w

choozmo/gsearch_top.py

@@ -0,0 +1,170 @@
+from selenium import webdriver
+from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
+import time
+import os
+import datetime
+import urllib.parse
+from selenium.webdriver.support.ui import WebDriverWait
+from selenium.webdriver.common.by import By
+from selenium.webdriver.support import expected_conditions as EC
+import codecs
+import random
+from bs4 import BeautifulSoup
+import requests
+import time
+import rpyc
+import sys
+import docker
+import  googlesearch
+import codecs
+import sys
+import time
+import dataset
+import os
+
+
+db = dataset.connect('mysql://choozmo:pAssw0rd@db.ptt.cx:3306/hhh?charset=utf8mb4')
+
+#cursor=db.query('SELECT kw FROM hhh.hhh_contentgap_serp where ranking is not null;')
+#cursor=db.query('SELECT kw FROM hhh.hhh_contentgap_serp where kw not in (select distinct kw from hhh_contentgap_serp where id >= 155)')
+
+kwlst={}
+#for c in cursor:
+#    kwlst[c['kw']]=1
+
+
+
+
+
+table=db['hhh_top_serp']
+curdir=os.path.realpath('.')
+
+#fr=codecs.open(curdir+os.sep+'contentgap.txt','r','utf-8')
+#fr=codecs.open(curdir+os.sep+'hhh\\seo\\contentgap.txt','r','utf-8')
+#fr=codecs.open('C:\\gitlab\\kw_tools\\top.csv','r','utf-8')
+#lines=fr.readlines()
+lst=[]
+cursor=db.query('select term from hhh.content_top_terms where term not in (SELECT kw FROM hhh.hhh_top_serp where datediff(now(),dt) =0 and ranking is not null )')
+for c in cursor:
+    lst.append(c['term'])
+
+
+
+#for l in lines:
+
+#for l in lines[35:]:
+#for l in lines[49:]:
+#for l in lines[34:]:
+#for l in lines[41:]:
+#
+#    lst.append(l.replace('\n',''))
+
+
+
+
+
+headers = {
+        "Authorization": "Bearer " + "t35vhZtWNgvDNWHc3DJh0OKll3mcB9GvC8K2EAkBug2",
+        "Content-Type": "application/x-www-form-urlencoded"
+}
+
+def send_msg(kw):
+    params = {"message": "處理關鍵字: "+kw}  
+    r = requests.post("https://notify-api.line.me/api/notify",headers=headers, params=params)
+
+
+def empty_query(q):
+    global driver
+    googleurl='https://www.google.com/search?q='+urllib.parse.quote(q)
+    driver.get(googleurl)
+    time.sleep(3)
+
+
+def process_query(qs,number_results=10,language_code='en',pat='hhh.com.tw'):
+    global driver
+    escaped_search_term=urllib.parse.quote(qs)
+#    escaped_search_term = qs.replace(' ', '+')
+#    googleurl='https://www.google.com/search?q='+
+    googleurl = 'https://www.google.com/search?q={}&num={}&hl={}'.format(escaped_search_term, number_results+1,language_code)
+
+    driver.get(googleurl)
+    elmts=driver.find_elements_by_xpath("//div[@class='yuRUbf']/a")
+    idx=0
+    for elmt in elmts:
+        try:
+            href=elmt.get_attribute('href')
+            print(str(idx)+': '+href)
+            if pat in href:
+                return idx
+            idx+=1
+        except:
+            print('href exception')
+
+    try:
+        elmt=driver.find_element_by_xpath("//a[@id='pnnext']")
+        webdriver.ActionChains(driver).move_to_element(elmt).perform()
+        webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
+    except:
+        print('pnnext exception')
+        return None
+
+    time.sleep(4)
+    elmts=driver.find_elements_by_xpath("//div[@class='yuRUbf']/a")
+
+    for elmt in elmts:
+        try:
+            href=elmt.get_attribute('href')
+            print(str(idx)+': '+href)
+            if pat in href:
+                return idx
+            idx+=1
+
+        except:
+            print('href2 exception')
+
+
+
+result=[]
+driver=None
+
+def restart_browser():
+    client = docker.from_env()
+    ls=client.containers.list()
+    print(ls)
+    ls[0].restart()
+    time.sleep(10)
+
+#    options = webdriver.ChromeOptions()
+    #driver=webdriver.Chrome(desired_capabilities=options.to_capabilities())
+    driver = webdriver.Remote(
+        command_executor='http://127.0.0.1:4444/wd/hub',
+#        command_executor='http://172.104.93.163:4444/wd/hub', 
+    #command_executor='http://dev2.choozmo.com:14444/wd/hub',
+#    desired_capabilities=options.to_capabilities())
+    desired_capabilities=DesiredCapabilities.CHROME)
+    driver.set_window_size(1400,1000)
+    return driver
+
+for l in lst:
+#for l in lst[2:]:
+    if True:
+#    if kwlst.get(l) is None:
+        driver=restart_browser()
+
+    #    l='房間 油漆'
+    #    idx=process_query(,number_results=100,language_code='zh-TW',pat='hhh.com.tw')
+        idx=process_query(l,number_results=100,language_code='zh-TW',pat='hhh.com.tw')
+
+
+        if idx==None:
+            print(driver.page_source)
+            if '我們的系統偵測到您的電腦網路送出的流量有異常情況' in driver.page_source:
+                print('baned.....')
+                sys.exit()
+
+        table.insert({'kw':l,'ranking':idx,'dt':datetime.datetime.now()})
+        print({'kw':l,'ranking':idx})
+        db.commit()
+    #    time.sleep(9999)
+
+#        time.sleep(4)