perf: 量化商品页打开耗时并去重检查 (#109)

client/src/pdd_collect_service.py

@@ -11,6 +11,7 @@ import math
 import re
 import time
 import xml.etree.ElementTree as ET
+from contextlib import nullcontext
 from dataclasses import dataclass
 from datetime import datetime, timezone
 from decimal import Decimal, InvalidOperation, ROUND_HALF_UP
@@ -19,6 +20,7 @@ from typing import Any, Callable, Iterable, Mapping, Optional, Sequence
 from urllib.parse import parse_qs, urlparse
 
 from .pdd_device_service import PDD_PACKAGE_NAME, PddDeviceError, PddDeviceService
+from .performance_timing import current_performance_trace
 from .util.get_size_panle_coord import get_size_panel_coord
 
 
@@ -691,8 +693,11 @@ class PddCollectService:
         self._check_cancelled()
 
         try:
-            with self._device_service.connect(self._device_address) as device:
-                self._open_goods(device, goods_url)
+            session = self._device_service.connect(self._device_address)
+            with session as device:
+                self._open_goods(
+                    device, goods_url, session.initial_app_state
+                )
                 goods = self._collect_goods_details(device)
                 if not goods.title:
                     raise PddCollectError("PDD_DATA_TITLE_MISSING", "商品页没有可识别的标题")
@@ -779,14 +784,32 @@ class PddCollectService:
         ):
             raise PddCollectError("PDD_PAGE_OVERALL_TIMEOUT", "PDD 采集超过 10 分钟")
 
-    def _open_goods(self, device: Any, goods_url: str) -> None:
+    def _open_goods(
+        self,
+        device: Any,
+        goods_url: str,
+        initial_app_state: Optional[Mapping[str, Any]] = None,
+    ) -> None:
+        trace = current_performance_trace()
         try:
-            current = device.app_current()
+            current = (
+                dict(initial_app_state)
+                if initial_app_state is not None
+                else device.app_current()
+            )
             if current.get("package") != PDD_PACKAGE_NAME:
-                device.app_start(PDD_PACKAGE_NAME)
-            if not device.app_wait(PDD_PACKAGE_NAME, timeout=10):
-                raise PddCollectError("DEVICE_APP_START_FAILED", "PDD 应用启动失败")
-            device.open_url(goods_url)
+                stage = trace.stage("pdd_start_or_wait") if trace else nullcontext()
+                with stage:
+                    device.app_start(PDD_PACKAGE_NAME)
+                    if not device.app_wait(PDD_PACKAGE_NAME, timeout=10):
+                        raise PddCollectError(
+                            "DEVICE_APP_START_FAILED", "PDD 应用启动失败"
+                        )
+            elif trace is not None:
+                trace.record("pdd_start_or_wait", 0, "already_foreground")
+            stage = trace.stage("open_url") if trace else nullcontext()
+            with stage:
+                device.open_url(goods_url)
         except PddCollectError:
             raise
         except Exception as exc:
@@ -797,30 +820,44 @@ class PddCollectService:
                 ) from exc
             raise PddCollectError("DEVICE_APP_START_FAILED", f"无法打开 PDD 商品链接：{exc}") from exc
 
-        deadline = self._monotonic() + self._page_timeout
-        while self._monotonic() < deadline:
-            self._check_cancelled()
-            current = device.app_current()
-            xml_data = self._dump_hierarchy(device)
-            root = _parse_xml(xml_data)
-            last_labels = _all_labels(root)
-            pdd_node_count = sum(
-                1
-                for node in root.iter("node")
-                if node.get("package") == PDD_PACKAGE_NAME
+        ready_stage = trace.stage("goods_page_ready") if trace else nullcontext()
+        with ready_stage:
+            deadline = self._monotonic() + self._page_timeout
+            first_dump = True
+            while self._monotonic() < deadline:
+                self._check_cancelled()
+                current = device.app_current()
+                if first_dump and trace is not None:
+                    with trace.stage("first_dump_hierarchy"):
+                        xml_data = self._dump_hierarchy(device)
+                    first_dump = False
+                else:
+                    xml_data = self._dump_hierarchy(device)
+                root = _parse_xml(xml_data)
+                last_labels = _all_labels(root)
+                pdd_node_count = sum(
+                    1
+                    for node in root.iter("node")
+                    if node.get("package") == PDD_PACKAGE_NAME
+                )
+                # 部分 OPPO/ColorOS 设备会一直把无线调试设置页报告为焦点，
+                # 即使屏幕和无障碍树已经是 PDD。此时以树中真实包名为准。
+                is_pdd_hierarchy = pdd_node_count >= 3
+                is_pdd_focused = current.get("package") == PDD_PACKAGE_NAME
+                if is_pdd_focused or is_pdd_hierarchy:
+                    _raise_special_page(last_labels)
+                    combined = " ".join(last_labels)
+                    loading = "加载中" in combined or "正在加载" in combined
+                    if not loading and any(
+                        marker in combined for marker in _READY_MARKERS
+                    ):
+                        if trace is not None:
+                            trace.checkpoint("end_to_end_total")
+                        return
+                self._sleep(0.5)
+            raise PddCollectError(
+                "PDD_PAGE_TIMEOUT", "等待 PDD 商品详情页加载超时"
             )
-            # 部分 OPPO/ColorOS 设备会一直把无线调试设置页报告为焦点，
-            # 即使屏幕和无障碍树已经是 PDD。此时以树中真实包名为准。
-            is_pdd_hierarchy = pdd_node_count >= 3
-            is_pdd_focused = current.get("package") == PDD_PACKAGE_NAME
-            if is_pdd_focused or is_pdd_hierarchy:
-                _raise_special_page(last_labels)
-                combined = " ".join(last_labels)
-                loading = "加载中" in combined or "正在加载" in combined
-                if not loading and any(marker in combined for marker in _READY_MARKERS):
-                    return
-            self._sleep(0.5)
-        raise PddCollectError("PDD_PAGE_TIMEOUT", "等待 PDD 商品详情页加载超时")
 
     def _collect_goods_details(self, device: Any) -> GoodsSnapshot:
         """保留首屏结果，并向下浏览补采评价数量和店铺名。"""