spider/EasySpider
1
0
Fork 0
mirror of https://github.com/NaiboWang/EasySpider.git synced 2025-04-22 15:06:32 +08:00
Code Issues Packages Projects Releases Wiki Activity

0.3.5收尾

Browse Source
This commit is contained in:
naibo 2023-07-10 09:59:59 +08:00
parent 8c5267d66c
commit adcfc95057
1 changed files with 24 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
ExecuteStage/easyspider_executestage.py
View File

@ -281,7 +281,10 @@ class BrowserThread(Thread):
except: except:
self.Log('Time out after set seconds when scrolling. ') self.Log('Time out after set seconds when scrolling. ')
self.recordLog('Time out after set seconds when scrolling') self.recordLog('Time out after set seconds when scrolling')
try:
self.browser.execute_script('window.stop()') self.browser.execute_script('window.stop()')
except:
pass
if scrollType != 0 and para["scrollCount"] > 0: # 控制屏幕向下滚动 if scrollType != 0 and para["scrollCount"] > 0: # 控制屏幕向下滚动
for i in range(para["scrollCount"]): for i in range(para["scrollCount"]):
self.Log("Wait for set second after screen scrolling") self.Log("Wait for set second after screen scrolling")
@ -679,7 +682,10 @@ class BrowserThread(Thread):
# 切换历史记录等待： # 切换历史记录等待：
self.Log("Change history back time or:", self.Log("Change history back time or:",
node["parameters"]["historyWait"]) node["parameters"]["historyWait"])
try:
self.browser.execute_script('window.stop()') self.browser.execute_script('window.stop()')
except:
pass
if int(node["parameters"]["breakMode"]) > 0: # 如果设置了退出循环的脚本条件 if int(node["parameters"]["breakMode"]) > 0: # 如果设置了退出循环的脚本条件
output = self.execute_code(int( output = self.execute_code(int(
node["parameters"]["breakMode"]) - 1, node["parameters"]["breakCode"], node["parameters"]["breakCodeWaitTime"], iframe=node["parameters"]["iframe"]) node["parameters"]["breakMode"]) - 1, node["parameters"]["breakCode"], node["parameters"]["breakCodeWaitTime"], iframe=node["parameters"]["iframe"])
@ -724,7 +730,10 @@ class BrowserThread(Thread):
# time.sleep(2) # time.sleep(2)
self.Log("Change history back time or:", self.Log("Change history back time or:",
node["parameters"]["historyWait"]) node["parameters"]["historyWait"])
try:
self.browser.execute_script('window.stop()') self.browser.execute_script('window.stop()')
except:
pass
except NoSuchElementException: except NoSuchElementException:
print("Loop element not found: ", path) print("Loop element not found: ", path)
print("找不到循环元素: ", path) print("找不到循环元素: ", path)
@ -997,7 +1006,10 @@ class BrowserThread(Thread):
self.history["index"] = self.browser.execute_script( self.history["index"] = self.browser.execute_script(
"return history.length") "return history.length")
except TimeoutException: except TimeoutException:
try:
self.browser.execute_script('window.stop()') self.browser.execute_script('window.stop()')
except:
pass
self.history["index"] = self.browser.execute_script( self.history["index"] = self.browser.execute_script(
"return history.length") "return history.length")
else: else:
@ -1005,7 +1017,10 @@ class BrowserThread(Thread):
self.history["index"] = self.browser.execute_script( self.history["index"] = self.browser.execute_script(
"return history.length") "return history.length")
except TimeoutException: except TimeoutException:
try:
self.browser.execute_script('window.stop()') self.browser.execute_script('window.stop()')
except:
pass
self.history["index"] = self.browser.execute_script( self.history["index"] = self.browser.execute_script(
"return history.length") "return history.length")
# 如果打开了新窗口，切换到新窗口 # 如果打开了新窗口，切换到新窗口
@ -1277,7 +1292,10 @@ class BrowserThread(Thread):
self.Log('Time out after set seconds when getting data') self.Log('Time out after set seconds when getting data')
self.recordLog( self.recordLog(
'Time out after set seconds when getting data') 'Time out after set seconds when getting data')
try:
self.browser.execute_script('window.stop()') self.browser.execute_script('window.stop()')
except:
pass
if p["relative"]: # 是否相对xpath if p["relative"]: # 是否相对xpath
if p["relativeXPath"] == "": # 相对xpath有时候就是元素本身，不需要二次查找 if p["relativeXPath"] == "": # 相对xpath有时候就是元素本身，不需要二次查找
element = loopElement element = loopElement