feat: 添加 WikipediaTopViewsSource 并增强查询源优先级系统

config.example.yaml

@@ -57,6 +57,11 @@ query_engine:
     wikipedia:
       enabled: true # Wikipedia 热门话题
       timeout: 15 # 超时（秒）
+    wikipedia_top_views:
+      enabled: true # Wikipedia 热门浏览（真实热搜数据）
+      timeout: 30 # 超时（秒）
+      lang: "en" # 语言
+      ttl: 21600 # 缓存 TTL（秒，默认 6 小时）
     bing_suggestions:
       enabled: true # Bing 建议API
   bing_api:

pyproject.toml

@@ -8,7 +8,7 @@ version = "1.0.0"
 requires-python = ">=3.10"
 dependencies = [
     "playwright>=1.49.0",
-    "playwright-stealth>=1.0.6",
+    "playwright-stealth>=1.0.6,<2.0",
     "pyyaml>=6.0.1",
     "aiohttp>=3.11.0",
     "beautifulsoup4>=4.12.3",
@@ -35,17 +35,7 @@ dev = [
     "pytest-xdist>=3.5.0",
     "hypothesis>=6.125.0",
     "faker>=35.0.0",
-]
-test = [
-    "pytest>=8.0.0",
-    "pytest-asyncio>=0.24.0",
-    "pytest-playwright>=0.5.0",
-    "pytest-benchmark>=5.0.0",
-    "pytest-cov>=6.0.0",
-    "pytest-timeout>=2.3.0",
-    "pytest-xdist>=3.5.0",
-    "hypothesis>=6.125.0",
-    "faker>=35.0.0",
+    "respx>=0.21.0",
 ]
 viz = [
     "streamlit>=1.41.0",

src/account/points_detector.py

@@ -8,15 +8,15 @@
 
 from playwright.async_api import Page
 
-from constants import REWARDS_URLS
+from api.dashboard_client import DashboardClient
 
 logger = logging.getLogger(__name__)
 
 
 class PointsDetector:
     """积分检测器类"""
 
-    DASHBOARD_URL = REWARDS_URLS["dashboard"]
+    DASHBOARD_URL = "https://rewards.bing.com/"
 
     POINTS_SELECTORS = [
         "p.text-title1.font-semibold",
@@ -90,11 +90,25 @@ async def get_current_points(self, page: Page, skip_navigation: bool = False) ->
                 logger.debug("跳过导航，使用当前页面")
                 await page.wait_for_timeout(1000)
 
+            # 优先使用 Dashboard API
+            try:
+                logger.debug("尝试使用 Dashboard API 获取积分...")
+                client = DashboardClient(page)
+                api_points: int | None = await client.get_current_points()
+                if api_points is not None and api_points >= 0:
+                    logger.debug(f"✓ 从 API 获取积分: {api_points:,}")
+                    return int(api_points)
+            except Exception as e:
+                logger.warning(
+                    f"API 获取积分失败（{type(e).__name__}: {e}），使用 HTML 解析作为备用"
+                )
+
+            # 备用：HTML 解析
             logger.debug("尝试从页面源码提取积分...")
             points = await self._extract_points_from_source(page)
 
             if points is not None:
-                logger.info(f"✓ 从源码提取积分: {points:,}")
+                logger.debug(f"✓ 从源码提取积分: {points:,}")
                 return points
 
             logger.debug("源码提取失败，尝试选择器...")
@@ -107,13 +121,14 @@ async def get_current_points(self, page: Page, skip_navigation: bool = False) ->
                         points_text = await element.text_content()
                         logger.debug(f"找到积分文本: {points_text}")
 
-                        points = self._parse_points(points_text)
+                        if points_text:
+                            points = self._parse_points(points_text)
 
-                        if points is not None and points >= 100:
-                            logger.info(f"✓ 当前积分: {points:,}")
-                            return points
-                        elif points is not None:
-                            logger.debug(f"积分值太小，可能是误识别: {points}")
+                            if points is not None and points >= 100:
+                                logger.debug(f"✓ 当前积分: {points:,}")
+                                return points
+                            elif points is not None:
+                                logger.debug(f"积分值太小，可能是误识别: {points}")
 
                 except Exception as e:
                     logger.debug(f"选择器 {selector} 失败: {e}")
@@ -143,7 +158,7 @@ def _parse_points(self, text: str) -> int | None:
         Returns:
             积分数量，失败返回 None
         """
-        if not text:
+        if not text or not text.strip():
             return None
 
         try:
@@ -310,7 +325,12 @@ async def _check_task_status(self, page: Page, selectors: list, task_name: str)
         Returns:
             任务状态字典
         """
-        status = {"found": False, "completed": False, "progress": None, "max_progress": None}
+        status: dict[str, bool | int | None] = {
+            "found": False,
+            "completed": False,
+            "progress": None,
+            "max_progress": None,
+        }
 
         try:
             for selector in selectors:
@@ -338,10 +358,12 @@ async def _check_task_status(self, page: Page, selectors: list, task_name: str)
                             # 查找类似 "15/30" 的进度
                             progress_match = re.search(r"(\d+)\s*/\s*(\d+)", text)
                             if progress_match:
-                                status["progress"] = int(progress_match.group(1))
-                                status["max_progress"] = int(progress_match.group(2))
+                                progress_val = int(progress_match.group(1))
+                                max_progress_val = int(progress_match.group(2))
+                                status["progress"] = progress_val
+                                status["max_progress"] = max_progress_val
 
-                                if status["progress"] >= status["max_progress"]:
+                                if progress_val >= max_progress_val:
                                     status["completed"] = True
 
                         logger.debug(f"{task_name} 状态: {status}")

src/api/__init__.py

@@ -0,0 +1,5 @@
+"""API clients module"""
+
+from .dashboard_client import DashboardClient
+
+__all__ = ["DashboardClient"]

src/api/dashboard_client.py

@@ -0,0 +1,170 @@
+"""
+Dashboard API Client
+
+Fetches points data from Microsoft Rewards Dashboard API.
+"""
+
+import logging
+import re
+from typing import Any
+
+from playwright.async_api import Page
+
+from constants import API_ENDPOINTS, REWARDS_URLS
+
+logger = logging.getLogger(__name__)
+
+
+class DashboardClient:
+    """Client for fetching data from Microsoft Rewards Dashboard API"""
+
+    def __init__(self, page: Page):
+        """
+        Initialize Dashboard client
+
+        Args:
+            page: Playwright Page object
+        """
+        self.page = page
+        self._cached_points: int | None = None
+        base = REWARDS_URLS.get("dashboard", "https://rewards.bing.com")
+        self._base_url = base.rstrip("/")
+
+    async def get_current_points(self) -> int | None:
+        """
+        Get current points from Dashboard API
+
+        Attempts to fetch points via API call first, falls back to
+        parsing page content if API fails.
+
+        Returns:
+            Points balance or None if unable to determine
+        """
+        try:
+            points = await self._fetch_points_via_api()
+            if points is not None and points >= 0:
+                self._cached_points = points
+                return points
+        except TimeoutError as e:
+            logger.warning(f"API request timeout: {e}")
+        except ConnectionError as e:
+            logger.warning(f"API connection error: {e}")
+        except Exception as e:
+            logger.warning(f"API call failed: {e}")
+
+        try:
+            points = await self._fetch_points_via_page_content()
+            if points is not None and points >= 0:
+                self._cached_points = points
+                return points
+        except Exception as e:
+            logger.debug(f"Page content parsing failed: {e}")
+
+        return self._cached_points
+
+    async def _fetch_points_via_api(self) -> int | None:
+        """
+        Fetch points via internal API endpoint
+
+        Returns:
+            Points balance or None
+        """
+        try:
+            api_url = f"{self._base_url}{API_ENDPOINTS['dashboard_balance']}"
+            response = await self.page.evaluate(
+                f"""
+                async () => {{
+                    try {{
+                        const resp = await fetch('{api_url}', {{
+                            method: 'GET',
+                            credentials: 'include'
+                        }});
+                        if (!resp.ok) return null;
+                        return await resp.json();
+                    }} catch {{
+                        return null;
+                    }}
+                }}
+                """
+            )
+
+            if response and isinstance(response, dict):
+                available = response.get("availablePoints")
+                balance = response.get("pointsBalance")
+                points = available if available is not None else balance
+                if points is not None:
+                    try:
+                        return int(points)
+                    except (ValueError, TypeError):
+                        pass
+
+        except Exception as e:
+            logger.debug(f"API fetch error: {e}")
+
+        return None
+
+    async def _fetch_points_via_page_content(self) -> int | None:
+        """
+        Extract points from page content as fallback
+
+        Returns:
+            Points balance or None
+        """
+        try:
+            content = await self.page.content()
+
+            patterns = [
+                r'"availablePoints"\s*:\s*(\d+)',
+                r'"pointsBalance"\s*:\s*(\d+)',
+                r'"totalPoints"\s*:\s*(\d+)',
+            ]
+
+            for pattern in patterns:
+                match = re.search(pattern, content)
+                if match:
+                    points = int(match.group(1))
+                    if 0 <= points <= 1000000:
+                        return points
+
+        except Exception as e:
+            logger.debug(f"Page content extraction error: {e}")
+
+        return None
+
+    async def get_dashboard_data(self) -> dict[str, Any] | None:
+        """
+        Fetch full dashboard data
+
+        Returns:
+            Dashboard data dict or None
+        """
+        try:
+            api_url = f"{self._base_url}{API_ENDPOINTS['dashboard_data']}"
+            response = await self.page.evaluate(
+                f"""
+                async () => {{
+                    try {{
+                        const resp = await fetch('{api_url}', {{
+                            method: 'GET',
+                            credentials: 'include'
+                        }});
+                        if (!resp.ok) return null;
+                        return await resp.json();
+                    }} catch {{
+                        return null;
+                    }}
+                }}
+                """
+            )
+
+            if response is not None and isinstance(response, dict):
+                return dict(response)
+
+        except TimeoutError as e:
+            logger.warning(f"Dashboard API timeout: {e}")
+        except ConnectionError as e:
+            logger.warning(f"Dashboard API connection error: {e}")
+        except Exception as e:
+            logger.warning(f"Dashboard API error: {e}")
+
+        return None

src/constants/urls.py

@@ -37,6 +37,8 @@
 
 API_ENDPOINTS = {
     "dashboard": "https://rewards.bing.com/api/getuserinfo",
+    "dashboard_balance": "/api/getuserbalance",
+    "dashboard_data": "/api/dashboard",
     "report_activity": "https://rewards.bing.com/api/reportactivity",
     "quiz": "https://www.bing.com/bingqa/ReportActivity",
     "app_dashboard": "https://prod.rewardsplatform.microsoft.com/dapi/me",

src/review/models.py

@@ -118,8 +118,10 @@ class ReviewMetadata(BaseModel):
     pr_number: int
     owner: str
     repo: str
+    branch: str = Field(default="", description="拉取评论时的分支名称")
+    head_sha: str = Field(default="", description="拉取评论时的 HEAD commit SHA（前7位）")
     last_updated: str = Field(default_factory=lambda: datetime.utcnow().isoformat())
-    version: str = "2.2"
+    version: str = "2.3"
     etag_comments: str | None = Field(None, description="GitHub ETag，用于条件请求")
     etag_reviews: str | None = Field(None, description="Reviews ETag")