单.png"],
                    "attachment_count": 1,
                    "ocr_documents": [
                        {
                            "filename": "滴滴行程单.png",
                            "summary": "滴滴出行电子行程单",
                            "text": "滴滴出行 实付 13.4 元",
                            "avg_score": 0.94,
                            "preview_kind": "image",
                            "preview_data_url": "data:image/png;base64,ZmFrZQ==",
                            "warnings": [],
                        },
                    ],
                },
            )
        )

        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我上传了打车票据，帮我生成报销草稿",
                ontology=ontology,
                context_json={
                    "attachment_names": ["滴滴行程单.png"],
                    "attachment_count": 1,
                    "ocr_documents": [
                        {
                            "filename": "滴滴行程单.png",
                            "summary": "滴滴出行电子行程单",
                            "text": "滴滴出行 实付 13.4 元",
                            "avg_score": 0.94,
                            "preview_kind": "image",
                            "preview_data_url": "data:image/png;base64,ZmFrZQ==",
                            "warnings": [],
                        },
                    ],
                },
                tool_payload={"draft_only": True},
            )
        )

        assert response.review_payload is not None
        assert response.review_payload.document_cards[0].preview_kind == "image"
        assert response.review_payload.document_cards[0].preview_data_url.startswith("data:image/png;base64,")


def test_user_agent_prompts_existing_draft_association_choice_for_multi_documents() -> None:
    session_factory = build_session_factory()
    with session_factory() as db:
        ontology = SemanticOntologyService(db).parse(
            OntologyParseRequest(
                query="我上传了两张票据，帮我生成报销草稿",
                user_id="pytest",
            )
        )

        response = UserAgentService(db).respond(
            UserAgentRequest(
                run_id=ontology.run_id,
                user_id="pytest",
                message="我上传了两张票据，帮我生成报销草稿",
                ontology=ontology,
                context_json={
                    "attachment_names": ["滴滴行程单.png", "餐饮发票.jpg"],
                    "attachment_count": 2,
                    "ocr_documents": [
                        {"filename": "滴滴行程单.png", "summary": "滴滴出行 金额 32 元", "text": "滴滴出行 金额 32 元"},
                        {"filename": "餐饮发票.jpg", "summary": "餐饮发票 金额 68 元", "text": "餐饮发票 金额 68 元"},
                    ],
                },
                tool_payload={
                    "pending_association_decision": True,
                    "association_candidate_claim_no": "EXP-202605-008",
                },
            )
        )

        assert response.review_payload is not None
        assert response.review_payload.can_proceed is False
        assert [item.action_type for item in response.review_payload.confirmation_actions] == [
            "cancel_review",
            "edit_review",
            "link_to_existing_draft",
            "create_new_claim_from_documents",
        ]
        assert "EXP-202605-008" in response.answer