'commit'
This commit is contained in:
@@ -196,11 +196,11 @@ class TeLaiDianCrawler(BaseCrawler):
|
||||
"""
|
||||
在详情页提取价格和状态信息
|
||||
"""
|
||||
# 1. 截取第一屏的基础信息图(名称、精确地址),并异步识别
|
||||
first_screen_path = take_screenshot(d, f"tld_detail_basic_{int(time.time())}.jpg")
|
||||
station_name = station_info.get("name")
|
||||
address = station_info.get("address")
|
||||
logger.info(f"已截取详情页首屏截图,启动异步基础信息识别任务: {station_name} | {address}")
|
||||
logger.info(f"[详情页] 进入 crawl_detail_logic,场站: {station_name} | 地址: {address}")
|
||||
logger.info(f"[详情页] 已截取首屏截图,启动异步基础信息识别任务: {first_screen_path}")
|
||||
|
||||
async def analyze_basic_info_background(image_path, fallback_name, fallback_address):
|
||||
try:
|
||||
@@ -219,25 +219,24 @@ class TeLaiDianCrawler(BaseCrawler):
|
||||
|
||||
asyncio.create_task(analyze_basic_info_background(first_screen_path, station_name, address))
|
||||
|
||||
# 2. 一次向上滑动,确保当前底部文字完全顶到上沿之外
|
||||
w, h = d.window_size()
|
||||
logger.info("执行一次向上滑动,将当前底部文字完全推至屏幕上沿之外...")
|
||||
d.swipe(w * 0.5, h * 0.75, w * 0.5, h * 0.25, 0.8)
|
||||
await asyncio.sleep(1.5)
|
||||
logger.info("[详情页] 先执行一次较大的向上滑动,将详情内容上移一屏左右")
|
||||
d.swipe(w * 0.5, h * 0.8, w * 0.5, h * 0.3, 0.5)
|
||||
await asyncio.sleep(1.0)
|
||||
|
||||
# 3. 使用模板 jgxx.jpg 匹配“价格信息 当前价”入口并点击
|
||||
template_jgxx = os.path.join(os.path.dirname(__file__), "Template", "jgxx.jpg")
|
||||
entrance_point = None
|
||||
|
||||
if os.path.exists(template_jgxx):
|
||||
logger.info(f"使用模板匹配价格入口: {template_jgxx}")
|
||||
logger.info(f"[详情页] 使用模板匹配价格入口: {template_jgxx},第 1 轮(只匹配一次)")
|
||||
try:
|
||||
match_res = d.image.match(template_jgxx)
|
||||
except Exception as e:
|
||||
match_res = None
|
||||
logger.error(f"模板匹配价格入口失败: {e}")
|
||||
logger.error(f"[详情页] 模板匹配价格入口失败: {e}")
|
||||
|
||||
if match_res:
|
||||
logger.info(f"[详情页] 第 1 轮模板匹配返回结果: {match_res}")
|
||||
if hasattr(match_res, "point") and match_res.point:
|
||||
entrance_point = match_res.point
|
||||
elif isinstance(match_res, dict):
|
||||
@@ -248,26 +247,25 @@ class TeLaiDianCrawler(BaseCrawler):
|
||||
elif isinstance(match_res, (list, tuple)) and len(match_res) >= 2:
|
||||
entrance_point = (match_res[0], match_res[1])
|
||||
|
||||
if entrance_point:
|
||||
logger.info(f"通过 jgxx.jpg 成功找到价格入口,点击坐标: {entrance_point}")
|
||||
d.click(int(entrance_point[0]), int(entrance_point[1]))
|
||||
await asyncio.sleep(WAIT_DETAIL_PAGE_LOAD)
|
||||
else:
|
||||
logger.warning("未能通过 jgxx.jpg 找到价格入口,使用坐标兜底点击页面中部偏下位置。")
|
||||
d.click(w // 2, int(h * 0.45))
|
||||
await asyncio.sleep(WAIT_DETAIL_PAGE_LOAD)
|
||||
if not entrance_point:
|
||||
logger.error("[详情页] 未能通过 jgxx.jpg 找到价格入口,本次流程直接中止,需人工检查详情页布局或模板。")
|
||||
raise RuntimeError("TeLaiDian: 价格入口模板匹配失败")
|
||||
|
||||
logger.info(f"[详情页] 通过 jgxx.jpg 成功找到价格入口,点击坐标: {entrance_point},即将进入电价页")
|
||||
d.click(int(entrance_point[0]), int(entrance_point[1]))
|
||||
await asyncio.sleep(WAIT_DETAIL_PAGE_LOAD)
|
||||
|
||||
# 4. 进入分时电价页面后的处理
|
||||
try:
|
||||
logger.info("已点击进入价格详情,等待小程序自动滚动定位完成 (4秒)...")
|
||||
logger.info("[电价页] 已点击价格入口,等待小程序自动滚动定位完成 (4秒)...")
|
||||
await asyncio.sleep(4.0)
|
||||
|
||||
logger.info("执行两次大幅向下拉动,确保回到 00:00 时段顶部...")
|
||||
for _ in range(2):
|
||||
logger.info("[电价页] 开始执行两次大幅向下拉动,目标是回到 00:00 时段顶部")
|
||||
for idx in range(2):
|
||||
logger.info(f"[电价页] 第 {idx+1} 次向下滑动(d.swipe_ext(\"down\", scale=0.9))")
|
||||
d.swipe_ext("down", scale=0.9)
|
||||
await asyncio.sleep(1.0)
|
||||
except Exception as e:
|
||||
logger.error(f"处理分时电价页面初始状态失败: {e}")
|
||||
logger.error(f"[电价页] 处理分时电价页面初始状态失败: {e}")
|
||||
|
||||
# 5. 循环滑动抓取完整分时电价
|
||||
all_prices = []
|
||||
@@ -277,7 +275,7 @@ class TeLaiDianCrawler(BaseCrawler):
|
||||
screenshot_tasks = [] # 用于异步分析图片的任务列表
|
||||
temp_screenshots = [] # 记录临时文件以便后续清理
|
||||
|
||||
logger.info("开始循环截图(UI操作优先,后台并行分析)...")
|
||||
logger.info("[电价页] 开始循环截图(UI操作优先,后台并行分析)...")
|
||||
|
||||
try:
|
||||
while price_page_count < max_price_pages:
|
||||
|
||||
Binary file not shown.
Binary file not shown.
Reference in New Issue
Block a user