chore: [workflow] 提交工作区代码

2026-09-30 10:46:15 +08:00
1 parent 2ff588a65d
commit 01f3d7aaf7
@@ -674,3 +674,52 @@ def test_wait_for_exported_excel_returns_immediately_when_present(monkeypatch):
    )

    assert chanmama._wait_for_exported_excel(0.0) == "/downloads/已有.xlsx"
+
+
+def test_wait_for_exported_excel_pumps_playwright_events(monkeypatch):
+    """回归: 纯 sleep 轮询不泵 Playwright 事件会让 download 事件排队，
+    download.save_as 到 quit 才执行，文件错过 deadline 误判超时。
+    等待期间必须周期性调用 driver 泵事件循环。"""
+
+    class _FakeDriver:
+        def __init__(self) -> None:
+            self.eval_calls = 0
+
+        def execute_script(self, script, *args):
+            self.eval_calls += 1
+            return None
+
+    now = [1000.0]
+    results = iter([None, None, "/downloads/泵事件后落盘.xlsx"])
+    driver = _FakeDriver()
+
+    monkeypatch.setattr(chanmama.time, "time", lambda: now[0])
+    monkeypatch.setattr(
+        chanmama.time,
+        "sleep",
+        lambda seconds: now.__setitem__(0, now[0] + seconds),
+    )
+    monkeypatch.setattr(chanmama, "_find_new_excel_after", lambda _pre_mtime: next(results))
+
+    assert chanmama._wait_for_exported_excel(0.0, driver=driver) == (
+        "/downloads/泵事件后落盘.xlsx"
+    )
+    assert driver.eval_calls >= 3
+
+
+def test_wait_for_exported_excel_tolerates_driver_errors(monkeypatch):
+    """泵事件失败(页面导航中等)不应中断文件轮询。"""
+
+    class _BrokenDriver:
+        def execute_script(self, script, *args):
+            raise RuntimeError("page navigating")
+
+    monkeypatch.setattr(
+        chanmama,
+        "_find_new_excel_after",
+        lambda _pre_mtime: "/downloads/已有.xlsx",
+    )
+
+    assert chanmama._wait_for_exported_excel(0.0, driver=_BrokenDriver()) == (
+        "/downloads/已有.xlsx"
+    )
@@ -358,22 +358,32 @@ def test_scheduled_collection_has_no_automatic_retry_and_syncs_once():
    )


-def test_daily_collection_includes_self_operated_pipeline():
+def test_daily_collection_separates_self_operated_pipeline():
    steps = _workflow("content.metrics.daily").steps
    entries = [step.entry for step in steps]
    run_step = next(step for step in steps if step.entry == "run_all.py")

    assert "--daily-scope" in run_step.args
    assert "--include-self-operated" not in run_step.args
-    assert entries.index("run_all.py") < entries.index(
-        "data/tools/refresh_self_mapping.py"
-    )
-    assert entries.index("data/tools/refresh_self_mapping.py") < entries.index(
-        "self_bilibili_scraper.py"
-    )
-    assert entries.index("self_bilibili_scraper.py") < entries.index(
-        "chanmama_scraper.py"
+    assert entries == ["run_all.py", "data/tools/sync_metrics_to_cmt_notes.py"]
+    assert steps[1].args == ("--scope", "collaborators")
+
+    self_steps = _workflow("content.metrics.self_operated.daily").steps
+    self_entries = [step.entry for step in self_steps]
+    assert self_entries == [
+        "data/tools/refresh_self_mapping.py",
+        "self_bilibili_scraper.py",
+        "chanmama_scraper.py",
+        "self_xhs_scraper.py",
+        "data/tools/sync_metrics_to_cmt_notes.py",
+        "data/tools/sync_self_video_metrics.py",
+    ]
+    sync_step = next(
+        step
+        for step in self_steps
+        if step.entry == "data/tools/sync_metrics_to_cmt_notes.py"
    )
+    assert sync_step.args == ("--scope", "self-operated")


@pytest.mark.parametrize(
@@ -0,0 +1,661 @@
+# -*- coding: utf-8 -*-
+"""自营内容全量指标(抖音/B站/小红书): 解析、账号映射、入库行构造的边界测试。"""
+from __future__ import annotations
+
+import datetime as dt
+import json
+import os
+import re
+from pathlib import Path
+
+import openpyxl
+
+from gyxx_flow.modules.content_marketing import (
+    chanmama_export,
+    self_operated_accounts,
+    self_xhs_scraper,
+)
+from gyxx_flow.modules.content_marketing.data.tools import (
+    sync_self_video_metrics as sync_tool,
+)
+from gyxx_flow.modules.content_marketing.runtime_paths import PATHS
+
+PROJECT_ROOT = Path(__file__).resolve().parents[3]
+TOOLS = (
+    PROJECT_ROOT / "src/gyxx_flow/modules/content_marketing/data/tools"
+)
+
+EXPORT_HEADERS = [
+    "视频名称", "蝉妈妈视频链接", "抖音视频链接", "达人昵称", "抖音号",
+    "视频时长", "发布时间", "预估曝光", "点赞数", "评论数", "转发数",
+    "预估销量", "预估销量-指数", "预估销售额", "预估销售额-指数",
+]
+
+STORE_FLAGSHIP = "光影行星箱包旗舰店"
+STORE_SPECIALTY = "GYXX光影行星箱包专卖店"
+XHS_SHOP = "光影行星GYXX旗舰店"
+
+
+def _write_export_xlsx(path: Path, rows: list[tuple]) -> str:
+    workbook = openpyxl.Workbook()
+    sheet = workbook.active
+    sheet.append(EXPORT_HEADERS)
+    for row in rows:
+        sheet.append(list(row))
+    workbook.save(path)
+    workbook.close()
+    return str(path)
+
+
+def _douyin_sample_rows() -> list[tuple]:
+    return [
+        ("总计", "-", "-", "-", "-", "-", "-", "-", "-", "-", "-", "0", "-", "0", "-"),
+        (
+            "轻户外，自己定义-光影行星 逐星",
+            "https://www.chanmama.com/awemeDetail/SbXQnFsF2tlZF5Wy7U1uG3izRQaBOwYT",
+            "https://www.iesdouyin.com/share/video/7687797200985869595/?region=CN",
+            STORE_FLAGSHIP, "gyxx168888", "1分0秒", "2026-09-21 09:39",
+            "3.2w+", 135, 8, 2, "12", 0, "3580.5", 0,
+        ),
+        (
+            "图文笔记标题",
+            "https://www.chanmama.com/awemeDetail/M7_SEJOk9P_Q1eafatlOCAQl4F2eI3FR",
+            "https://www.iesdouyin.com/share/note/7687432318584143857/?region=CN",
+            STORE_SPECIALTY, "GYXX168168", "--", "2026-09-20 10:03",
+            "0", 1, 0, 0, "0", 0, "0", 0,
+        ),
+    ]
+
+
+def test_parse_chanmama_excel_extracts_all_metric_fields(tmp_path):
+    xlsx = _write_export_xlsx(tmp_path / "flagship.xlsx", _douyin_sample_rows())
+
+    records, headers = chanmama_export.parse_chanmama_excel(xlsx)
+
+    assert headers == EXPORT_HEADERS
+    assert len(records) == 2  # 总计行被跳过
+    first = records[0]
+    assert first["creator_name"] == STORE_FLAGSHIP
+    assert first["exposure"] == 32000
+    assert first["like_count"] == 135
+    assert first["comment_count"] == 8
+    assert first["share_count"] == 2
+    assert first["est_sales_amount"] == 3580.5
+    assert chanmama_export.extract_video_id(
+        first["chanmama_url"], first["video_url"]
+    ) == "7687797200985869595"
+
+
+def test_self_operated_accounts_config_covers_three_platforms():
+    path = self_operated_accounts.accounts_config_path(PATHS.config_root)
+    all_accounts = self_operated_accounts.load_all_accounts(path)
+    assert set(all_accounts) == {"douyin", "bilibili", "xiaohongshu"}
+    # 用户约定的身份映射: 抖音·机能号=旗舰店, 抖音·高智号=专卖店
+    assert self_operated_accounts.resolve_alias(
+        all_accounts["douyin"], account_name=STORE_FLAGSHIP
+    ) == "抖音·机能号"
+    assert self_operated_accounts.resolve_alias(
+        all_accounts["douyin"], platform_account_id="GYXX168168"
+    ) == "抖音·高智号"
+    # 小红书账号必须带登录手机号
+    xhs = self_operated_accounts.find_account(
+        "xiaohongshu", path, account_name=XHS_SHOP
+    )
+    assert xhs is not None and re.fullmatch(r"1\d{10}", xhs["phone"])
+    assert self_operated_accounts.resolve_alias(
+        all_accounts["xiaohongshu"], account_name=XHS_SHOP
+    ) == "小红书·自营号"
+    # 未知账号返回空串,不抛异常
+    assert self_operated_accounts.resolve_alias(
+        all_accounts["bilibili"], account_name="未知UP主"
+    ) == ""
+
+
+def test_load_accounts_tolerates_missing_or_broken_config(tmp_path):
+    assert self_operated_accounts.load_accounts(
+        "douyin", tmp_path / "missing.json"
+    ) == []
+    broken = tmp_path / "broken.json"
+    broken.write_text("not-json", encoding="utf-8")
+    assert self_operated_accounts.load_accounts("douyin", broken) == []
+    empty = tmp_path / "empty.json"
+    empty.write_text("{}", encoding="utf-8")
+    assert self_operated_accounts.load_accounts("bilibili", empty) == []
+    with __import__("pytest").raises(ValueError):
+        self_operated_accounts.load_accounts("kuaishou", empty)
+
+
+# ============================================================
+# B站: 公开 API 全字段
+# ============================================================
+class _FakeBiliResponse:
+    def __init__(self, payload):
+        self._payload = payload
+
+    def json(self):
+        return self._payload
+
+
+class _FakeBiliSession:
+    """固定返回一个 B站 view API 响应的假 HTTP 会话。"""
+
+    PAYLOAD = {
+        "code": 0,
+        "data": {
+            "aid": 114514,
+            "bvid": "BV1xT4y1Q7zz",
+            "title": "光影行星 逐星 开箱",
+            "duration": 92,
+            "pubdate": 1761026400,
+            "owner": {"mid": 5201314, "name": "光影行星官方号"},
+            "stat": {
+                "view": 35821, "like": 1204, "coin": 88, "favorite": 301,
+                "share": 57, "danmaku": 145, "reply": 96,
+            },
+        },
+    }
+
+    def get(self, url, **_kwargs):
+        return _FakeBiliResponse(self.PAYLOAD)
+
+
+def test_fetch_bili_video_extracts_full_metrics():
+    from gyxx_flow.modules.content_marketing.self_bilibili_scraper import (
+        fetch_bili_video,
+        format_duration,
+    )
+
+    info = fetch_bili_video("https://www.bilibili.com/video/BV1xT4y1Q7zz",
+                            _FakeBiliSession())
+    assert info is not None
+    assert info["video_id"] == "BV1xT4y1Q7zz"
+    assert info["video_url"].endswith("/video/BV1xT4y1Q7zz")
+    assert info["title"] == "光影行星 逐星 开箱"
+    assert info["owner_name"] == "光影行星官方号"
+    assert info["view_count"] == 35821
+    assert info["like_count"] == 1204
+    assert info["coin_count"] == 88
+    assert info["favorite_count"] == 301
+    assert info["danmaku_count"] == 145
+    assert info["reply_count"] == 96
+    assert info["publish_time"] == dt.datetime.fromtimestamp(1761026400)
+    assert format_duration(92) == "1:32"
+    assert format_duration(3725) == "1:02:05"
+    assert format_duration(None) == ""
+
+
+def test_collect_bilibili_rows_from_artifact(tmp_path, monkeypatch):
+    artifact = tmp_path / "daily.json"
+    artifact.write_text(
+        json.dumps({
+            "platform": "bilibili",
+            "business_date": "2026-09-28",
+            "videos": [{
+                "video_id": "BV1xT4y1Q7zz",
+                "video_url": "https://www.bilibili.com/video/BV1xT4y1Q7zz",
+                "title": "光影行星 逐星 开箱",
+                "owner_name": "光影行星官方号",
+                "owner_mid": "5201314",
+                "duration_seconds": 92,
+                "publish_time": "2026-09-21 09:39",
+                "view_count": 35821, "like_count": 1204, "coin_count": 88,
+                "favorite_count": 301, "share_count": 57,
+                "danmaku_count": 145, "reply_count": 96,
+            }],
+        }, ensure_ascii=False),
+        encoding="utf-8",
+    )
+    monkeypatch.setattr(sync_tool, "BILIBILI_ARTIFACT", artifact)
+    accounts = [
+        {"alias": "B站·自营号", "account_name": "光影行星官方号",
+         "platform_account_id": "", "phone": ""}
+    ]
+
+    rows, stats, source = sync_tool.collect_bilibili_rows(accounts, dt.date(2026, 9, 28))
+
+    assert stats["rows"] == 1
+    assert source == artifact
+    row = rows[("bilibili", "光影行星官方号", "BV1xT4y1Q7zz")]
+    assert row[sync_tool.ROW_FIELDS.index("account_alias")] == "B站·自营号"
+    assert row[sync_tool.ROW_FIELDS.index("exposure")] == 35821
+    assert row[sync_tool.ROW_FIELDS.index("comment_count")] == 96  # reply→评论
+    assert row[sync_tool.ROW_FIELDS.index("coin_count")] == 88
+    assert row[sync_tool.ROW_FIELDS.index("duration")] == "1:32"
+
+
+def test_collect_bilibili_rows_missing_artifact(tmp_path, monkeypatch):
+    monkeypatch.setattr(
+        sync_tool, "BILIBILI_ARTIFACT", tmp_path / "missing.json"
+    )
+    rows, stats, source = sync_tool.collect_bilibili_rows([], dt.date(2026, 9, 28))
+    assert rows == {} and stats["rows"] == 0 and source is None
+
+
+# ============================================================
+# 小红书: ARK 管理页表格解析
+# ============================================================
+GOODS_HEADERS = ["", "笔记信息", "挂载商品", "笔记阅读数", "笔记支付金额",
+                 "笔记支付订单数", "发布状态", "操作"]
+NORMAL_HEADERS = ["", "笔记信息", "阅读数", "点赞数", "引流主页次数",
+                  "引流直播间次数", "发布状态", "操作"]
+
+
+def test_bilibili_wbi_signing_matches_documented_vector():
+    from gyxx_flow.modules.content_marketing.self_bilibili_scraper import (
+        derive_mixin_key,
+        wbi_sign,
+    )
+
+    # B站 WBI 公开示例密钥 → 文档给定的 mixin key
+    mixin = derive_mixin_key(
+        "7cd084941338484aae1ad9425b84077c",
+        "4932caff0ff746eab6f01bf08b70ac45",
+    )
+    assert mixin == "ea1db124af3c7062474693fa704f4ff8"
+
+    signed = wbi_sign(
+        {"platform": "web", "pn": 1, "ps": 30, "mid": "1008229967",
+         "order": "pubdate"},
+        mixin,
+        1702204800,
+    )
+    assert signed["wts"] == 1702204800
+    assert re.fullmatch(r"[0-9a-f]{32}", signed["w_rid"])
+    # 确定性: 同参数同时间戳签名一致
+    again = wbi_sign(
+        {"mid": "1008229967", "order": "pubdate", "pn": 1, "ps": 30,
+         "platform": "web"},
+        mixin,
+        1702204800,
+    )
+    assert again["w_rid"] == signed["w_rid"]
+
+
+def test_bilibili_artifact_covers_full_history(tmp_path, monkeypatch):
+    """主页全量采集后 artifact 应含全部投稿(含飞书未登记的历史视频)。"""
+    artifact = tmp_path / "daily.json"
+    videos = [
+        {
+            "video_id": f"BV1xT4y1Q7z{i:02d}",
+            "video_url": f"https://www.bilibili.com/video/BV1xT4y1Q7z{i:02d}",
+            "title": f"视频{i}",
+            "owner_name": "光影行星",
+            "owner_mid": "1008229967",
+            "duration_seconds": 60 + i,
+            "publish_time": f"2026-09-{i + 1:02d} 09:00:00",
+            "view_count": 100 * i, "like_count": i, "coin_count": i,
+            "favorite_count": i, "share_count": i, "danmaku_count": i,
+            "reply_count": i,
+        }
+        for i in range(3)
+    ]
+    artifact.write_text(
+        json.dumps({"platform": "bilibili", "business_date": "2026-09-29",
+                    "videos": videos}, ensure_ascii=False),
+        encoding="utf-8",
+    )
+    monkeypatch.setattr(sync_tool, "BILIBILI_ARTIFACT", artifact)
+    accounts = [
+        {"alias": "B站·自营号", "account_name": "光影行星",
+         "platform_account_id": "1008229967", "phone": ""}
+    ]
+
+    rows, stats, _source = sync_tool.collect_bilibili_rows(
+        accounts, dt.date(2026, 9, 29)
+    )
+
+    assert stats["rows"] == 3
+    key = ("bilibili", "光影行星", "BV1xT4y1Q7z00")
+    assert rows[key][sync_tool.ROW_FIELDS.index("account_alias")] == "B站·自营号"
+    assert rows[key][sync_tool.ROW_FIELDS.index("publish_time")] == dt.datetime(
+        2026, 9, 1, 9, 0
+    )
+
+
+def test_parse_info_cell_variants():
+    parse = self_xhs_scraper._parse_info_cell
+    # 多行,发布时间带前缀(线上真实形态)
+    assert parse("标题A\n发布时间：2026-09-15 14:23\n手动创作") == (
+        "标题A", "2026-09-15 14:23", "手动创作"
+    )
+    # 单行压扁(历史兜底)
+    title, publish, method = parse("标题B 发布时间：2026-09-08 13:41 手动创作")
+    assert title == "标题B"
+    assert publish == "2026-09-08 13:41"
+    assert method == "手动创作"
+    # 无日期无方式的纯标题
+    assert parse("纯标题") == ("纯标题", "", "")
+    assert parse("") == ("", "", "")
+
+
+def test_parse_goods_note_rows():
+    rows = [
+        {
+            "cells": [
+                "",
+                "斜挎包通勤好物\n2026-09-25 10:23\n手动创作",
+                "光影行星 极星单肩包男女…",
+                "1.2万",
+                "3,586.50",
+                "12",
+                "已发布",
+                "数据 编辑",
+            ],
+            "href": "https://www.xiaohongshu.com/explore/66e4b2c3000000001d00f1a2",
+        },
+        {
+            "cells": ["", "另一条\n2026-09-20 08:00", "暂无关联商品", "356",
+                      "—", "—", "已发布", "编辑"],
+            "href": "",
+        },
+    ]
+
+    records, stats = self_xhs_scraper.parse_management_rows(
+        GOODS_HEADERS, rows, "goods"
+    )
+
+    assert stats["records"] == 2
+    first = records[0]
+    assert first["note_type"] == "goods"
+    assert first["title"] == "斜挎包通勤好物"
+    assert first["publish_time"] == "2026-09-25 10:23"
+    assert first["creation_method"] == "手动创作"
+    assert first["mounted_products"] == "光影行星 极星单肩包男女…"
+    assert first["read_count"] == 12000
+    assert first["payment_amount"] == 3586.50
+    assert first["payment_orders"] == 12
+    assert first["publish_status"] == "已发布"
+    assert first["note_id"] == "66e4b2c3000000001d00f1a2"
+    assert first["video_id"] == "66e4b2c3000000001d00f1a2"
+    # 页面原始列全量保留
+    assert first["raw_columns"]["笔记阅读数"] == "1.2万"
+    # 无链接行回退标题指纹,并跨日稳定
+    second = records[1]
+    assert second["video_id"].startswith("xhs-")
+    assert second["video_id"] == self_xhs_scraper.fallback_note_fingerprint(
+        "另一条", "2026-09-20 08:00"
+    )
+    assert second["mounted_products"] == "暂无关联商品"
+    assert second["payment_amount"] is None
+    assert second["payment_orders"] is None
+
+
+def test_parse_normal_note_rows():
+    rows = [
+        {
+            "cells": ["", "双肩包种草\n2026-09-22 18:30\n可能是AI创作",
+                      "6,592", "483", "35", "8", "已发布", "数据"],
+            "href": "https://www.xiaohongshu.com/discovery/item/66e0abc5000000001900c3b1",
+        },
+    ]
+
+    records, stats = self_xhs_scraper.parse_management_rows(
+        NORMAL_HEADERS, rows, "normal"
+    )
+
+    assert stats["records"] == 1
+    record = records[0]
+    assert record["note_type"] == "normal"
+    assert record["read_count"] == 6592
+    assert record["like_count"] == 483
+    assert record["homepage_redirect_count"] == 35
+    assert record["live_redirect_count"] == 8
+    assert record["creation_method"] == "可能是AI创作"
+    assert record["video_id"] == "66e0abc5000000001900c3b1"
+    # 标题含「创作」字样不会被误判为创作方式标签
+    assert record["title"] == "双肩包种草"
+
+
+def test_parse_count_and_amount_edge_cases():
+    assert self_xhs_scraper.parse_count("1.2万") == 12000
+    assert self_xhs_scraper.parse_count("6,592") == 6592
+    assert self_xhs_scraper.parse_count("0") == 0
+    assert self_xhs_scraper.parse_count("—") is None
+    assert self_xhs_scraper.parse_count("暂无") is None
+    assert self_xhs_scraper.parse_count("") is None
+    assert self_xhs_scraper.parse_amount("3,586.50") == 3586.50
+    assert self_xhs_scraper.parse_amount("0.9万") == 9000.0
+    assert self_xhs_scraper.parse_amount("-") is None
+
+
+def test_collect_xiaohongshu_rows_from_artifact(tmp_path, monkeypatch):
+    artifact = tmp_path / "daily.json"
+    artifact.write_text(
+        json.dumps({
+            "platform": "xiaohongshu",
+            "business_date": "2026-09-28",
+            "account_name": XHS_SHOP,
+            "notes": [
+                {
+                    "note_type": "goods",
+                    "video_id": "66e4b2c3000000001d00f1a2",
+                    "title": "斜挎包通勤好物",
+                    "publish_time": "2026-09-25 10:23",
+                    "creation_method": "手动创作",
+                    "mounted_products": "光影行星 极星单肩包男女…",
+                    "read_count": 12000, "like_count": 431,
+                    "payment_amount": 3586.5, "payment_orders": 12,
+                    "publish_status": "已发布",
+                    "note_url": "https://www.xiaohongshu.com/explore/66e4b2c3000000001d00f1a2",
+                    "raw_columns": {"笔记阅读数": "1.2万"},
+                },
+            ],
+        }, ensure_ascii=False),
+        encoding="utf-8",
+    )
+    monkeypatch.setattr(sync_tool, "XHS_ARTIFACT", artifact)
+    accounts = [
+        {"alias": "小红书·自营号", "account_name": XHS_SHOP,
+         "platform_account_id": "18933246725", "phone": "18933246725"}
+    ]
+
+    rows, stats, source = sync_tool.collect_xiaohongshu_rows(
+        accounts, dt.date(2026, 9, 28)
+    )
+
+    assert stats["rows"] == 1
+    row = rows[("xiaohongshu", XHS_SHOP, "66e4b2c3000000001d00f1a2")]
+    fields = sync_tool.ROW_FIELDS
+    assert row[fields.index("account_alias")] == "小红书·自营号"
+    assert row[fields.index("exposure")] == 12000  # 阅读数 → 主触达
+    assert row[fields.index("like_count")] == 431
+    assert row[fields.index("payment_amount")] == 3586.5
+    assert row[fields.index("payment_orders")] == 12
+    assert row[fields.index("note_type")] == "goods"
+    assert row[fields.index("publish_time")] == dt.datetime(2026, 9, 25, 10, 23)
+    assert row[fields.index("page_metrics")].obj["笔记阅读数"] == "1.2万"
+
+
+def test_douyin_history_backfill_uses_file_mtime_and_filename_account(tmp_path):
+    """历史回填: 按导出落盘日期各留快照,旧版12列文件按文件名前缀补账号。"""
+    # 旧版 12 列表头(无 达人昵称/抖音号/蝉妈妈链接)
+    legacy_headers = [
+        "视频名称", "抖音视频链接", "视频时长", "发布时间", "预估曝光",
+        "点赞数", "评论数", "转发数", "预估销量", "预估销量-指数",
+        "预估销售额", "预估销售额-指数",
+    ]
+    legacy_file = tmp_path / "光影行星箱包旗舰店_视频记录_周期内发布视频_20260803-20260901.xlsx"
+    workbook = openpyxl.Workbook()
+    sheet = workbook.active
+    sheet.append(legacy_headers)
+    sheet.append([
+        "旧版格式视频", "https://www.iesdouyin.com/share/video/7686000000000000001/",
+        "46秒", "2026-08-27 16:40:56", "23000", 1321, 46, 14, "0", 0, "0", 0,
+    ])
+    workbook.save(legacy_file)
+    workbook.close()
+    import os
+
+    historical = dt.datetime(2026, 9, 2, 17, 0, 0).timestamp()
+    os.utime(legacy_file, (historical, historical))
+    accounts = self_operated_accounts.load_accounts(
+        "douyin", self_operated_accounts.accounts_config_path(PATHS.config_root)
+    )
+
+    rows, stats, _sources = sync_tool.collect_douyin_rows(
+        accounts, dt.date(2026, 9, 29), [str(legacy_file)],
+        use_file_mtime=True,
+    )
+
+    assert stats["skip_no_identity"] == 0
+    assert stats["rows"] == 1
+    row = next(iter(rows.values()))
+    fields = sync_tool.ROW_FIELDS
+    assert row[fields.index("account_name")] == STORE_FLAGSHIP  # 文件名前缀回退
+    assert row[fields.index("account_alias")] == "抖音·机能号"
+    assert row[fields.index("metric_date")] == dt.date(2026, 9, 2)  # mtime 日期
+    assert row[fields.index("exposure")] == 23000
+
+
+def test_collect_douyin_rows_keeps_video_note_type(tmp_path):
+    xlsx = _write_export_xlsx(tmp_path / "flagship.xlsx", _douyin_sample_rows())
+    accounts = self_operated_accounts.load_accounts(
+        "douyin", self_operated_accounts.accounts_config_path(PATHS.config_root)
+    )
+
+    rows, stats, _sources = sync_tool.collect_douyin_rows(
+        accounts, dt.date(2026, 9, 28), [xlsx]
+    )
+
+    assert stats["rows"] == 2
+    fields = sync_tool.ROW_FIELDS
+    video_row = rows[
+        ("douyin", STORE_FLAGSHIP, "7687797200985869595", dt.date(2026, 9, 28))
+    ]
+    note_row = rows[
+        ("douyin", STORE_SPECIALTY, "7687432318584143857", dt.date(2026, 9, 28))
+    ]
+    assert video_row[fields.index("note_type")] == "video"
+    assert note_row[fields.index("note_type")] == "note"
+    assert video_row[fields.index("account_alias")] == "抖音·机能号"
+
+
+def test_make_row_validates_fields_and_truncates():
+    row = sync_tool.make_row(
+        "douyin", dt.date(2026, 9, 28), "a.xlsx",
+        title="超长标题" * 200, video_id="1",
+    )
+    fields = sync_tool.ROW_FIELDS
+    assert len(row) == len(fields)
+    assert row[fields.index("title")] == ("超长标题" * 200)[:512]
+    assert row[fields.index("account_alias")] == ""
+    import pytest
+
+    with pytest.raises(KeyError):
+        sync_tool.make_row("douyin", dt.date(2026, 9, 28), "a.xlsx",
+                           unknown_field=1)
+
+
+# ============================================================
+# SQL 安全 / 表结构 / 配置登记
+# ============================================================
+def test_upsert_sql_is_fully_parameterized():
+    source = (TOOLS / "sync_self_video_metrics.py").read_text(encoding="utf-8")
+    matches = re.findall(
+        r"cur\.execute\(\s*\"\"\"\s*(INSERT\b.*?)\"\"\",\s*row,",
+        source,
+        re.DOTALL,
+    )
+    assert len(matches) == 1, "入库 upsert 必须以字面量 SQL + 参数元组形式调用"
+    sql = matches[0]
+    assert sql.count("%s") == len(sync_tool.ROW_FIELDS)  # 全部参数绑定
+    assert "{" not in sql and ".format(" not in sql
+    assert "ON CONFLICT (platform, account_name, video_id, metric_date)" in sql
+
+
+def test_schema_and_migrations_declare_self_video_metrics_consistently():
+    schema = (TOOLS / "schema_gyxx_super_data.sql").read_text(encoding="utf-8")
+    migration_009 = (TOOLS / "migrations" / "009_self_video_metrics.sql").read_text(
+        encoding="utf-8"
+    )
+    migration_010 = (
+        TOOLS / "migrations" / "010_self_video_metrics_platforms.sql"
+    ).read_text(encoding="utf-8")
+    for sql in (schema, migration_009, migration_010):
+        assert "cmt_self_video_metrics" in sql
+    for sql in (schema, migration_009):
+        assert "CREATE TABLE IF NOT EXISTS cmt_self_video_metrics" in sql
+        assert "UNIQUE (platform, account_name, video_id, metric_date)" in sql
+    combined = schema + "\n" + migration_009 + "\n" + migration_010
+    for column in (
+        "account_alias", "video_id", "video_url", "publish_time",
+        "exposure", "like_count", "comment_count", "share_count",
+        "est_sales_count", "est_sales_amount", "note_type",
+        "creation_method", "publish_status", "mounted_products",
+        "homepage_redirect_count", "live_redirect_count", "payment_amount",
+        "payment_orders", "coin_count", "favorite_count", "danmaku_count",
+        "page_metrics",
+    ):
+        assert re.search(rf"\b{column}\b", combined), f"{column} 缺失"
+
+
+def test_workflow_commands_and_bindings_declare_self_content_pipeline():
+    workflows = json.loads(
+        (PROJECT_ROOT / "config" / "workflows.json").read_text(encoding="utf-8")
+    )
+    workflow = next(
+        item
+        for item in workflows["workflows"]
+        if item["id"] == "content.metrics.self_operated.daily"
+    )
+    steps = {step["id"]: step for step in workflow["execution"]["steps"]}
+    assert set(steps) == {
+        "refresh_self_mapping", "collect_self_bilibili", "collect_self_douyin",
+        "collect_self_xiaohongshu", "sync", "sync_self_video_metrics",
+    }
+    assert steps["collect_self_xiaohongshu"]["entry"] == "self_xhs_scraper.py"
+    assert steps["collect_self_xiaohongshu"]["depends_on"] == ["collect_self_douyin"]
+    assert steps["sync_self_video_metrics"]["depends_on"] == [
+        "sync", "collect_self_xiaohongshu",
+    ]
+    module_root = PROJECT_ROOT / "src/gyxx_flow/modules/content_marketing"
+    for entry in (
+        "self_xhs_scraper.py",
+        "data/tools/sync_self_video_metrics.py",
+    ):
+        assert (module_root / entry).is_file()
+
+    commands = json.loads(
+        (PROJECT_ROOT / "config" / "commands.json").read_text(encoding="utf-8")
+    )
+    command_entries = {item["id"]: item for item in commands["commands"]}
+    assert command_entries["content.metrics.collect_self_xiaohongshu"]["entry"] == (
+        "self_xhs_scraper.py"
+    )
+    assert command_entries["content.metrics.sync_self_video_metrics"]["entry"] == (
+        "data/tools/sync_self_video_metrics.py"
+    )
+
+    bindings = json.loads(
+        (PROJECT_ROOT / "config" / "runtime-bindings.json").read_text(
+            encoding="utf-8"
+        )
+    )
+    xhs_binding = bindings["scripts"]["content.metrics.collect_self_xiaohongshu"]
+    assert xhs_binding["script_id"] == "content_marketing:self_xhs_scraper.py"
+    assert xhs_binding["required_cookie_domains"] == ["xiaohongshu.com"]
+    assert "XHS_ARK_ACCOUNT" in xhs_binding["credential_env_names"]
+    sync_binding = bindings["scripts"]["content.metrics.sync_self_video_metrics"]
+    assert sync_binding["cdp_port"] == 22164
+    ports = [entry["cdp_port"] for entry in bindings["scripts"].values()]
+    assert len(ports) == len(set(ports)), "CDP 端口必须唯一"
+
+
+def test_metric_date_resolution_and_stale_guard(tmp_path, monkeypatch):
+    monkeypatch.delenv("GYXX_BUSINESS_DATE", raising=False)
+    assert sync_tool.runtime_business_date() is None
+    monkeypatch.setenv("GYXX_BUSINESS_DATE", "2026-09-28")
+    assert sync_tool.runtime_business_date() == dt.date(2026, 9, 28)
+    assert sync_tool.resolve_metric_date("2026-09-01") == dt.date(2026, 9, 1)
+
+    xlsx = Path(_write_export_xlsx(tmp_path / "a.xlsx", _douyin_sample_rows()))
+    fresh = dt.datetime.now().timestamp()
+    os.utime(xlsx, (fresh, fresh))
+    assert sync_tool.is_current_artifact(xlsx, dt.date.today())
+    assert not sync_tool.is_current_artifact(xlsx, dt.date(2026, 9, 21))
+
+
+def test_runtime_business_date_rejects_invalid_value(monkeypatch):
+    monkeypatch.setenv("GYXX_BUSINESS_DATE", "2026/09/28")
+    with __import__("pytest").raises(RuntimeError):
+        sync_tool.runtime_business_date()
@@ -161,7 +161,7 @@ def test_erp_login_navigates_before_probing_initial_blank_page(
    action = erp_login.make_page_action(
        login_url="https://erp.test/login.aspx",
        username="user",
-        password="pw",
+        password="secret",
        home_url="https://erp.test/epaas",
        timeout_ms=1000,
        captcha_checks=0,
@@ -1,5 +1,6 @@
 from __future__ import annotations

+from datetime import date
 from pathlib import Path

 import collect_douyin_qianchuan_ads as collector
@@ -90,3 +91,128 @@ def test_report_ready_wait_guards_missing_body() -> None:
        "arg": list(collector._required_page_markers()),
        "timeout": 90_000,
    }
+
+
+def test_matching_task_rows_uses_scene_and_exact_date_range() -> None:
+    filters: list[str] = []
+
+    class _Locator:
+        def filter(self, *, has_text: str):
+            filters.append(has_text)
+            return self
+
+    class _Page:
+        def locator(self, selector: str) -> _Locator:
+            assert selector == "tr.ovui-tr"
+            return _Locator()
+
+    collector._matching_task_rows(
+        _Page(), date(2026, 9, 9), date(2026, 9, 9)
+    )
+
+    assert filters == [
+        collector.TASK_SCENE_MARKER,
+        "2026-09-09 00:00:00-2026-09-09 23:59:59",
+    ]
+
+
+def test_trigger_report_download_returns_direct_download() -> None:
+    expected_download = object()
+
+    class _Button:
+        def __init__(self, page) -> None:
+            self.page = page
+
+        def click(self) -> None:
+            self.page.listener(expected_download)
+
+    class _Page:
+        listener = None
+
+        def on(self, event: str, listener) -> None:
+            assert event == "download"
+            self.listener = listener
+
+        def remove_listener(self, event: str, listener) -> None:
+            assert event == "download"
+            assert listener == self.listener
+
+        def locator(self, selector: str):
+            assert selector == collector.DOWNLOAD_SELECTOR
+            return _Button(self)
+
+    assert collector._trigger_report_download(_Page()) is expected_download
+
+
+def test_trigger_report_download_detects_async_task_notification() -> None:
+    class _Locator:
+        def click(self) -> None:
+            pass
+
+        def inner_text(self, *, timeout: int) -> str:
+            assert timeout == 10_000
+            return "已生成下载任务，可前往任务中心查看"
+
+    class _Page:
+        def on(self, event: str, listener) -> None:
+            assert event == "download"
+
+        def remove_listener(self, event: str, listener) -> None:
+            assert event == "download"
+
+        def locator(self, selector: str) -> _Locator:
+            assert selector in {collector.DOWNLOAD_SELECTOR, "body"}
+            return _Locator()
+
+    assert collector._trigger_report_download(_Page()) is None
+
+
+def test_download_task_from_page_downloads_completed_matching_row() -> None:
+    expected_download = object()
+
+    class _DownloadInfo:
+        value = expected_download
+
+        def __enter__(self):
+            return self
+
+        def __exit__(self, exc_type, exc, traceback) -> None:
+            pass
+
+    class _DownloadButton:
+        def click(self) -> None:
+            pass
+
+    class _Row:
+        def inner_text(self) -> str:
+            return "全域推广数据-推商品\n生成完成\n下载"
+
+        def get_by_text(self, text: str, *, exact: bool) -> _DownloadButton:
+            assert (text, exact) == ("下载", True)
+            return _DownloadButton()
+
+    class _Rows:
+        first = _Row()
+
+        def count(self) -> int:
+            return 1
+
+    class _Page:
+        def expect_download(self, *, timeout: int) -> _DownloadInfo:
+            assert timeout == 60_000
+            return _DownloadInfo()
+
+    page = _Page()
+    original = collector._matching_task_rows
+    collector._matching_task_rows = lambda *args: _Rows()
+    try:
+        actual = collector._download_task_from_page(
+            page,
+            date(2026, 9, 9),
+            date(2026, 9, 9),
+            wait_timeout_ms=0,
+        )
+    finally:
+        collector._matching_task_rows = original
+
+    assert actual is expected_download
@@ -173,3 +173,173 @@ def test_safe_extract_zip_keeps_csv_inside_output(tmp_path: Path) -> None:
    files = collector._safe_extract_zip(archive, tmp_path / "extracted")

    assert files == [(tmp_path / "extracted" / "nested" / "full.csv").resolve()]
+
+
+FULL_REPORT_CSV = """查询条件,广告主PIN: 光影行星旗舰店霖; 时间: 2026-09-19至2026-09-19
+日期,'当前时间',SPU ID,商品计划名称,SpuImgUrl,花费
+20260919~20260919,当前时间,10032366483452,波塞冬托特包,img,43.68
+20260919~20260919,当前时间,10033333485894,宙斯双肩包,img,104.78
+汇总
+0,当前时间,0,,,148.46
+
+日期,'对比时间',SPU ID,商品计划名称,SpuImgUrl,花费
+20260918~20260918,对比时间,10032366483452,波塞冬托特包,img,40.00
+汇总
+0,对比时间,0,,,40.00
+"""
+
+NOT_READY_FULL_REPORT_CSV = """查询条件,广告主PIN: 光影行星旗舰店霖; 时间: 2026-09-20至2026-09-20
+日期,'当前时间',SPU ID,商品计划名称,SpuImgUrl,花费
+汇总
+0,当前时间,0,,,0.00,0.00,0.00,0,0.00
+
+日期,'对比时间',SPU ID,商品计划名称,SpuImgUrl,花费
+20260919~20260919,对比时间,10032366483452,波塞冬托特包,img,43.68
+汇总
+0,对比时间,0,,,148.46
+"""
+
+NON_FULL_REPORT_CSV = """查询条件,广告主PIN: 光影行星旗舰店霖; 时间: 2026-09-20至2026-09-20
+SKU ID,SKU 名称,点击时间,'当前时间',花费
+10144206623346,阿波罗斜挎包,20260920~20260920,当前时间,105.50
+汇总
+0,当前时间,0,,,105.50
+"""
+
+
+def _write_report(tmp_path: Path, name: str, content: str) -> Path:
+    path = tmp_path / name
+    path.write_text(content, encoding="utf-8-sig")
+    return path
+
+
+def test_report_content_rows_counts_real_current_time_spu_rows(tmp_path: Path) -> None:
+    path = _write_report(tmp_path, "full_20260919.csv", FULL_REPORT_CSV)
+
+    assert collector._report_content_rows(path, "full") == 2
+
+
+def test_report_content_rows_flags_report_generated_before_data_ready(
+    tmp_path: Path,
+) -> None:
+    path = _write_report(tmp_path, "full_20260920.csv", NOT_READY_FULL_REPORT_CSV)
+
+    assert collector._report_content_rows(path, "full") == 0
+
+
+def test_report_content_rows_counts_non_full_sku_rows(tmp_path: Path) -> None:
+    path = _write_report(tmp_path, "non_full_20260920.csv", NON_FULL_REPORT_CSV)
+
+    assert collector._report_content_rows(path, "non_full") == 1
+
+
+def test_report_content_rows_rejects_csv_without_marker_header(tmp_path: Path) -> None:
+    path = _write_report(tmp_path, "broken.csv", "查询条件\n花费\n1.00\n")
+
+    with pytest.raises(collector.JDAdCostCollectionError, match="表头"):
+        collector._report_content_rows(path, "full")
+
+
+def test_ensure_report_data_accepts_empty_full_when_non_full_has_data(
+    monkeypatch, tmp_path: Path
+) -> None:
+    empty_full = _write_report(tmp_path, "full_20260920.csv", NOT_READY_FULL_REPORT_CSV)
+    sku_report = _write_report(tmp_path, "non_full_20260920.csv", NON_FULL_REPORT_CSV)
+    resubmits: list[collector.ReportType] = []
+    sleeps: list[float] = []
+
+    monkeypatch.setattr(
+        collector,
+        "_report_content_rows",
+        lambda path, rt: 0 if rt == "full" else 1,
+    )
+    monkeypatch.setattr(
+        collector,
+        "_resubmit_report",
+        lambda _page, rt, _td: resubmits.append(rt),
+    )
+    monkeypatch.setattr(collector.time, "sleep", sleeps.append)
+
+    downloaded = collector._ensure_report_data(
+        object(),
+        {"non_full": sku_report, "full": empty_full},
+        date(2026, 9, 20),
+        tmp_path,
+        timeout_seconds=1,
+    )
+
+    assert downloaded == {"non_full": sku_report, "full": empty_full}
+    assert resubmits == []
+    assert sleeps == []
+
+
+def test_ensure_report_data_resubmits_both_reports_when_neither_has_data(
+    monkeypatch, tmp_path: Path
+) -> None:
+    empty_full = _write_report(tmp_path, "full_20260920.csv", NOT_READY_FULL_REPORT_CSV)
+    sku_report = _write_report(tmp_path, "non_full_20260920.csv", NON_FULL_REPORT_CSV)
+    actions: list[object] = []
+    row_counts = iter((0, 0, 1, 2))
+
+    monkeypatch.setattr(
+        collector, "_report_content_rows", lambda _path, _rt: next(row_counts)
+    )
+    monkeypatch.setattr(
+        collector,
+        "_resubmit_report",
+        lambda _page, rt, _td: actions.append(("resubmit", rt)),
+    )
+
+    def fake_download(_page, rt, _td, _od, *, timeout_seconds):
+        del timeout_seconds
+        actions.append(("download", rt))
+        return {"non_full": sku_report, "full": empty_full}[rt]
+
+    monkeypatch.setattr(collector, "download_ready_report", fake_download)
+    monkeypatch.setattr(collector.time, "sleep", lambda _seconds: None)
+
+    downloaded = collector._ensure_report_data(
+        object(),
+        {"non_full": sku_report, "full": empty_full},
+        date(2026, 9, 20),
+        tmp_path,
+        timeout_seconds=1,
+    )
+
+    assert downloaded == {"non_full": sku_report, "full": empty_full}
+    assert actions == [
+        ("resubmit", "non_full"),
+        ("download", "non_full"),
+        ("resubmit", "full"),
+        ("download", "full"),
+    ]
+
+
+def test_ensure_report_data_fails_when_non_full_never_arrives(
+    monkeypatch, tmp_path: Path
+) -> None:
+    empty_full = _write_report(tmp_path, "full_20260920.csv", NOT_READY_FULL_REPORT_CSV)
+    sku_report = _write_report(tmp_path, "non_full_20260920.csv", NON_FULL_REPORT_CSV)
+    resubmits: list[collector.ReportType] = []
+
+    monkeypatch.setattr(collector, "_report_content_rows", lambda _path, _rt: 0)
+    monkeypatch.setattr(
+        collector,
+        "_resubmit_report",
+        lambda _page, rt, _td: resubmits.append(rt),
+    )
+    monkeypatch.setattr(
+        collector, "download_ready_report", lambda _page, rt, _td, _od, *, timeout_seconds: {"non_full": sku_report, "full": empty_full}[rt]
+    )
+    monkeypatch.setattr(collector.time, "sleep", lambda _seconds: None)
+
+    with pytest.raises(collector.JDAdCostCollectionError, match="尚未就绪"):
+        collector._ensure_report_data(
+            object(),
+            {"non_full": sku_report, "full": empty_full},
+            date(2026, 9, 20),
+            tmp_path,
+            timeout_seconds=1,
+        )
+
+    assert resubmits == ["non_full", "full", "non_full", "full"]
@@ -1,5 +1,6 @@
 from __future__ import annotations

+import io
 import json
 import zipfile
 from datetime import date
@@ -171,6 +172,75 @@ def test_safe_extract_zip_returns_csv_and_writes_manifest(tmp_path: Path) -> Non
    assert (tmp_path / "extracted" / "商品报表_20260817.csv").is_file()


+def test_save_signed_download_retries_transient_dns_error(
+    monkeypatch: pytest.MonkeyPatch, tmp_path: Path
+) -> None:
+    class _Response:
+        ok = True
+        status = 200
+
+        def body(self) -> bytes:
+            payload = io.BytesIO()
+            with zipfile.ZipFile(payload, "w") as archive:
+                archive.writestr("商品报表.csv", "日期,主体ID\n")
+            return payload.getvalue()
+
+    class _Request:
+        def __init__(self) -> None:
+            self.calls = 0
+
+        def get(self, url: str, *, timeout: int) -> _Response:
+            del timeout
+            self.calls += 1
+            if self.calls == 1:
+                raise RuntimeError(
+                    f"APIRequestContext.get: getaddrinfo ENOTFOUND {url}"
+                )
+            return _Response()
+
+    request = _Request()
+
+    class _Page:
+        context = type("_Context", (), {"request": request})()
+
+    monkeypatch.setattr(collector.time, "sleep", lambda _seconds: None)
+    download_url = "https://example.test/report.zip?Signature=secret"
+
+    path = collector._save_signed_download(
+        _Page(), download_url, tmp_path, date(2026, 8, 17)
+    )
+
+    assert request.calls == 2
+    assert zipfile.is_zipfile(path)
+
+
+def test_save_signed_download_redacts_signed_url_after_network_retries(
+    monkeypatch: pytest.MonkeyPatch, tmp_path: Path
+) -> None:
+    class _Request:
+        def get(self, url: str, *, timeout: int) -> object:
+            del timeout
+            raise RuntimeError(f"getaddrinfo ENOTFOUND {url}")
+
+    class _Page:
+        context = type(
+            "_Context", (), {"request": _Request()}
+        )()
+
+    monkeypatch.setattr(collector.time, "sleep", lambda _seconds: None)
+    download_url = "https://example.test/report.zip?Signature=secret"
+
+    with pytest.raises(collector.WanxiangDownloadError) as exc_info:
+        collector._save_signed_download(
+            _Page(), download_url, tmp_path, date(2026, 8, 17)
+        )
+
+    message = str(exc_info.value)
+    assert "getaddrinfo ENOTFOUND" in message
+    assert "Signature=secret" not in message
+    assert "https://example.test/report.zip" in message
+
+
 def test_safe_extract_zip_rejects_path_traversal(tmp_path: Path) -> None:
    archive_path = tmp_path / "unsafe.zip"
    with zipfile.ZipFile(archive_path, "w") as archive:
@@ -120,7 +120,7 @@ def test_get_pool_retries_transient_initial_connection(
        "PG_PORT": "5432",
        "PG_DB": "warehouse",
        "PG_USER": "collector",
-        "PG_PASSWORD": "unit-test-only",
+        "PG_PASSWORD": "secret",
    }.items():
        monkeypatch.setenv(name, value)
    monkeypatch.setattr(db, "_pool", None)
@@ -213,3 +213,64 @@ def test_jd_dry_run_accepts_explicit_desktop_style_files(tmp_path: Path, capsys)
    assert "全站 SPU 1" in output
    assert "非全站 SKU 1" in output
    assert "DRY-RUN" in output
+
+
+def test_parse_files_treats_empty_full_report_as_zero_when_sku_has_rows(
+    tmp_path: Path,
+    capsys,
+) -> None:
+    full = tmp_path / "full_20260920.csv"
+    full.write_text(
+        "查询条件,广告主PIN: 光影行星旗舰店霖; 时间: 2026-09-20\n"
+        "日期,'当前时间',SPU ID,商品计划名称,SpuImgUrl,花费,全站投产比\n"
+        "汇总\n"
+        "0,当前时间,0,,,0.00,0.00\n",
+        encoding="utf-8-sig",
+    )
+    sku = tmp_path / "non_full_20260920.csv"
+    sku.write_text(
+        "查询条件,广告主PIN: 光影行星旗舰店霖; 时间: 2026-09-20\n"
+        "SKU ID,SKU 名称,点击时间,'当前时间',花费,展现数,点击数,投产比\n"
+        "sku-1,光影行星 盖亚款式 黑色,20260920~20260920,当前时间,3.20,100,5,2.50\n",
+        encoding="utf-8-sig",
+    )
+
+    full_records, sku_records = importer._parse_files(
+        date(2026, 9, 20),
+        full_file=full,
+        non_full_file=sku,
+        advertiser_pin="光影行星旗舰店霖",
+        shop_account="光影行星研发",
+    )
+
+    assert full_records == []
+    assert [record["sku_id"] for record in sku_records] == ["sku-1"]
+    assert "零投放" in capsys.readouterr().out
+
+
+def test_parse_files_still_fails_when_both_reports_are_empty(tmp_path: Path) -> None:
+    empty = tmp_path / "full_20260920.csv"
+    empty.write_text(
+        "查询条件,广告主PIN: 光影行星旗舰店霖; 时间: 2026-09-20\n"
+        "日期,'当前时间',SPU ID,商品计划名称,SpuImgUrl,花费,全站投产比\n"
+        "0,当前时间,0,,,0.00,0.00\n",
+        encoding="utf-8-sig",
+    )
+    empty_sku = tmp_path / "non_full_20260920.csv"
+    empty_sku.write_text(
+        "查询条件,广告主PIN: 光影行星旗舰店霖; 时间: 2026-09-20\n"
+        "SKU ID,SKU 名称,点击时间,'当前时间',花费,展现数,点击数,投产比\n"
+        "0,汇总占位,20260920~20260920,当前时间,0.00,0,0,0.00\n",
+        encoding="utf-8-sig",
+    )
+
+    import pytest
+
+    with pytest.raises(ValueError, match="当前时间 SKU 行"):
+        importer._parse_files(
+            date(2026, 9, 20),
+            full_file=empty,
+            non_full_file=empty_sku,
+            advertiser_pin="光影行星旗舰店霖",
+            shop_account="光影行星研发",
+        )
@@ -0,0 +1,96 @@
+from __future__ import annotations
+
+import json
+from datetime import datetime
+
+from jd_main_image_collector import (
+    DETAIL_URL,
+    _goto,
+    _load_reusable_style_output,
+    _locate_product_detail_search,
+    _wait_for_product_detail,
+)
+
+
+class _Locator:
+    def __init__(self, visible: bool) -> None:
+        self.visible = visible
+
+    @property
+    def first(self) -> _Locator:
+        return self
+
+    def is_visible(self, timeout: int = 0) -> bool:
+        return self.visible
+
+
+class _Page:
+    def __init__(self, visible_selectors: set[str], goto_errors: int = 0) -> None:
+        self.url = DETAIL_URL
+        self.visible_selectors = visible_selectors
+        self.goto_errors = goto_errors
+        self.waits: list[str] = []
+        self.timeouts: list[int] = []
+
+    def locator(self, selector: str) -> _Locator:
+        return _Locator(selector in self.visible_selectors)
+
+    def wait_for_timeout(self, milliseconds: int) -> None:
+        self.timeouts.append(milliseconds)
+
+    def goto(self, url: str, **kwargs: object) -> None:
+        self.url = url
+        self.waits.append(str(kwargs["wait_until"]))
+        if self.goto_errors:
+            self.goto_errors -= 1
+            raise RuntimeError("net::ERR_CONNECTION_CLOSED")
+
+
+def test_product_detail_search_uses_new_jd_input() -> None:
+    page = _Page({"input.jd-input-multiple-input-input"})
+
+    assert _locate_product_detail_search(page) is not None
+
+
+def test_product_detail_wait_accepts_legacy_wrapper_fallback() -> None:
+    page = _Page({".jmtd-multiple-input"})
+
+    _wait_for_product_detail(page, timeout_ms=1_000)
+
+
+def test_shangzhi_navigation_uses_domcontentloaded() -> None:
+    page = _Page(set())
+
+    _goto(page, DETAIL_URL)
+
+    assert page.waits == ["domcontentloaded"]
+
+
+def test_product_detail_navigation_uses_new_jd_host() -> None:
+    assert DETAIL_URL == "https://jdsz.jd.com/szweb/view/product/productDetail.html"
+
+
+def test_transient_detail_navigation_is_retried() -> None:
+    page = _Page(set(), goto_errors=1)
+
+    _goto(page, DETAIL_URL)
+
+    assert page.waits == ["domcontentloaded", "domcontentloaded"]
+    assert page.timeouts == [1000]
+
+
+def test_same_date_result_is_reused_only_with_local_image(tmp_path) -> None:
+    style_dir = tmp_path / datetime.now().strftime("%Y-%m-%d") / "同款"
+    style_dir.mkdir(parents=True)
+    image_path = style_dir / "jd_images" / "main.jpg"
+    image_path.parent.mkdir()
+    image_path.write_bytes(b"image")
+    (style_dir / "jd_main_images.json").write_text(
+        json.dumps({"records": [{"image_key": "main.jpg", "local_image_path": str(image_path)}]}),
+        encoding="utf-8",
+    )
+
+    reusable = _load_reusable_style_output(tmp_path, "同款")
+
+    assert reusable is not None
+    assert reusable["records"][0]["image_key"] == "main.jpg"
@@ -34,6 +34,34 @@ class MainImagePathTests(unittest.TestCase):
            self.assertEqual(resolve_images_dir(style_dir, "jd"), current)


+class MainImageLarkRetryTests(unittest.TestCase):
+    def test_transient_feishu_transport_error_is_retried(self):
+        with (
+            patch.object(
+                tmall_insert,
+                "run_lark",
+                side_effect=[(1, "", "EOF"), (0, "ok", "")],
+            ) as run_lark,
+            patch.object(tmall_insert.time, "sleep") as sleep,
+        ):
+            result = tmall_insert.run_lark_with_retry(["base", "+field-list"])
+
+        self.assertEqual(result, (0, "ok", ""))
+        self.assertEqual(run_lark.call_count, 2)
+        sleep.assert_called_once_with(1.0)
+
+    def test_non_transient_feishu_error_is_not_retried(self):
+        with patch.object(
+            tmall_insert,
+            "run_lark",
+            return_value=(4, "", "missing field"),
+        ) as run_lark:
+            result = tmall_insert.run_lark_with_retry(["base", "+field-list"])
+
+        self.assertEqual(result, (4, "", "missing field"))
+        run_lark.assert_called_once()
+
+
 class MainImageWorkflowPreflightTests(unittest.TestCase):
    def test_zero_valid_styles_is_a_formal_failure(self):
        for runner in (tmall_runner, jd_runner):
@@ -8,6 +8,7 @@ from gyxx_flow.modules.product_commerce.market_rank_product_import import (
    extract_document_labels,
    normalize_scene_name,
    normalize_style_name,
+    reuse_saved_categories,
 )

 RUNTIME_ROOT = (
@@ -87,6 +88,28 @@ def test_missing_tmall_price_keeps_product_but_excludes_it_from_amount():
    assert snapshot["quality_status"] == "missing_price"


+def test_saved_category_is_reused_by_platform_product_id_without_changing_raw_data():
+    snapshots = [
+        {
+            "platform_product_id": "123",
+            "category_name": "旅行箱",
+            "raw_data": {"category_name": "旅行箱", "rank": 1},
+        },
+        {
+            "platform_product_id": "456",
+            "category_name": "双肩包",
+            "raw_data": {"category_name": "双肩包", "rank": 2},
+        },
+    ]
+
+    reused, count = reuse_saved_categories(snapshots, {"123": "商务旅行箱"})
+
+    assert reused[0]["category_name"] == "商务旅行箱"
+    assert reused[0]["raw_data"] == {"category_name": "旅行箱", "rank": 1}
+    assert reused[1]["category_name"] == "双肩包"
+    assert count == 1
+
+
 def test_feishu_document_labels_keep_raw_values_and_normalized_values():
    content = """
    <title>京东榜单</title>
@@ -111,6 +134,7 @@ def test_feishu_document_labels_keep_raw_values_and_normalized_values():
    assert labels["123456"]["style_name"] == "都市新旅"
    assert labels["123456"]["raw_scene_name"] == "上班通勤"
    assert labels["123456"]["scene_name"] == "通勤上班"
+    assert labels["123456"]["category_name"] == "旅行箱"
    assert labels["123456"]["unit_price_text"] == "￥399"


@@ -121,3 +145,4 @@ def test_collector_schema_contains_idempotent_market_snapshot_tables():
    assert "CREATE TABLE IF NOT EXISTS fact_market_product_snapshot" in schema
    assert "CREATE TABLE IF NOT EXISTS market_product_classification" in schema
    assert "UNIQUE (report_id, market_product_id, category_name)" in schema
+    assert "ADD COLUMN IF NOT EXISTS category_name TEXT" in schema
@@ -11,6 +11,7 @@ import collect_jd_persona_to_bitable as jd_worker
 import collect_persona_to_bitable as tm_worker
 import dy_audience_profile_collect as dy_collector
 import pytest
+import run_daily_persona as persona_launcher
 import taobao_dmp_item_crowd_insight_screenshots as dmp

 from gyxx_flow.adapters import acceptance_policy
@@ -315,6 +316,15 @@ def test_dy_persistent_profile_is_not_overwritten_by_exported_cookies() -> None:
    }


+def test_dy_account_profile_receives_vault_cookies_without_sharing_profile() -> None:
+    cookies = [{"name": "session", "value": "new", "domain": ".jinritemai.com"}]
+
+    assert dy_collector.account_profile_launch_options("bound-profile", cookies) == {
+        "user_data_dir": "bound-profile",
+        "cookies": cookies,
+    }
+
+
 def test_dy_profileless_fallback_can_restore_exported_cookies() -> None:
    cookies = [{"name": "session", "value": "new", "domain": ".jinritemai.com"}]

@@ -397,3 +407,48 @@ def test_jd_persona_login_home_navigation_does_not_wait_for_networkidle() -> Non
    assert 'wait_until="networkidle"' not in home_block[:400]
    assert 'wait_until="domcontentloaded"' in home_block[:400]
    assert "timeout=60_000" in home_block[:400]
+
+
+@pytest.mark.parametrize(
+    ("url", "expected"),
+    (
+        (
+            "https://sz.jd.com/szweb/sz/view/productAnalysis/productDetail.html",
+            True,
+        ),
+        (
+            "https://jdsz.jd.com/szweb/view/product/productDetail.html",
+            True,
+        ),
+        ("https://passport.jd.com/new/login.aspx?ReturnUrl=productDetail", False),
+        ("https://sz.jd.com/sz/view/notAuthorized.html", False),
+        ("https://example.com/szweb/view/product/productDetail.html", False),
+    ),
+)
+def test_jd_product_detail_route_accepts_vendor_redirect_only(
+    url: str,
+    expected: bool,
+) -> None:
+    assert jd_worker._is_jd_product_detail_url(url) is expected
+
+
+def test_persona_launcher_can_select_failed_platforms_and_styles() -> None:
+    assert persona_launcher.selected_platforms("dy,jd,dy") == ["dy", "jd"]
+    assert persona_launcher.selected_platforms("") == ["tm", "dy", "jd"]
+    assert persona_launcher.build_child_command(
+        "collector.py",
+        BUSINESS_DATE,
+        "布谷,白鹭,蓝鹊",
+    ) == [
+        sys.executable,
+        "collector.py",
+        "--date",
+        "2026-08-04",
+        "--styles",
+        "布谷,白鹭,蓝鹊",
+    ]
+
+
+def test_persona_launcher_rejects_unknown_platform() -> None:
+    with pytest.raises(ValueError, match="不支持的平台"):
+        persona_launcher.selected_platforms("tm,unknown")
@@ -98,6 +98,26 @@ def test_persona_workers_rebind_their_nested_browser_collectors(monkeypatch) ->
        _assert_child_binding(observed["environment"])


+def test_dy_persona_default_uses_inner_account_profile(monkeypatch) -> None:
+    observed: dict[str, object] = {}
+    monkeypatch.setattr(
+        dy_persona_worker,
+        "environment_for_child_script",
+        _fake_rebind(observed),
+    )
+
+    def fake_run(command, **kwargs):
+        observed["command"] = command
+        observed["environment"] = kwargs["env"]
+        return subprocess.CompletedProcess(command, 0)
+
+    monkeypatch.setattr(dy_persona_worker.subprocess, "run", fake_run)
+
+    assert dy_persona_worker.run_collect_all("", False, date(2026, 8, 4)) is True
+    assert "--user-data-dir" not in observed["command"]
+    _assert_child_binding(observed["environment"])
+
+
 def test_daily_run_step_rebinds_to_command_target(monkeypatch) -> None:
    observed: dict[str, object] = {}
    monkeypatch.setattr(daily, "environment_for_child_script", _fake_rebind(observed))
@@ -592,6 +592,44 @@ class FeishuDocumentWriteTests(unittest.TestCase):
        self.assertEqual(doc_id, "doc_token")
        run.assert_called_once()

+    def test_ambiguous_doc_create_reconciles_and_retries_after_missing_user_error(self):
+        created = {
+            "data": {"document": {
+                "url": "https://example.feishu.cn/docx/doc_token",
+                "document_id": "doc_token",
+            }},
+            "permission_grant": {"status": "granted", "perm": "full_access"},
+        }
+        responses = [
+            RuntimeError(
+                'lark-cli failed (exit=1): {"error":{"code":20008,'
+                '"message":"The user does not exist."}}'
+            ),
+            {"ok": True, "data": {"results": []}},
+            {"identities": {"user": {"openId": "ou_current_user"}}},
+            created,
+        ]
+
+        with (
+            patch.object(analysis, "run_lark_cli", side_effect=responses) as run,
+            patch.object(analysis, "DOC_CREATE_RECONCILE_DELAYS_SECONDS", (0,)),
+        ):
+            url, doc_id = analysis.write_feishu_doc("测试分析", "# 内容")
+
+        self.assertEqual((url, doc_id), (
+            "https://example.feishu.cn/docx/doc_token",
+            "doc_token",
+        ))
+        self.assertEqual(
+            [call.args[0][:2] for call in run.call_args_list],
+            [
+                ["docs", "+create"],
+                ["docs", "+search"],
+                ["auth", "status"],
+                ["docs", "+create"],
+            ],
+        )
+
    def test_missing_current_user_identity_stops_before_base_write(self):
        responses = [
            {"data": {"document": {
@@ -720,20 +758,39 @@ class ExternalEffectCheckpointTests(unittest.TestCase):
                    "permission_granted",
                )

-    def test_ambiguous_and_corrupt_checkpoints_fail_closed(self):
+    def test_unconfirmed_checkpoint_reconciles_before_retrying_and_corrupt_fails_closed(self):
        with tempfile.TemporaryDirectory() as tmp:
            with patch.object(analysis, "CURATED_DATA_ROOT", Path(tmp)):
                ambiguous = self.checkpoint_spec()
                state, owns_creation = analysis.claim_style_analysis_checkpoint(ambiguous)
                self.assertTrue(owns_creation)
                self.assertEqual(state["phase"], "doc_creating")
-                with patch.object(analysis, "run_lark_cli") as run:
-                    with self.assertRaisesRegex(
-                        analysis.StyleAnalysisCheckpointError,
-                        "doc_creating",
-                    ):
-                        analysis.write_feishu_doc(ambiguous.title, "# 内容", ambiguous)
-                run.assert_not_called()
+                responses = [
+                    {"ok": True, "data": {"results": []}},
+                    {"data": {"document": {
+                        "url": "https://example.feishu.cn/docx/doc_token",
+                        "document_id": "doc_token",
+                    }}, "permission_grant": {"status": "granted"}},
+                ]
+                with (
+                    patch.object(analysis, "run_lark_cli", side_effect=responses) as run,
+                    patch.object(analysis, "DOC_CREATE_RECONCILE_DELAYS_SECONDS", (0,)),
+                ):
+                    url, doc_id = analysis.write_feishu_doc(
+                        ambiguous.title,
+                        "# 内容",
+                        ambiguous,
+                    )
+                self.assertEqual(doc_id, "doc_token")
+                self.assertTrue(url.endswith("/docx/doc_token"))
+                self.assertEqual(
+                    analysis.load_style_analysis_checkpoint(ambiguous)["phase"],
+                    "permission_granted",
+                )
+                self.assertEqual(
+                    [call.args[0][:2] for call in run.call_args_list],
+                    [["docs", "+search"], ["docs", "+create"]],
+                )

            with patch.object(analysis, "CURATED_DATA_ROOT", Path(tmp) / "corrupt"):
                corrupt = self.checkpoint_spec()
@@ -21,6 +21,78 @@ def _entry() -> baibu.BaibuEntry:
    )


+def test_export_template_retries_transient_network_failure(
+    monkeypatch,
+    tmp_path: Path,
+) -> None:
+    calls: list[object] = []
+    sleeps: list[int] = []
+
+    def fake_run(_command, **_kwargs):
+        calls.append(_command)
+        if len(calls) == 2:
+            (tmp_path / "demo.xlsx").write_bytes(b"xlsx")
+            return SimpleNamespace(returncode=0, stdout='{"ok":true}', stderr="")
+        return SimpleNamespace(
+            returncode=4,
+            stdout=(
+                '{"ok":false,"error":{"type":"network",'
+                '"subtype":"timeout"}}'
+            ),
+            stderr="",
+        )
+
+    monkeypatch.setattr(baibu, "_lark_cli", lambda: "lark-cli")
+    monkeypatch.setattr(baibu.subprocess, "run", fake_run)
+    monkeypatch.setattr(baibu.time, "sleep", sleeps.append)
+
+    result = baibu._export_template(_entry(), tmp_path)
+
+    assert result == tmp_path / "demo.xlsx"
+    assert len(calls) == 2
+    assert sleeps == [5]
+
+
+def test_export_template_does_not_retry_non_transient_failure(
+    monkeypatch,
+    tmp_path: Path,
+) -> None:
+    calls: list[object] = []
+    sleeps: list[int] = []
+
+    def fake_run(_command, **_kwargs):
+        calls.append(_command)
+        return SimpleNamespace(
+            returncode=4,
+            stdout=(
+                '{"ok":false,"error":{"type":"permission",'
+                '"message":"forbidden"}}'
+            ),
+            stderr="",
+        )
+
+    monkeypatch.setattr(baibu, "_lark_cli", lambda: "lark-cli")
+    monkeypatch.setattr(baibu.subprocess, "run", fake_run)
+    monkeypatch.setattr(baibu.time, "sleep", sleeps.append)
+
+    with pytest.raises(baibu.BaibuApplyError, match="exit=4"):
+        baibu._export_template(_entry(), tmp_path)
+
+    assert len(calls) == 1
+    assert sleeps == []
+
+
+def test_safe_cli_detail_redacts_feishu_tokens() -> None:
+    detail = baibu._safe_cli_detail(
+        '{"token":"secret-token","message":"?token=query-token"}',
+        "",
+    )
+
+    assert "secret-token" not in detail
+    assert "query-token" not in detail
+    assert detail.count("<redacted>") == 2
+
+
 def test_latest_batch_import_status_parses_processing_and_complete_summary() -> None:
    processing = baibu._latest_batch_import_status(
        "您于2026-08-26 09:23:37执行了商品批量导入操作，正在处理中。"
@@ -123,13 +195,13 @@ class _LoginLocator:
 class _LoginFrame:
    def __init__(self) -> None:
        self.username = _LoginLocator()
-        self.password = _LoginLocator()
+        self.password_input = _LoginLocator()
        self.submit = _LoginLocator()

    def locator(self, selector: str):
        return {
            "input[placeholder='账号名/邮箱/手机号']": self.username,
-            "input[placeholder='请输入登录密码']": self.password,
+            "input[placeholder='请输入登录密码']": self.password_input,
            "button[type='submit']": _LoginLocator(visible=False),
            "input[type='submit']": _LoginLocator(visible=False),
        }.get(selector, _LoginLocator(visible=False))
@@ -161,7 +233,7 @@ def test_login_if_needed_fills_the_password_form_inside_an_iframe(monkeypatch) -

    assert attempted is True
    assert frame.username.filled == ["demo-account"]
-    assert frame.password.filled == ["demo-password"]
+    assert frame.password_input.filled == ["demo-password"]
    assert frame.submit.click_count == 1


@@ -181,7 +253,7 @@ def test_manual_login_waits_without_filling_the_password_form(monkeypatch) -> No

    assert attempted is True
    assert frame.username.filled == []
-    assert frame.password.filled == []
+    assert frame.password_input.filled == []
    assert frame.submit.click_count == 0


@@ -208,7 +280,7 @@ def test_login_if_needed_reports_missing_credentials_without_filling(monkeypatch
        raise AssertionError("missing credentials must stop before filling")

    assert frame.username.filled == []
-    assert frame.password.filled == []
+    assert frame.password_input.filled == []


 def test_wait_for_authenticated_stops_on_explicit_login_error(monkeypatch) -> None:
@@ -555,18 +627,18 @@ def test_complete_draft_products_runs_required_sequence_until_count_is_zero(
        "draft",
        "open:True",
        "否",
-        "下一步",
+        "提交",
        "draft",
        "open:False",
        "否",
-        "下一步",
+        "提交",
        "draft",
    ]
    assert clicked == [
        ("否",),
-        ("下一步",),
+        ("提交", "下一步"),
        ("否",),
-        ("下一步",),
+        ("提交", "下一步"),
    ]


@@ -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(
@@ -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
+    )
@@ -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
@@ -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
@@ -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
@@ -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
@@ -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"
@@ -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") == {}
@@ -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