chore: [workflow] 提交工作区代码

tests/modules/product_commerce/test_browser_profile_reuse.py

@@ -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,

tests/modules/product_commerce/test_collect_douyin_qianchuan_ads.py

@@ -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

tests/modules/product_commerce/test_collect_jd_ad_costs.py

@@ -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"]

tests/modules/product_commerce/test_collect_tmall_wanxiang_ads.py

@@ -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:

tests/modules/product_commerce/test_db_get_conn.py

@@ -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)

tests/modules/product_commerce/test_import_jd_ad_costs.py

@@ -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",
+        )

tests/modules/product_commerce/test_jd_main_image_navigation.py

@@ -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"

tests/modules/product_commerce/test_main_image_concurrency.py

@@ -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):

tests/modules/product_commerce/test_market_rank_product_import.py

@@ -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

tests/modules/product_commerce/test_persona_integrity.py

@@ -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")

tests/modules/product_commerce/test_product_nested_child_bindings.py

@@ -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))

tests/modules/product_commerce/test_style_analysis_orchestration.py

@@ -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()

tests/modules/product_commerce/test_tmall_baibu_apply.py

@@ -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"),
     ]