chore: [workflow] 提交工作区代码

tests/modules/shop_intelligence/test_douyin_price_appeal.py

@@ -496,12 +496,14 @@ class _ActionFingerprintLocator:
         row_key: str | None,
         action_text: str,
         previous_text: str,
+        product_scope_text: str = "",
     ) -> None:
         self.payload = {
             "log_extra": log_extra,
             "row_key": row_key,
             "action_text": action_text,
             "previous_text": previous_text,
+            "product_scope_text": product_scope_text,
         }
 
     def evaluate(self, script: str) -> dict[str, str | None]:
@@ -1187,6 +1189,29 @@ def test_action_fingerprint_falls_back_to_previous_product_row_and_row_key() ->
     assert "<countdown>" in fingerprint
 
 
+def test_action_fingerprint_falls_back_to_nested_product_card_context() -> None:
+    locator = _ActionFingerprintLocator(
+        log_extra=json.dumps(
+            {
+                "sub_activity_id": "7610636843016552714",
+                "record_id": "7617013918167138569",
+            }
+        ),
+        row_key="7617013918167138569",
+        action_text="报名异常 待改价 20:22:31 立即改价",
+        previous_text="",
+        product_scope_text=(
+            "GYXX/光影行星星背包男双肩包2026新款旅行大... "
+            "类目：箱包-时尚箱包-双肩包 ID: 3757136031745441873"
+        ),
+    )
+
+    fingerprint = appeal._action_fingerprint(locator)  # noqa: SLF001
+
+    assert appeal._extract_item_id(fingerprint) == "3757136031745441873"  # noqa: SLF001
+    assert "record_id: 7617013918167138569" in fingerprint
+
+
 def test_same_product_in_multiple_activities_has_distinct_fingerprints() -> None:
     product_id = "3774363265803616739"
     first = _ActionFingerprintLocator(

tests/modules/shop_intelligence/test_dy_after_sales_quality.py

@@ -0,0 +1,381 @@
+from __future__ import annotations
+
+import argparse
+from datetime import date, datetime
+from pathlib import Path
+
+import pytest
+
+from gyxx_flow.modules.shop_intelligence.collectors import (
+    dy_after_sales_quality as collector,
+)
+
+_LIST_ROW_REFUND = """订单编号 6929344525347945945 售后编号 147664497007198455 退货退款 平台主动处理
+GYXX/光影行星阿波罗x1斜挎包男款单肩包电脑包新款2026爆款
+逆光白/加大号(适合16寸游戏本电脑)
+商家编码： 10348008
+小件运费险 7天 15天售后
+应付金额 ¥369 ◎
+购买件数 1件
+售后退款 ¥369 ◎
+申请件数 1件
+申请原因 商品品质不好
+申请时间 2026/09/14 21:26:01
+同意退款，退款成功
+钱款去向 >
+未介入
+买家退货 运输中 >
+订单发货 1/1已签收 >
+查看详情"""
+
+_LIST_HEADERS = [
+    "",
+    "商品信息",
+    "订单信息",
+    "售后信息",
+    "售后状态",
+    "纠纷仲裁",
+    "物流信息",
+    "操作",
+]
+
+_PRODUCT_CELL = """GYXX/光影行星阿波罗x1斜挎包男款单肩包电脑包新款2026爆款
+逆光白/加大号(适合16寸游戏本电脑)
+商家编码： 10348008
+小件运费险 7天 15天售后"""
+
+_STATUS_CELL = """同意退款，退款成功
+钱款去向 >"""
+
+_NEGOTIATION_PANEL = """协商记录
+时间排序↑
+消费者 发起了退货退款申请，等待商家处理
+2026-09-07 21:59:55
+操作人 消费者
+售后类型 退货退款
+货物状态 已收到货
+退货原因 商品品质不好
+退款金额 ¥29.9
+退款说明 不值
+退货方式 上门取件（2026-09-08 18:00 - 20:00）
+凭证信息
+系统 因该用户信誉良好，收到商品后请及时处理售后，系统自动同意退货申请，等待消费者退货
+2026-09-07 22:00:00
+操作人 系统
+退货地址 广东省惠州市惠城区水口街道惠州市惠城区湖溪大道一号 光影行星地球工作舱3楼（抖仓库）
+系统 预约上门取件
+2026-09-07 22:00:02
+操作人 系统
+预约时间 2026-09-08 18:00:00 至 2026-09-08 20:00:00"""
+
+_DETAIL_PRODUCT_BLOCK = """GYXX/光影行星星干湿分离袋 洗漱包 收纳袋
+ID: 3791074652722343083 颜色分类: 黑色 商家编码: 10399079
+小件运费险 7天 15天售后"""
+
+_DETAIL_AFTERSALE_BLOCK = """售后信息
+退货退款 平台主动处理
+举报异常售后 售后历史
+订单编号 6929012177139563577 售后编号 1476543384326189817
+下单时间 2026-08-28 10:14:18 申请金额 ¥29.9 ◎ 消费者应退金额 ¥29.9
+申请时间 2026-09-07 21:59:55
+售后原因 商品品质不好（商品品质原因）
+售后说明 不值
+消费者凭证
+收货状态 已收到货
+申请件数 1件
+消费者需回寄数量 1件
+展开全部"""
+
+
+def _args(**kwargs: object) -> argparse.Namespace:
+    base: dict[str, object] = {
+        "date": None,
+        "from_date": None,
+        "to_date": None,
+    }
+    base.update(kwargs)
+    return argparse.Namespace(**base)
+
+
+def test_parse_total_count() -> None:
+    assert collector.parse_total_count("共161条 | 批量同意退款") == 161
+    assert collector.parse_total_count("共 22 条") == 22
+
+    with pytest.raises(RuntimeError, match="售后单总数"):
+        collector.parse_total_count("暂无数据")
+
+
+def test_parse_list_row_extracts_refund_fields_from_full_text() -> None:
+    row = collector.parse_list_row(_LIST_ROW_REFUND)
+
+    assert row["aftersale_id"] == "147664497007198455"
+    assert row["order_id"] == "6929344525347945945"
+    assert row["aftersale_type"] == "退货退款"
+    assert row["status_tag"] == "平台主动处理"
+    assert row["aftersale_status"] == "同意退款，退款成功"
+    assert row["apply_at"] == "2026/09/14 21:26:01"
+    assert row["apply_reason"] == "商品品质不好"
+    assert row["apply_quantity"] == "1"
+    assert row["pay_amount"] == "369"
+    assert row["refund_amount"] == "369"
+    assert row["product_name"] == (
+        "GYXX/光影行星阿波罗x1斜挎包男款单肩包电脑包新款2026爆款"
+    )
+    assert row["product_code"] == "10348008"
+
+
+def test_parse_list_row_uses_cell_mapping_for_product_and_status() -> None:
+    cells = {
+        "商品信息": _PRODUCT_CELL,
+        "售后状态": _STATUS_CELL,
+    }
+
+    row = collector.parse_list_row(_LIST_ROW_REFUND, cells)
+
+    assert row["product_name"] == (
+        "GYXX/光影行星阿波罗x1斜挎包男款单肩包电脑包新款2026爆款"
+    )
+    assert row["spec_name"] == "逆光白/加大号(适合16寸游戏本电脑)"
+    assert row["product_code"] == "10348008"
+    assert row["aftersale_status"] == "同意退款，退款成功"
+
+
+def test_map_row_cells_requires_matching_lengths() -> None:
+    mapping = collector.map_row_cells(_LIST_HEADERS, ["", "a", "b", "c", "d", "e", "f", "g"])
+    assert mapping["商品信息"] == "a"
+    assert mapping["操作"] == "g"
+    assert collector.map_row_cells(_LIST_HEADERS, ["only"]) == {}
+
+
+def test_parse_list_row_returns_empty_without_aftersale_id() -> None:
+    row = collector.parse_list_row("无关文本")
+    assert row["aftersale_id"] is None
+
+
+def test_parse_negotiation_entries_extracts_reason_note_and_evidence() -> None:
+    entries = collector.parse_negotiation_entries(
+        _NEGOTIATION_PANEL,
+        {0: ["https://obj.example/evidence-1.jpg"]},
+    )
+
+    assert [entry["entry_seq"] for entry in entries] == [0, 1, 2]
+    first = entries[0]
+    assert first["actor_role"] == "消费者"
+    assert first["action_text"] == "发起了退货退款申请，等待商家处理"
+    assert first["happened_at"] == datetime(2026, 9, 7, 21, 59, 55, tzinfo=collector.SHANGHAI)
+    assert first["operator"] == "消费者"
+    assert first["return_reason"] == "商品品质不好"
+    assert first["refund_note"] == "不值"
+    assert first["evidence_urls"] == ["https://obj.example/evidence-1.jpg"]
+    assert first["extra_fields"]["售后类型"] == "退货退款"
+    assert first["extra_fields"]["货物状态"] == "已收到货"
+    assert first["extra_fields"]["退款金额"] == "¥29.9"
+    assert first["extra_fields"]["退货方式"].startswith("上门取件")
+
+    second = entries[1]
+    assert second["actor_role"] == "系统"
+    assert second["action_text"].startswith("因该用户信誉良好")
+    assert second["return_reason"] is None
+    assert second["extra_fields"]["退货地址"].endswith("（抖仓库）")
+    assert second["evidence_urls"] == []
+
+    third = entries[2]
+    assert third["action_text"] == "预约上门取件"
+    assert third["extra_fields"]["预约时间"] == (
+        "2026-09-08 18:00:00 至 2026-09-08 20:00:00"
+    )
+
+
+def test_parse_negotiation_entries_handles_block_label_values() -> None:
+    text = """消费者 发起了退货退款申请，等待商家处理
+2026-09-07 21:59:55
+操作人 消费者
+退货原因
+商品品质不好
+凭证信息"""
+
+    entries = collector.parse_negotiation_entries(text)
+
+    assert len(entries) == 1
+    assert entries[0]["operator"] == "消费者"
+    assert entries[0]["return_reason"] == "商品品质不好"
+
+
+def test_parse_detail_product_block_extracts_id_and_title() -> None:
+    product = collector.parse_detail_product_block(_DETAIL_PRODUCT_BLOCK)
+
+    assert product["product_name"] == "GYXX/光影行星星干湿分离袋 洗漱包 收纳袋"
+    assert product["product_id"] == "3791074652722343083"
+    assert product["spec_name"] == "黑色"
+    assert product["product_code"] == "10399079"
+
+
+def test_parse_detail_aftersale_block_parses_inline_and_block_fields() -> None:
+    fields = collector.parse_detail_aftersale_block(_DETAIL_AFTERSALE_BLOCK)
+
+    assert fields["订单编号"] == "6929012177139563577"
+    assert fields["售后编号"] == "1476543384326189817"
+    assert fields["下单时间"] == "2026-08-28 10:14:18"
+    assert fields["申请金额"] == "¥29.9"
+    assert fields["消费者应退金额"] == "¥29.9"
+    assert fields["申请时间"] == "2026-09-07 21:59:55"
+    assert fields["售后原因"] == "商品品质不好（商品品质原因）"
+    assert fields["售后说明"] == "不值"
+    assert fields["收货状态"] == "已收到货"
+    assert fields["申请件数"] == "1件"
+    assert fields["消费者需回寄数量"] == "1件"
+
+
+def test_parse_apply_time_supports_list_and_detail_formats() -> None:
+    parsed = collector.parse_apply_time("2026/09/14 21:26:01")
+    assert parsed is not None
+    assert parsed.isoformat() == "2026-09-14T21:26:01+08:00"
+
+    parsed = collector.parse_apply_time("2026-09-07 21:59:55")
+    assert parsed is not None
+    assert parsed.isoformat() == "2026-09-07T21:59:55+08:00"
+
+    assert collector.parse_apply_time("unknown") is None
+    assert collector.parse_apply_time(None) is None
+
+
+def test_resolve_query_range_defaults_to_full_mode() -> None:
+    assert collector._resolve_query_range(_args()) == (None, None)
+
+
+def test_resolve_query_range_single_day_and_env_range(
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    single = collector._resolve_query_range(_args(date=date(2026, 9, 20)))
+    assert single == (date(2026, 9, 20), date(2026, 9, 20))
+
+    monkeypatch.setenv("GYXX_DY_AFTER_SALES_TO", "2026-09-21")
+    ranged = collector._resolve_query_range(_args(date=date(2026, 9, 1)))
+    assert ranged == (date(2026, 9, 1), date(2026, 9, 21))
+
+    monkeypatch.setenv("GYXX_DY_AFTER_SALES_TO", "2026-08-31")
+    with pytest.raises(ValueError, match="不能早于"):
+        collector._resolve_query_range(_args(date=date(2026, 9, 1)))
+
+    monkeypatch.setenv("GYXX_DY_AFTER_SALES_TO", "bad-date")
+    with pytest.raises(ValueError, match="YYYY-MM-DD"):
+        collector._resolve_query_range(_args(date=date(2026, 9, 1)))
+
+
+def test_resolve_query_range_from_to_arguments() -> None:
+    ranged = collector._resolve_query_range(
+        _args(from_date=date(2026, 3, 1), to_date=date(2026, 9, 1))
+    )
+    assert ranged == (date(2026, 3, 1), date(2026, 9, 1))
+
+    with pytest.raises(ValueError, match="同时提供"):
+        collector._resolve_query_range(_args(from_date=date(2026, 3, 1)))
+    with pytest.raises(ValueError, match="不能早于"):
+        collector._resolve_query_range(
+            _args(from_date=date(2026, 9, 1), to_date=date(2026, 3, 1))
+        )
+
+
+def test_normalize_order_builds_order_and_negotiation_records() -> None:
+    row = collector.parse_list_row(
+        _LIST_ROW_REFUND,
+        {"商品信息": _PRODUCT_CELL, "售后状态": _STATUS_CELL},
+    )
+    detail = {
+        "entries": collector.parse_negotiation_entries(
+            _NEGOTIATION_PANEL, {0: ["https://obj.example/evidence-1.jpg"]}
+        ),
+        "product": collector.parse_detail_product_block(_DETAIL_PRODUCT_BLOCK),
+        "aftersale_fields": collector.parse_detail_aftersale_block(
+            _DETAIL_AFTERSALE_BLOCK
+        ),
+        "panel_text": _NEGOTIATION_PANEL,
+        "product_text": _DETAIL_PRODUCT_BLOCK,
+        "detail_url": "https://fxg.jinritemai.com/ffa/merchant-aftersale-workbench/aftersale/detail",
+    }
+    captured_at = datetime(2026, 9, 21, 10, 0, 0, tzinfo=collector.SHANGHAI)
+
+    order, negotiations = collector.normalize_order(
+        row,
+        detail,
+        account="douyin-shop",
+        source_url=collector.TARGET_URL,
+        captured_at=captured_at,
+    )
+
+    assert order["account"] == "douyin-shop"
+    assert order["aftersale_id"] == "147664497007198455"
+    assert order["order_id"] == "6929344525347945945"
+    assert order["business_date"].isoformat() == "2026-09-07"
+    assert order["apply_at"] == datetime(2026, 9, 7, 21, 59, 55, tzinfo=collector.SHANGHAI)
+    assert order["apply_reason"] == "商品品质不好（商品品质原因）"
+    assert order["product_id"] == "3791074652722343083"
+    # 详情页交易信息优先于列表行；夹具中详情块来自另一条售后单，
+    # 因此商家编码/商品名以详情块为准。
+    assert order["product_code"] == "10399079"
+    assert order["product_name"] == "GYXX/光影行星星干湿分离袋 洗漱包 收纳袋"
+    assert order["negotiation_count"] == 3
+    assert order["detail_url"].endswith("/aftersale/detail")
+    assert len(negotiations) == 3
+    assert negotiations[0]["aftersale_id"] == "147664497007198455"
+    assert negotiations[0]["return_reason"] == "商品品质不好"
+    assert negotiations[0]["refund_note"] == "不值"
+    assert negotiations[0]["captured_at"] == captured_at
+
+
+def test_normalize_order_requires_aftersale_id() -> None:
+    with pytest.raises(RuntimeError, match="售后编号"):
+        collector.normalize_order(
+            {"aftersale_id": None, "raw_text": "broken"},
+            {"entries": [], "product": {}, "aftersale_fields": {}},
+            account="douyin-shop",
+            source_url=collector.TARGET_URL,
+            captured_at=datetime(2026, 9, 21, tzinfo=collector.SHANGHAI),
+        )
+
+
+def test_normalize_order_requires_apply_time() -> None:
+    with pytest.raises(RuntimeError, match="申请时间"):
+        collector.normalize_order(
+            collector.parse_list_row("订单编号 6929344525347945945 售后编号 147664497007198455"),
+            {"entries": [], "product": {}, "aftersale_fields": {}},
+            account="douyin-shop",
+            source_url=collector.TARGET_URL,
+            captured_at=datetime(2026, 9, 21, tzinfo=collector.SHANGHAI),
+        )
+
+
+def test_save_progress_round_trip(tmp_path: Path) -> None:
+    output = tmp_path / "抖音品退_全量.json"
+    collector.save_progress(
+        output,
+        account="douyin-shop",
+        query_start=None,
+        query_end=None,
+        expected_count=2,
+        collected={"147664497007198455": {"negotiation_count": 3}},
+        raw_rows={"147664497007198455": {"list_row": "text"}},
+        complete=False,
+        failed={"147664497007198456": "timeout"},
+    )
+
+    saved = collector.load_progress(output)
+    assert saved is not None
+    assert saved["complete"] is False
+    assert saved["query_mode"] == "full"
+    assert saved["expected_count"] == 2
+    assert saved["records"]["147664497007198455"]["negotiation_count"] == 3
+    assert saved["failed_records"] == {"147664497007198456": "timeout"}
+
+
+def test_after_sales_quality_schema_has_tables_and_idempotent_keys() -> None:
+    schema = Path(collector.__file__).parents[1] / "db" / "schema.sql"
+    text = schema.read_text(encoding="utf-8")
+
+    assert "dy_after_sales_quality_orders" in text
+    assert "dy_after_sales_quality_negotiations" in text
+    assert "CONSTRAINT dy_after_sales_quality_order_uniq UNIQUE (account, aftersale_id)" in text
+    assert (
+        "CONSTRAINT dy_after_sales_quality_negotiation_uniq\n"
+        "        UNIQUE (account, aftersale_id, entry_seq)" in text
+    )

tests/modules/shop_intelligence/test_dy_customer_service_messages.py

@@ -0,0 +1,584 @@
+from __future__ import annotations
+
+import json
+from datetime import date
+from pathlib import Path
+
+import pytest
+
+from gyxx_flow.modules.shop_intelligence.collectors import (
+    dy_customer_service_messages as collector,
+)
+
+
+def _row() -> dict[str, object]:
+    return {
+        "conversation_id": "76865431791638131343",
+        "session_time_label": "09/18 00:33-00:49",
+        "customer_nickname": "go",
+        "consult_entry": "直播间",
+        "reception_type": "纯人工",
+        "initiator_scene": "消费者发起咨询（纯人工/智能转人工/纯智能）",
+        "end_scene": "消费者-留言关闭会话",
+        "rating_time_label": "-",
+        "rating_result": "未评价",
+        "rating_tags": "-",
+        "conversation_intent": "-",
+        "smart_reception_label": "-",
+        "smart_rating_time_label": "-",
+        "smart_rating_result": "未评价",
+        "smart_rating_tags": "-",
+        "transfer_human_duration": "00:34:27",
+        "raw_cells": ["ID: 76865431791638131343\n09/18 00:33-00:49", "go"],
+    }
+
+
+def _messages() -> list[dict[str, object]]:
+    return [
+        {
+            "sender_role": "customer",
+            "sender_name": None,
+            "content": "盖亚，有没有？",
+            "message_type": "text",
+            "card": {},
+            "time_label": "09/18 00:33",
+            "message_at": None,
+            "raw_message": {"text": "盖亚，有没有？", "time_label": "09/18 00:33"},
+        },
+        {
+            "sender_role": "service",
+            "sender_name": None,
+            "content": "有！点45号盖亚双肩3，曜夜黑款能装1机2镜",
+            "message_type": "text",
+            "card": {},
+            "time_label": "09/18 00:33",
+            "message_at": None,
+            "raw_message": {"text": "有！点45号盖亚双肩3", "time_label": "09/18 00:33"},
+        },
+        {
+            "sender_role": "service",
+            "sender_name": None,
+            "content": "GYXX/光影行星盖亚双肩3相机包男士摄影\n¥684.00 券后价 已售14",
+            "message_type": "product_card",
+            "card": {
+                "kind": "product_card",
+                "title": "GYXX/光影行星盖亚双肩3相机包男士摄影",
+                "price": "684.00",
+                "text": "GYXX/光影行星盖亚双肩3相机包男士摄影\n¥684.00 券后价 已售14",
+            },
+            "time_label": None,
+            "message_at": None,
+            "raw_message": {"text": "product card"},
+        },
+    ]
+
+
+def test_parse_session_time_label_anchors_to_business_date() -> None:
+    start, end = collector.parse_session_time_label(
+        "09/18 00:33-00:49", date(2026, 9, 18)
+    )
+
+    assert start is not None and end is not None
+    assert start.isoformat() == "2026-09-18T00:33:00+08:00"
+    assert end.isoformat() == "2026-09-18T00:49:00+08:00"
+
+
+def test_parse_session_time_label_rolls_back_previous_year() -> None:
+    start, end = collector.parse_session_time_label(
+        "12/31 23:50-00:10", date(2026, 1, 1)
+    )
+
+    assert start is not None and end is not None
+    assert start.isoformat() == "2025-12-31T23:50:00+08:00"
+    assert end.isoformat() == "2026-01-01T00:10:00+08:00"
+
+
+def test_parse_session_time_label_extends_past_midnight() -> None:
+    start, end = collector.parse_session_time_label(
+        "09/18 23:50-00:10", date(2026, 9, 19)
+    )
+
+    assert start is not None and end is not None
+    assert end > start
+    assert end.isoformat() == "2026-09-19T00:10:00+08:00"
+
+
+def test_parse_total_count_and_page_size() -> None:
+    assert collector.parse_total_count("共22条数据") == 22
+    assert collector.parse_page_size("20 条/页") == 20
+
+    with pytest.raises(RuntimeError, match="会话总数"):
+        collector.parse_total_count("暂无数据")
+
+
+def test_resolve_reception_types_defaults_to_human_and_transfer() -> None:
+    assert collector._resolve_reception_types(
+        "纯人工客服接待,智能客服转人工接待"
+    ) == ("纯人工客服接待", "智能客服转人工接待")
+
+    with pytest.raises(ValueError, match="不支持的接待类型"):
+        collector._resolve_reception_types("纯人工客服接待,全部")
+
+
+def test_normalize_conversation_keeps_business_and_session_times() -> None:
+    result = collector.normalize_conversation(
+        _row(),
+        {
+            "consult_entry": "直播间",
+            "rating_result": "未评价",
+            "service_problem_tags": None,
+            "unmet_metrics": None,
+        },
+        _messages(),
+        account="douyin-shop",
+        business_date=date(2026, 9, 18),
+        source_url=collector.TARGET_URL,
+    )
+
+    conversation = result["conversation"]
+    assert result["message_count"] == 3
+    assert conversation["business_date"] == date(2026, 9, 18)
+    assert conversation["conversation_id"] == "76865431791638131343"
+    assert conversation["customer_nickname"] == "go"
+    assert conversation["session_start_at"].isoformat() == "2026-09-18T00:33:00+08:00"
+    assert conversation["session_end_at"].isoformat() == "2026-09-18T00:49:00+08:00"
+    assert conversation["reception_type"] == "纯人工"
+    assert conversation["transfer_human_duration"] == "00:34:27"
+    messages = result["messages"]
+    assert [item["sender_role"] for item in messages] == [
+        "customer",
+        "service",
+        "service",
+    ]
+    assert [item["message_seq"] for item in messages] == [0, 1, 2]
+    assert messages[2]["message_type"] == "product_card"
+    assert messages[2]["card"]["price"] == "684.00"
+    assert len({item["message_key"] for item in messages}) == 3
+
+
+def test_normalize_conversation_fails_closed_without_messages() -> None:
+    with pytest.raises(RuntimeError, match="未采集到任何消息"):
+        collector.normalize_conversation(
+            _row(),
+            {},
+            [],
+            account="douyin-shop",
+            business_date=date(2026, 9, 18),
+            source_url=collector.TARGET_URL,
+        )
+
+
+def test_save_progress_marks_incomplete_until_all_conversations_collected(
+    tmp_path: Path,
+) -> None:
+    output = tmp_path / "抖音客服消息_2026-09-18.json"
+    collector.save_progress(
+        output,
+        account="douyin-shop",
+        business_date=date(2026, 9, 18),
+        reception_types=("纯人工客服接待", "智能客服转人工接待"),
+        expected_count=2,
+        collected={"76865431791638131343": {"message_count": 3}},
+        raw_rows={"76865431791638131343": _row()},
+        complete=False,
+    )
+
+    saved = json.loads(output.read_text(encoding="utf-8"))
+    assert saved["complete"] is False
+    assert saved["expected_conversation_count"] == 2
+    assert saved["collected_conversation_count"] == 1
+    assert saved["reception_types"] == ["纯人工客服接待", "智能客服转人工接待"]
+    assert saved["raw_rows"]["76865431791638131343"]["reception_type"] == "纯人工"
+    assert saved["failed_conversations"] == {}
+
+
+class _FakePage:
+    """模拟虚拟滚动表格：视口相交的行才挂载，滚动容器由脚本 JS 驱动。"""
+
+    row_height = 60.0
+    client_height = 240.0
+
+    def __init__(self, rows: list[dict[str, object]], batch: int = 2) -> None:
+        self._rows = rows
+        self._batch = batch
+        self._top = 0.0
+        self.evaluate_calls = 0
+
+    @property
+    def _scroll_height(self) -> float:
+        return max(len(self._rows) * self.row_height, 300.0)
+
+    def _mounted_rows(self) -> list[dict[str, object]]:
+        start = int(self._top // self.row_height)
+        end = -(-int(self._top + self.client_height) // int(self.row_height))
+        return self._rows[start : max(end, start + self._batch)]
+
+    def evaluate(self, script: str, *args: object) -> object:
+        self.evaluate_calls += 1
+        if "qa-history-conversation-item" not in script:
+            return None
+        if "scrollTop" not in script:
+            # read_current_page_rows 的读取脚本。
+            return {
+                "headers": ["会话信息", "用户信息", "接待类型"],
+                "rows": [
+                    {
+                        "key": row["conversation_id"],
+                        "cells": [
+                            f"ID: {row['conversation_id']}\n09/18 00:33-00:49",
+                            str(row.get("customer_nickname", "")),
+                            str(row.get("reception_type", "")),
+                        ],
+                    }
+                    for row in self._mounted_rows()
+                ],
+            }
+        if "found:" in script:
+            return {
+                "found": 1,
+                "top": self._top,
+                "height": self._scroll_height,
+                "client": self.client_height,
+            }
+        if "scrollTop = Math.min" in script:
+            step = float(args[0]) if args else 200.0
+            max_top = max(self._scroll_height - self.client_height, 0.0)
+            self._top = min(self._top + step, max_top)
+            return None
+        if "scrollTop = 0" in script:
+            self._top = 0.0
+            return None
+        return None
+
+    def wait_for_timeout(self, _ms: int) -> None:  # pragma: no cover - 同步桩
+        return None
+
+
+def test_scan_current_page_rows_walks_virtual_scrolled_table() -> None:
+    rows = [
+        {
+            "conversation_id": str(7686000000000000000 + index),
+            "customer_nickname": f"u{index}",
+        }
+        for index in range(6)
+    ]
+    page = _FakePage(rows, batch=2)
+
+    scanned = collector.scan_current_page_rows(page, max_steps=20)
+
+    assert [row["conversation_id"] for row in scanned] == [
+        row["conversation_id"] for row in rows
+    ]
+    # 扫描完成后滚动容器应回到顶部。
+    assert page._top == 0.0
+
+
+def test_scan_current_page_rows_without_virtual_scroll_reads_once() -> None:
+    class _PlainPage(_FakePage):
+        def evaluate(self, script: str, *args: object) -> object:  # noqa: D102
+            if "found:" in script:
+                return {"found": 0}
+            return super().evaluate(script, *args)
+
+    rows = [
+        {
+            "conversation_id": str(7686000000000000000 + index),
+            "customer_nickname": f"u{index}",
+        }
+        for index in range(3)
+    ]
+    page = _PlainPage(rows, batch=10)
+
+    scanned = collector.scan_current_page_rows(page)
+
+    assert len(scanned) == 3
+
+
+_ORDER_CARD_FULL = """已完成
+(售后完成)
+咨询订单
++0
+6929386983188954550
+耳机包;换货(11535611)#;换货物流公司:中通速递物流单号:76979169163382;补发物流公司:中通速递物流单号:76980427930406;(收) 10507005*1#
+修改
+承诺服务工单
+[承诺同意换货申请]
+已完结
+GYXX/光影行星曜影双肩男女背包通勤旅游轻便出行大学生电脑包
+[已发1/1]
+运费险
+7天
+极速退
+小店自卖
+直播
++2
+规格
+【标准版】曜夜黑
+编码
+10507011
+实付金额
+¥
+414.
+00
+(含运费 ¥ 0.00)
+优惠 ¥ 15.00
+付款时间
+2026/09/11 11:12:03 (抖音月付)
+物流信息
+中通快递 76977842019514
+[已签收]2026-09-13 21:01:46
+收货信息
+陈*，1***
+，北京市北京市通州区梨园镇***
+发货时间
+2026-09-11 14:27"""
+
+_ORDER_CARD_PLAIN = """已发货
+咨询订单
++0
+6929558440707981209
+GYXX/光影行星阿波罗2斜挎包男款单肩包男生电脑背包新款2026爆款
+[已发1/1]
+运费险
+7天
+极速退
+小店自卖
+商品卡
++2
+规格
+曜夜黑 加大
+编码
+10442024
+实付金额
+¥
+379.
+00
+(含运费 ¥ 0.00)
+付款时间
+2026/09/17 16:35:15 (微信)
+物流信息
+中通快递 76980363689790
+[待取件]2026-09-19 09:24:37
+收货信息
+[已改地址]
+许*，1***
+，上海市上海市闵行区吴泾镇***
+发货时间
+2026-09-17 17:29"""
+
+
+def test_parse_order_card_extracts_product_and_payment() -> None:
+    parsed = collector.parse_order_card(_ORDER_CARD_FULL)
+
+    assert parsed is not None
+    assert parsed["order_id"] == "6929386983188954550"
+    assert parsed["order_status"] == "已完成"
+    assert parsed["after_sale_status"] == "售后完成"
+    assert parsed["order_badge"] == "咨询订单"
+    assert parsed["product_title"] == (
+        "GYXX/光影行星曜影双肩男女背包通勤旅游轻便出行大学生电脑包"
+    )
+    assert parsed["product_code"] == "10507011"
+    assert parsed["spec"] == "【标准版】曜夜黑"
+    assert parsed["paid_amount"] == "414.00"
+    assert parsed["pay_channel"] == "抖音月付"
+    assert parsed["paid_at"] is not None
+    assert parsed["paid_at"].isoformat() == "2026-09-11T11:12:03+08:00"
+    assert parsed["discount_text"] == "15.00"
+    assert parsed["freight_text"] == "0.00"
+    assert parsed["logistics_company"] == "中通快递"
+    assert parsed["logistics_no"] == "76977842019514"
+    assert parsed["logistics_status"] == "已签收 2026-09-13 21:01:46"
+    assert parsed["receiver_info"] is not None
+    assert "北京市" in parsed["receiver_info"]
+    assert parsed["shipped_at"] is not None
+    assert parsed["shipped_at"].isoformat() == "2026-09-11T14:27:00+08:00"
+    assert "运费险" in (parsed["service_tags"] or "")
+    assert parsed["remark_text"] is not None
+    assert "换货" in parsed["remark_text"]
+
+
+def test_parse_order_card_handles_plain_shipped_variant() -> None:
+    parsed = collector.parse_order_card(_ORDER_CARD_PLAIN)
+
+    assert parsed is not None
+    assert parsed["order_id"] == "6929558440707981209"
+    assert parsed["order_status"] == "已发货"
+    assert parsed["after_sale_status"] is None
+    assert parsed["product_title"].startswith("GYXX/光影行星阿波罗2斜挎包")
+    assert parsed["product_code"] == "10442024"
+    assert parsed["spec"] == "曜夜黑 加大"
+
+
+def test_parse_order_card_keeps_code_and_spec_clean_with_trailing_badges() -> None:
+    card = """已发货
+6929461460816854023
+GYXX/光影行星逐星斜挎包男士单肩胸包
+规格
+逐影黑小号
+编码
+10411028
+退货退款
+售后关闭
+实付金额
+¥
+299.
+00
+付款时间
+2026/09/14 17:25:53 (微信)
+发货时间
+2026-09-15 11:18"""
+
+    parsed = collector.parse_order_card(card)
+
+    assert parsed is not None
+    assert parsed["product_code"] == "10411028"
+    assert parsed["spec"] == "逐影黑小号"
+    assert parsed["paid_amount"] == "299.00"
+    assert parsed["pay_channel"] == "微信"
+
+
+def test_parse_order_card_handles_card_without_sent_marker() -> None:
+    card = """已发货
+(售后失败)
+6929461460816854023
+GYXX/光影行星逐星斜挎包男士单肩胸包多功能时尚潮流骑行弓箭包
+规格
+逐影黑小号(适合7.9寸平板)
+编码
+10411028
+实付金额
+¥
+299.
+00
+付款时间
+2026/09/14 17:25:53 (微信)
+物流信息
+中通快递 76979187011548
+[已签收]2026-09-17 20:06:37
+发货时间
+2026-09-15 11:18"""
+
+    parsed = collector.parse_order_card(card)
+
+    assert parsed is not None
+    assert parsed["order_id"] == "6929461460816854023"
+    assert parsed["product_title"].startswith("GYXX/光影行星逐星斜挎包")
+    assert parsed["product_code"] == "10411028"
+    assert parsed["paid_amount"] == "299.00"
+    assert parsed["service_tags"] is None
+
+
+def test_parse_order_card_returns_none_without_order_id() -> None:
+    assert collector.parse_order_card("暂无相关订单") is None
+
+
+def test_parse_order_card_skips_status_line_before_title() -> None:
+    card = """待发货
+6929558440707981209
+待发货
+GYXX/光影行星阿波罗2斜挎包男款单肩包男生电脑背包新款2026爆款
+规格
+曜夜黑 加大
+编码
+10442024
+实付金额
+¥
+379.
+00
+付款时间
+2026/09/17 16:35:15 (微信)
+发货时间
+2026-09-17 17:29"""
+
+    parsed = collector.parse_order_card(card)
+
+    assert parsed is not None
+    assert parsed["product_title"].startswith("GYXX/光影行星阿波罗2斜挎包")
+    assert parsed["product_code"] == "10442024"
+
+
+def test_build_product_records_groups_by_code_with_variant_counts() -> None:
+    orders = [
+        {"product_code": "10442024", "product_title": "阿波罗2"},
+        {"product_code": "10442024", "product_title": "阿波罗2"},
+        {"product_code": "10442024", "product_title": "【直播】阿波罗2"},
+        {"product_code": "", "product_title": "无编码商品"},
+    ]
+
+    batch = collector.build_product_records(orders, date(2026, 9, 17))
+
+    assert set(batch) == {"10442024"}
+    entry = batch["10442024"]
+    assert entry["title_variants"] == {"阿波罗2": 2, "【直播】阿波罗2": 1}
+    assert entry["first_seen_date"] == entry["last_seen_date"] == date(2026, 9, 17)
+
+
+def test_merge_product_records_accumulates_variants_and_keeps_stable_title() -> None:
+    existing = [
+        {
+            "product_code": "10442024",
+            "product_title": "阿波罗2",
+            "first_seen_date": date(2026, 8, 1),
+            "last_seen_date": date(2026, 8, 20),
+            "title_variants": {"阿波罗2": 9},
+        }
+    ]
+    batch = collector.build_product_records(
+        [
+            {"product_code": "10442024", "product_title": "【直播】阿波罗2"},
+            {"product_code": "10442024", "product_title": "【直播】阿波罗2"},
+            {"product_code": "10507011", "product_title": "曜影双肩"},
+        ],
+        date(2026, 9, 17),
+    )
+
+    merged = collector.merge_product_records(existing, batch)
+    by_code = {row["product_code"]: row for row in merged}
+
+    assert set(by_code) == {"10442024", "10507011"}
+    apollo = by_code["10442024"]
+    assert apollo["product_title"] == "阿波罗2"
+    assert apollo["title_variants"] == {"阿波罗2": 9, "【直播】阿波罗2": 2}
+    assert apollo["first_seen_date"] == date(2026, 8, 1)
+    assert apollo["last_seen_date"] == date(2026, 9, 17)
+    assert by_code["10507011"]["product_title"] == "曜影双肩"
+
+
+def test_normalize_conversation_binds_orders_to_conversation() -> None:
+    result = collector.normalize_conversation(
+        _row(),
+        {},
+        _messages(),
+        account="douyin-shop",
+        business_date=date(2026, 9, 17),
+        source_url=collector.TARGET_URL,
+        order_cards=[_ORDER_CARD_FULL],
+    )
+
+    conversation = result["conversation"]
+    assert conversation["order_count"] == 1
+    assert "6929386983188954550" in (conversation["order_panel_text"] or "")
+    orders = result["orders"]
+    assert len(orders) == 1
+    order = orders[0]
+    assert order["account"] == "douyin-shop"
+    assert order["business_date"] == date(2026, 9, 17)
+    assert order["conversation_id"] == "76865431791638131343"
+    assert order["order_id"] == "6929386983188954550"
+    assert order["product_code"] == "10507011"
+    assert order["raw_card"]["text"].startswith("已完成")
+
+
+def test_customer_service_schema_has_douyin_tables_and_idempotent_keys() -> None:
+    schema = Path(collector.__file__).parents[1] / "db" / "schema.sql"
+    text = schema.read_text(encoding="utf-8")
+
+    assert "dy_customer_service_conversations" in text
+    assert "dy_customer_service_messages" in text
+    assert "dy_customer_service_conversation_orders" in text
+    assert "dy_customer_service_products" in text
+    assert "UNIQUE (account, business_date, conversation_id, order_id)" in text
+    assert "UNIQUE (account, business_date, conversation_id)" in text
+    assert "UNIQUE (account, business_date, conversation_id, message_key)" in text
+    assert "session_start_at         TIMESTAMPTZ" in text

tests/modules/shop_intelligence/test_dy_negative_reviews.py

@@ -0,0 +1,131 @@
+from __future__ import annotations
+
+import json
+from datetime import date, datetime
+from pathlib import Path
+from zoneinfo import ZoneInfo
+
+from gyxx_flow.modules.shop_intelligence.collectors import (
+    dy_negative_reviews as collector,
+)
+
+SHANGHAI = ZoneInfo("Asia/Shanghai")
+
+
+def _row() -> dict[str, object]:
+    return {
+        "order_id": "6929366182773751285",
+        "text": (
+            "订单ID:6929366182773751285\n"
+            "闻**.\n"
+            "差评\n"
+            "2026/09/13 19:04:49\n"
+            "发来又脏又皱 跟二手的一样 logo那都有异常凸起\n"
+            "更多\n"
+            "体验分不计入\n"
+            "展示中\n"
+            "GYXX/光影行星阿波罗2斜挎包男款单肩包男生电脑背包新款2026爆款\n"
+            "商品ID:3764111277610369503\n"
+            "星河银 中号\n"
+            "规格ID:3697688238906114\n"
+            "回复\n"
+            "订单详情\n"
+            "举报"
+        ),
+        "content": "发来又脏又皱 跟二手的一样 logo那都有异常凸起",
+        "rankTag": "差评",
+        "productName": "GYXX/光影行星阿波罗2斜挎包男款单肩包男生电脑背包新款2026爆款",
+        "productId": "商品ID:3764111277610369503",
+        "specName": "星河银 中号",
+        "images": [
+            "https://p26-product-comment-sign.byteimg.com/tos-cn-i-8y5sebgt5i/a.jpg"
+        ],
+        "status": "展示中",
+        "experienceExcluded": True,
+    }
+
+
+def test_parse_total_count_and_empty_detection() -> None:
+    assert collector.parse_total_count("共255条 批量回复") == 255
+
+    try:
+        collector.parse_total_count("评价管理")
+    except RuntimeError as exc:
+        assert "差评总数" in str(exc)
+    else:
+        raise AssertionError("missing total must raise")
+
+
+def test_normalize_review_extracts_text_images_and_product() -> None:
+    row = collector.normalize_review(
+        _row(),
+        account="douyin-shop",
+        business_date=date(2026, 9, 13),
+        source_url=collector.TARGET_URL,
+    )
+
+    assert row is not None
+    assert row["review_id"] == "6929366182773751285"
+    assert row["order_id"] == "6929366182773751285"
+    assert row["rating_level"] == "差评"
+    assert row["stars"] == 1
+    assert row["review_at"] == datetime(
+        2026, 9, 13, 19, 4, 49, tzinfo=SHANGHAI
+    )
+    assert row["content"].startswith("发来又脏又皱")
+    assert row["image_urls"] == [
+        "https://p26-product-comment-sign.byteimg.com/tos-cn-i-8y5sebgt5i/a.jpg"
+    ]
+    assert row["display_status"] == "展示中"
+    assert row["experience_score_excluded"] is True
+    assert row["product_id"] == "3764111277610369503"
+    assert row["product_name"].startswith("GYXX/光影行星阿波罗2斜挎包")
+    assert row["spec_name"] == "星河银 中号"
+    assert row["spec_id"] == "3697688238906114"
+    assert row["raw_review"]["text"].startswith("订单ID:")
+
+
+def test_normalize_review_returns_none_without_order_id() -> None:
+    row = dict(_row())
+    row["order_id"] = ""
+    row["text"] = "与商品无关的文本"
+
+    assert collector.normalize_review(
+        row,
+        account="douyin-shop",
+        business_date=date(2026, 9, 13),
+        source_url=collector.TARGET_URL,
+    ) is None
+
+
+def test_save_progress_marks_incomplete_until_all_reviews_collected(
+    tmp_path: Path,
+) -> None:
+    output = tmp_path / "抖音差评_2026-09-13.json"
+    collector.save_progress(
+        output,
+        account="douyin-shop",
+        business_date=date(2026, 9, 13),
+        expected_count=2,
+        collected={"6929366182773751285": {"content": "demo"}},
+        complete=False,
+    )
+
+    saved = json.loads(output.read_text(encoding="utf-8"))
+    assert saved["complete"] is False
+    assert saved["expected_count"] == 2
+    assert saved["collected_count"] == 1
+
+    restored = collector.load_progress(output)
+    assert restored is not None
+    assert list(restored["reviews"]) == ["6929366182773751285"]
+
+
+def test_negative_review_schema_has_idempotency_and_product_columns() -> None:
+    schema = Path(collector.__file__).parents[1] / "db" / "schema.sql"
+    text = schema.read_text(encoding="utf-8")
+
+    assert "dy_negative_reviews" in text
+    assert "UNIQUE (account, business_date, review_id)" in text
+    assert "product_id                TEXT" in text
+    assert "image_urls                JSONB NOT NULL DEFAULT '[]'::jsonb" in text

tests/modules/shop_intelligence/test_jd_after_sales_issues.py

@@ -0,0 +1,169 @@
+from __future__ import annotations
+
+import json
+from datetime import date
+from pathlib import Path
+
+import pytest
+
+from gyxx_flow.modules.shop_intelligence.collectors import (
+    jd_after_sales_issues as collector,
+)
+
+
+def _row(
+    service_id: str,
+    *,
+    user_original: str = "卡扣有一边插不进去",
+    product: str = "光影行星曜影双肩包 ID：10217847513327",
+) -> dict[str, str]:
+    return {
+        "服务单ID": service_id,
+        "订单ID": "351467676840",
+        "商品": product,
+        "一级品类": "箱包皮具",
+        "二级品类": "男包",
+        "三级品类": "男士单肩/斜挎包",
+        "申请时间": "2026-09-15 21:41:46",
+        "售后原因分类": "商品出现异常",
+        "售后一级原因": "商品问题",
+        "售后二级原因": "配件异常",
+        "用户期望处理方式": "退货",
+        "用户原声": user_original,
+        "操作": "服务单详情",
+    }
+
+
+def test_normalize_keeps_product_sku_and_filters_unknown_user_original() -> None:
+    rows, filtered_count = collector._normalize_issue_rows(
+        [_row("service-1"), _row("service-2", user_original="未知")],
+        account="光影行星研发",
+        business_date=date(2026, 9, 15),
+        source_url=collector.TARGET_URL,
+    )
+
+    assert filtered_count == 1
+    assert len(rows) == 1
+    row = rows[0]
+    assert row["service_id"] == "service-1"
+    assert row["product_name"] == "光影行星曜影双肩包"
+    assert row["sku_id"] == "10217847513327"
+    assert row["reason_category"] == "商品出现异常"
+    assert row["apply_at"].isoformat() == "2026-09-15T21:41:46+08:00"
+    assert row["user_original"] == "卡扣有一边插不进去"
+
+
+def test_normalize_deduplicates_identical_service_rows() -> None:
+    rows = collector.normalize_issue_rows(
+        [_row("service-1"), _row("service-1")],
+        account="光影行星研发",
+        business_date=date(2026, 9, 15),
+        source_url=collector.TARGET_URL,
+    )
+
+    assert [row["service_id"] for row in rows] == ["service-1"]
+
+
+def test_normalize_filters_non_target_reason() -> None:
+    row = _row("service-1")
+    row["售后原因分类"] = "7天无理由"
+
+    rows, filtered_count = collector._normalize_issue_rows(
+        [row],
+        account="光影行星研发",
+        business_date=date(2026, 9, 15),
+        source_url=collector.TARGET_URL,
+    )
+
+    assert rows == []
+    assert filtered_count == 0
+
+
+def test_custom_range_uses_application_date_as_daily_idempotency_date() -> None:
+    first = _row("service-1")
+    first["申请时间"] = "2026-06-02 08:15:00"
+    second = _row("service-2")
+    second["申请时间"] = "2026-09-15 21:41:46"
+    outside = _row("service-3")
+    outside["申请时间"] = "2026-09-16 00:00:00"
+
+    rows = collector.normalize_issue_rows(
+        [first, second, outside],
+        account="光影行星研发",
+        business_date=date(2026, 6, 1),
+        business_date_end=date(2026, 9, 15),
+        source_url=collector.TARGET_URL,
+    )
+
+    assert [(row["business_date"], row["service_id"]) for row in rows] == [
+        (date(2026, 6, 2), "service-1"),
+        (date(2026, 9, 15), "service-2"),
+    ]
+
+
+def test_custom_range_is_split_into_minimum_31_day_queries() -> None:
+    assert collector.split_custom_ranges(
+        date(2026, 6, 1), date(2026, 9, 15)
+    ) == (
+        (date(2026, 6, 1), date(2026, 7, 1)),
+        (date(2026, 7, 2), date(2026, 8, 1)),
+        (date(2026, 8, 2), date(2026, 9, 1)),
+        (date(2026, 9, 2), date(2026, 9, 15)),
+    )
+
+
+def test_normalize_rejects_conflicting_duplicate_service_rows() -> None:
+    first = _row("service-1")
+    second = _row("service-1", user_original="另一条用户原声")
+
+    with pytest.raises(RuntimeError, match="冲突服务单 ID"):
+        collector.normalize_issue_rows(
+            [first, second],
+            account="光影行星研发",
+            business_date=date(2026, 9, 15),
+            source_url=collector.TARGET_URL,
+        )
+
+
+def test_progress_contains_filter_counts_and_no_password(tmp_path: Path) -> None:
+    output = tmp_path / "品退问题_2026-09-15.json"
+    rows = collector.normalize_issue_rows(
+        [_row("service-1")],
+        account="光影行星研发",
+        business_date=date(2026, 9, 15),
+        source_url=collector.TARGET_URL,
+    )
+    collector.save_progress(
+        output,
+        account="光影行星研发",
+        business_date=date(2026, 9, 15),
+        raw_rows=[_row("service-1"), _row("service-2", user_original="未知")],
+        records=rows,
+        filtered_unknown_count=1,
+        downloaded_path=None,
+        db_rows=1,
+        complete=True,
+    )
+
+    saved = json.loads(output.read_text(encoding="utf-8"))
+    assert saved["complete"] is True
+    assert saved["filter"]["售后原因"] == "商品出现异常"
+    assert saved["raw_issue_count"] == 2
+    assert saved["valid_issue_count"] == 1
+    assert saved["filtered_unknown_user_original_count"] == 1
+    assert saved["db_upsert_count"] == 1
+    assert "aa123456" not in output.read_text(encoding="utf-8")
+
+
+def test_schema_and_upsert_use_daily_service_id_idempotency() -> None:
+    schema_path = Path(collector.__file__).parents[1] / "db" / "schema.sql"
+    db_path = Path(collector.__file__).parents[1] / "db" / "db.py"
+    schema = schema_path.read_text(encoding="utf-8")
+    db_source = db_path.read_text(encoding="utf-8")
+
+    assert "CREATE TABLE IF NOT EXISTS jd_after_sales_issues" in schema
+    assert "user_original         TEXT NOT NULL" in schema
+    assert "UNIQUE (account, business_date, service_id)" in schema
+    assert "def upsert_jd_after_sales_issues" in db_source
+    assert '"jd_after_sales_issues"' in db_source
+    assert '"account", "business_date", "service_id"' in db_source

tests/modules/shop_intelligence/test_jd_customer_service_messages.py

@@ -0,0 +1,197 @@
+from __future__ import annotations
+
+import json
+from datetime import date
+from pathlib import Path
+
+import pytest
+
+from gyxx_flow.modules.shop_intelligence.collectors import (
+    jd_customer_service_messages as collector,
+)
+
+
+def _payload() -> dict[str, object]:
+    return {
+        "code": "success",
+        "totalPage": 1,
+        "chatLogList": [
+            {
+                "chatLogMessageList": [
+                    {
+                        "sid": "session-1",
+                        "uuid": "message-1",
+                        "customer": "customer-demo",
+                        "created": "2026-09-10 10:01:02",
+                        "closeSessionTime": "2026-09-10 10:10:00",
+                        "content": "请问这个商品有现货吗？",
+                        "type": "text",
+                        "waiter": "示例客服",
+                        "waiterSend": 0,
+                        "imgUrl": "",
+                    },
+                    {
+                        "sid": "session-1",
+                        "uuid": "message-2",
+                        "customer": "customer-demo",
+                        "created": "2026-09-10 10:02:03",
+                        "closeSessionTime": "2026-09-10 10:10:00",
+                        "content": "有现货，今天可以安排发出。",
+                        "type": "text",
+                        "waiter": "示例客服",
+                        "waiterSend": 1,
+                        "imgUrl": "",
+                    },
+                ]
+            }
+        ],
+    }
+
+
+def test_normalize_chat_payload_keeps_business_and_message_times() -> None:
+    result = collector.normalize_chat_payload(
+        _payload(),
+        account="demo-shop",
+        business_date=date(2026, 9, 10),
+        customer_id="customer-demo",
+        source_url=collector.TARGET_URL,
+    )
+
+    assert result["session_count"] == 1
+    assert result["message_count"] == 2
+    conversation = result["conversations"][0]
+    assert conversation["business_date"] == date(2026, 9, 10)
+    assert conversation["session_id"] == "session-1"
+    assert conversation["session_start_at"].isoformat() == "2026-09-10T10:01:02+08:00"
+    messages = result["messages"]
+    assert messages[0]["sender_role"] == "customer"
+    assert messages[0]["waiter_send"] is False
+    assert messages[1]["sender_role"] == "waiter"
+    assert messages[1]["waiter_send"] is True
+    assert messages[1]["message_at"].isoformat() == "2026-09-10T10:02:03+08:00"
+
+
+def test_normalize_chat_payload_fails_closed_on_unhandled_message_paging() -> None:
+    payload = _payload()
+    payload["totalPage"] = 2
+
+    with pytest.raises(RuntimeError, match="拒绝截断入库"):
+        collector.normalize_chat_payload(
+            payload,
+            account="demo-shop",
+            business_date=date(2026, 9, 10),
+            customer_id="customer-demo",
+            source_url=collector.TARGET_URL,
+        )
+
+
+def test_normalize_chat_payload_filters_templates_and_stickers_but_keeps_real_media() -> None:
+    payload = _payload()
+    payload["chatLogList"][0]["chatLogMessageList"].extend(
+        [
+            {
+                "sid": "session-1",
+                "uuid": "message-template",
+                "created": "2026-09-10 10:03:00",
+                "content": "【此消息为欢迎卡片或富文本模板答案，聊天记录中暂不支持展示】",
+                "type": "template2",
+                "waiter": "示例客服",
+                "waiterSend": 1,
+                "imgUrl": "",
+            },
+            {
+                "sid": "session-1",
+                "uuid": "message-sticker-image",
+                "created": "2026-09-10 10:04:00",
+                "content": "",
+                "type": "image",
+                "waiter": "示例客服",
+                "waiterSend": 1,
+                "imgUrl": "https://img11.360buyimg.com/stickermanage/demo.png",
+            },
+            {
+                "sid": "session-1",
+                "uuid": "message-sticker-file",
+                "created": "2026-09-10 10:05:00",
+                "content": "file:///C:/Users/Public/Documents/AliWorkbench/emotions/MyEmotion/demo.png",
+                "type": "text",
+                "waiter": "示例客服",
+                "waiterSend": 1,
+                "imgUrl": "",
+            },
+            {
+                "sid": "session-1",
+                "uuid": "message-real-image",
+                "created": "2026-09-10 10:06:00",
+                "content": "",
+                "type": "image",
+                "waiter": "示例客服",
+                "waiterSend": 1,
+                "imgUrl": "https://img10.360buyimg.com/img/jfs/demo.jpg",
+            },
+        ]
+    )
+
+    result = collector.normalize_chat_payload(
+        payload,
+        account="demo-shop",
+        business_date=date(2026, 9, 10),
+        customer_id="customer-demo",
+        source_url=collector.TARGET_URL,
+    )
+
+    assert result["session_count"] == 1
+    assert result["message_count"] == 3
+    assert result["filtered_message_count"] == 3
+    assert [message["message_type"] for message in result["messages"]] == [
+        "text",
+        "text",
+        "image",
+    ]
+    assert result["messages"][-1]["image_url"].endswith("demo.jpg")
+    assert result["conversations"][0]["message_count"] == 3
+    assert len(result["conversations"][0]["raw_session"]["chatLogMessageList"]) == 3
+
+
+def test_save_progress_marks_unresolved_customers_and_keeps_raw_response(
+    tmp_path: Path,
+) -> None:
+    output = tmp_path / "客服消息_2026-09-10.json"
+    collector.save_progress(
+        output,
+        account="demo-shop",
+        business_date=date(2026, 9, 10),
+        expected_customers=["customer-demo", "customer-pending"],
+        collected={"customer-demo": {"message_count": 2}},
+        raw_payloads={"customer-demo": _payload()},
+        complete=False,
+    )
+
+    saved = json.loads(output.read_text(encoding="utf-8"))
+    assert saved["store"] == "flagship"
+    assert saved["complete"] is False
+    assert saved["unresolved_customers"] == ["customer-pending"]
+    assert saved["raw_payloads"]["customer-demo"]["chatLogList"]
+
+
+def test_customer_service_store_routes_to_the_self_operated_identity() -> None:
+    store, config = collector._customer_service_store("self-operated")
+
+    assert store == "self-operated"
+    assert config["account_id"] == "jd-self-operated"
+    assert collector._configured_identity_markers(
+        {
+            "GYXX_JD_CUSTOMER_SERVICE_IDENTITY_MARKERS": '["光影自营售前"]',
+        }
+    ) == ("光影自营售前",)
+
+
+def test_customer_service_schema_has_date_and_actual_message_time_columns() -> None:
+    schema = Path(collector.__file__).parents[1] / "db" / "schema.sql"
+    text = schema.read_text(encoding="utf-8")
+
+    assert "jd_customer_service_conversations" in text
+    assert "jd_customer_service_messages" in text
+    assert "business_date     DATE NOT NULL" in text
+    assert "message_at          TIMESTAMPTZ" in text
+    assert "UNIQUE (account, business_date, customer_id, session_id, message_key)" in text

tests/modules/shop_intelligence/test_jd_negative_reviews.py

@@ -0,0 +1,99 @@
+from __future__ import annotations
+
+import json
+from datetime import date
+from pathlib import Path
+
+from gyxx_flow.modules.shop_intelligence.collectors import (
+    jd_negative_reviews as collector,
+)
+
+
+def _payload() -> dict[str, object]:
+    return {
+        "code": 200,
+        "msg": "成功",
+        "data": {
+            "total": "1",
+            "pageIndex": 1,
+            "pageSize": 100,
+            "results": [
+                {
+                    "commentId": "review-1",
+                    "commentLevel": 1,
+                    "score": 1,
+                    "createDate": "2026-09-12 10:01:02",
+                    "nickName": "demo-buyer",
+                    "content": "商品与描述不一致",
+                    "orderId": "order-1",
+                    "skuId": "sku-1",
+                    "skuName": "光影行星示例商品 黑色",
+                    "productColor": "黑色",
+                    "productSize": "标准",
+                    "productImageUrl": "jfs/example.jpg",
+                    "images": [{"imgUrl": "jfs/review.jpg"}],
+                    "isVenderReply": False,
+                    "replyCount": 0,
+                    "secPin": "should-not-be-persisted",
+                }
+            ],
+        },
+    }
+
+
+def test_normalize_negative_review_keeps_product_fields_and_date() -> None:
+    rows = collector.normalize_review_results(
+        _payload(),
+        account="demo-shop",
+        business_date=date(2026, 9, 12),
+        source_url=collector.TARGET_URL,
+    )
+
+    assert len(rows) == 1
+    row = rows[0]
+    assert row["business_date"] == date(2026, 9, 12)
+    assert row["review_id"] == "review-1"
+    assert row["review_level"] == "差评"
+    assert row["content"] == "商品与描述不一致"
+    assert row["product_name"] == "光影行星示例商品 黑色"
+    assert row["sku_id"] == "sku-1"
+    assert row["review_at"].isoformat() == "2026-09-12T10:01:02+08:00"
+    assert row["is_vender_reply"] is False
+    assert "secPin" not in row["raw_review"]
+
+
+def test_save_progress_keeps_page_count_and_redacts_sensitive_fields(tmp_path: Path) -> None:
+    output = tmp_path / "差评_2026-09-12.json"
+    collector.save_progress(
+        output,
+        account="demo-shop",
+        store="flagship",
+        business_date=date(2026, 9, 12),
+        total=1,
+        raw_pages=[collector._sanitize_review_payload(_payload())],
+        collected={"review-1": {"sku_id": "sku-1"}},
+        complete=True,
+    )
+
+    saved = json.loads(output.read_text(encoding="utf-8"))
+    assert saved["complete"] is True
+    assert saved["collected_review_count"] == 1
+    assert saved["raw_pages"][0]["data"]["results"][0].get("secPin") is None
+
+
+def test_negative_review_schema_has_date_product_and_idempotency_columns() -> None:
+    schema = Path(collector.__file__).parents[1] / "db" / "schema.sql"
+    text = schema.read_text(encoding="utf-8")
+
+    assert "jd_negative_reviews" in text
+    assert "business_date       DATE NOT NULL" in text
+    assert "product_name        TEXT" in text
+    assert "sku_id              TEXT" in text
+    assert "UNIQUE (account, business_date, review_id)" in text
+
+
+def test_negative_review_store_reuses_customer_service_account_routing() -> None:
+    store, config = collector._store("self-operated")
+
+    assert store == "self-operated"
+    assert config["account_id"] == "jd-self-operated"

tests/modules/shop_intelligence/test_jd_old_datepicker_click_resilience.py

@@ -4,7 +4,9 @@ from datetime import date
 
 import pytest
 
-from gyxx_flow.modules.shop_intelligence.collectors import jd_data_collector as collector
+from gyxx_flow.modules.shop_intelligence.collectors import (
+    jd_data_collector as collector,
+)
 
 
 class _FlakyCellLocator:
@@ -104,3 +106,18 @@ def test_old_datepicker_happy_path_uses_playwright_click(
 
     assert page.cell.click_calls == 1
     assert page.native_click_script is None
+
+
+def test_merchant_ranking_skips_known_missing_module_permission(
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    monkeypatch.setattr(collector, "goto_with_retry", lambda *_args, **_kwargs: None)
+
+    class PermissionPage:
+        def wait_for_timeout(self, _milliseconds: int) -> None:
+            return None
+
+        def evaluate(self, _script: str):
+            return "当前身份下暂无此菜单\n请在顶部导航栏选择其他菜单。"
+
+    assert collector.step_merchant_ranking(PermissionPage(), "男包") == {}

tests/modules/shop_intelligence/test_source_business_logic_parity.py

@@ -139,15 +139,23 @@ TARGET_AST_SHA256_OVERRIDES = {
     ("jd_data_collector.py", "select_last_week_in_old_datepicker"): (
         "43d1056eeac1faeff32f0c01d0867e6673d2ab57999a14a3ebeb1b51c40075d4"
     ),
+    # Reviewed after the 2026-09-15 production run showed the account's
+    # merchant-ranking module is not entitled; skip that optional section while
+    # retaining the store-level metrics and marketing collection.
+    ("jd_data_collector.py", "step_merchant_ranking"): (
+        "b1cb2893f6e3a2378eca0e889980a06ffd2dcccc71b8c7714be6714958b274dd"
+    ),
     ("jd_data_collector.py", "step_shop_star_and_trade"): (
         "77aa873543861f1fad63d0bd1ace60528ab20592b0bffef3d46ad192f2103161"
     ),
-    ("jd_data_collector.py", "step_merchant_ranking"): (
-        "187cd3cfffcb97a5bf9f1ed4dde0d117ee65f82c1cf431df4c7c9c9ee2a7b282"
-    ),
     ("jd_data_collector.py", "step_jzt_non_full_site"): (
         "1bd53eba5d0123e63d2e628d9772c0efa57fc8d2f87d79a4f3c41a311132952d"
     ),
+    # Reviewed after the 2026-09-15 production run showed FXG had moved from
+    # the old "我的体验分" labels to the new "我的总分/商品体验" labels.
+    ("dy_store_competitor_store_scraping.py", "_extract_experience_score"): (
+        "c2eb1513c9aa510e9cf2f2c6baa155b7d6538c038ce5ff4782caed32850221c6"
+    ),
 }
 
 # These two source functions were reviewed after a production run proved that