'commit'

Apps/XinDianTu/Crawler.py

@@ -156,12 +156,10 @@ async def get_station_list(d, service, uploader, max_stations_count=MAX_STATIONS
         image_uuid = str(uuid.uuid4())
         screenshot_path = take_screenshot(d, image_uuid, save_dir=TEMP_IMAGE_DIR)
 
-        # 【优化】不再在每页滚动时检查兔子广告，仅在 Opener 进入时检查一次
-        # 如果后续发现有其它非兔子广告弹出，可在此处恢复非兔子广告的检测逻辑
         ad_top_y_norm = 0.78  # 默认的点击边界 (0.78)
-        # 新策略：滚动后的页面完全依赖“速度躲兔子”，不再因为靠近底部而丢弃卡片
         if scroll_count > 1:
             ad_top_y_norm = 1.0
+        logger.info(f"当前列表页广告安全下边界: {ad_top_y_norm:.2f} (scroll_count={scroll_count})")
         
         # 检查是否已经滚动到底部 (排除状态栏后，内容与上次一致)
         current_md5 = Kit.get_image_content_md5(

Apps/XinDianTu/Kit.py

@@ -4,7 +4,7 @@ import cv2
 import numpy as np
 import time
 import hashlib
-from Apps.XinDianTu.Config.Setting import BOTTOM_SAFE_EXCLUDE_RATIO
+from Apps.XinDianTu.Config.Setting import SAFE_EXCLUDE_RATIO, BOTTOM_SAFE_EXCLUDE_RATIO
 from Config.Config import TEMP_IMAGE_DIR
 
 logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s")
@@ -748,26 +748,7 @@ def crop_cards_from_image(img_path, output_dir=None, save_debug=True):
 
     logger.info(f"  Found {len(groups)} segment groups.")
 
-    # Select the Target Group
-    # Strategy: Select the LAST group that has at least one 'substantial' segment (H > 150).
-    # This assumes the Station List is the main content and usually at the bottom (scrolling area).
-    target_group = []
-    for g in reversed(groups):
-        # Check if group has substantial content
-        has_content = False
-        for (y1, y2) in g:
-            if (y2 - y1) > 150:
-                has_content = True
-                break
-
-        if has_content:
-            target_group = g
-            break
-
-    if not target_group and segments:
-        # Fallback to all segments if no substantial group found
-        target_group = segments
-
+    target_group = segments
     logger.info(f"  Selected group with {len(target_group)} segments.")
 
     # 2. 确定每个卡片的左右边界 (Width Refinement)
@@ -881,10 +862,7 @@ def crop_cards_from_image(img_path, output_dir=None, save_debug=True):
             max_h = max(heights)
             median_h = np.median(heights)
 
-            # 策略：如果高度差异较大，丢弃过小的卡片
-            # 阈值设定为 Median 的 75% (允许一定程度的偏差，如 220 vs 258 是允许的，但 150 vs 250 应被过滤)
-            # 使用 Median 比 Max 更稳健，避免被单个超高卡片(如广告)带偏
-            threshold_h = median_h * 0.70
+            threshold_h = median_h * 0.80
 
             for (y1, y2), card_h in zip(temp_valid_segments, heights):
                 if card_h < threshold_h:
@@ -914,6 +892,19 @@ def crop_cards_from_image(img_path, output_dir=None, save_debug=True):
         "cards": []
     }
 
+    # 顶部安全过滤：丢弃中心落在 SAFE_EXCLUDE_RATIO 以上区域的卡片
+    filtered_cards = []
+    for y1, y2, x1, x2 in final_cards:
+        center_y = (y1 + y2) / 2.0
+        if center_y < h * SAFE_EXCLUDE_RATIO:
+            logger.info(
+                f"  Dropping card Y={y1}-{y2} (center={center_y:.1f}) "
+                f"because it is within top safe area ({SAFE_EXCLUDE_RATIO*100:.0f}%)."
+            )
+            continue
+        filtered_cards.append((y1, y2, x1, x2))
+    final_cards = filtered_cards
+
     # 准备 _vl.jpg (只画框，不画红点)
     vl_img = img.copy()
 

Apps/XinDianTu/ReadImageKit.py

@@ -1144,40 +1144,89 @@ class ReadImageKit:
         # 寻找按钮坐标任务
         button_task = asyncio.create_task(ReadImageKit.find_all_time_button_coordinate(image_url_or_path, device_info))
 
-        prompt = (
-            "仅输出JSON对象（不含任何说明文字）。"
-            "任务1：识别图片中充电站的完整名称（full_station_name）。"
-            f"提示：列表中看到的名称可能是截断的（例如“{station_name}”），请在图片上方找到最匹配的完整名称。"
-            "任务2：识别充电站的详细地址（address）。"
-            "寻找规则：地址通常紧跟在场站名称下方，或者在‘距离’图标（定位小图表）附近，或者在带有‘导航’按钮的同一行。"
-            "返回包含 full_station_name 和 address 字段的JSON对象，例如 {\"full_station_name\": \"完整名称\", \"address\": \"详细地址\"}。"
-            "如果找不到，对应字段返回空字符串。"
-            "严格返回纯JSON格式。"
-        )
+        prompt = (
+            "仅输出JSON对象（不含任何说明文字）。"
+            "任务1：识别图片中充电站的完整名称（full_station_name）。"
+            f"提示：列表中看到的名称可能是截断的（例如“{station_name}”），请在图片上方找到最匹配的完整名称。"
+            "任务2：识别充电站的详细地址（address）。"
+            "寻找规则：地址通常紧跟在场站名称下方，或者在“距离”图标（定位小图标）附近，或者在带有“导航”按钮的同一行。"
+            "若地址跨多行，请合并为一行返回，不要省略。"
+            "返回包含 full_station_name 和 address 字段的JSON对象，例如 {\"full_station_name\": \"完整名称\", \"address\": \"详细地址\"}。"
+            "如果图片中能看到任何疑似地址的文字（如街道名、路名、园区名），请尽量提取出来并填入 address，而不是返回空字符串。"
+            "实在无法确定时，对应字段返回空字符串。"
+            "严格返回纯JSON格式。"
+        )
+
+        result = {}
+        try:
+            resp = await ReadImageKit._client.chat.completions.create(
+                model=VL_MODEL_NAME,
+                timeout=30,
+                messages=[
+                    {
+                        "role": "user",
+                        "content": [
+                            {"type": "image_url", "image_url": img_input},
+                            {"type": "text", "text": json.dumps(device_info, ensure_ascii=False)},
+                            {"type": "text", "text": prompt},
+                        ],
+                    },
+                ],
+            )
+            content = resp.choices[0].message.content or ""
+            raw = ReadImageKit._extract_json(content)
+            result = json.loads(raw)
+        except Exception as e:
+            logger.error(f"解析地址失败 ({station_name}): {e}")
+
+        # 首次结果兜底与必要时的重试
+        if not isinstance(result, dict):
+            result = {}
+
+        full_name = str(result.get("full_station_name") or "").strip()
+        address = str(result.get("address") or "").strip()
+
+        if not full_name and not address:
+            try:
+                logger.info(f"地址首次解析为空，针对场站“{station_name}”进行重试。")
+                retry_prompt = (
+                    "仅输出JSON对象（不含任何说明文字）。"
+                    "请再次仔细查找图片中的充电站名称和详细地址，尽量不要返回空字符串。"
+                    "即便地址不完整，也请返回你能看到的主要关键信息（如道路名、园区名、街道名等）。"
+                    "返回格式仍为 {\"full_station_name\": \"完整名称或列表中的截断名称\", \"address\": \"尽量完整的详细地址\"}。"
+                    "严格返回纯JSON格式。"
+                )
+                resp = await ReadImageKit._client.chat.completions.create(
+                    model=VL_MODEL_NAME,
+                    timeout=30,
+                    messages=[
+                        {
+                            "role": "user",
+                            "content": [
+                                {"type": "image_url", "image_url": img_input},
+                                {"type": "text", "text": json.dumps(device_info, ensure_ascii=False)},
+                                {"type": "text", "text": retry_prompt},
+                            ],
+                        },
+                    ],
+                )
+                content = resp.choices[0].message.content or ""
+                raw = ReadImageKit._extract_json(content)
+                retry_result = json.loads(raw)
+                if isinstance(retry_result, dict):
+                    result = retry_result
+                    full_name = str(result.get("full_station_name") or "").strip()
+                    address = str(result.get("address") or "").strip()
+            except Exception as e:
+                logger.error(f"地址重试解析失败 ({station_name}): {e}")
+
+        # 最终兜底：至少回填一个名称，避免 None
+        if not full_name:
+            result["full_station_name"] = station_name or ""
+        if "address" not in result:
+            result["address"] = address
 
-        result = {}
-        try:
-            resp = await ReadImageKit._client.chat.completions.create(
-                model=VL_MODEL_NAME,
-                timeout=30,
-                messages=[
-                    {
-                        "role": "user",
-                        "content": [
-                            {"type": "image_url", "image_url": img_input},
-                            {"type": "text", "text": json.dumps(device_info, ensure_ascii=False)},
-                            {"type": "text", "text": prompt},
-                        ],
-                    },
-                ],
-            )
-            content = resp.choices[0].message.content or ""
-            raw = ReadImageKit._extract_json(content)
-            result = json.loads(raw)
-        except Exception as e:
-            logger.error(f"解析地址失败 ({station_name}): {e}")
-
-        # 合并按钮坐标
+        # 合并按钮坐标
         try:
             button_result = await button_task
             if button_result: