kw_tools/choozmo/gen_seo.py

#import redis
import time
import traceback
#import json
from selenium import webdriver
from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
import time
import os
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.common.by import By
from selenium.webdriver.support import expected_conditions as EC
import dataset
from selenium.webdriver.common.keys import Keys
import json
import random
import time
import redis
import sys
import codecs
import random
import os
import time
from userAgentRandomizer import userAgents

driver=None

db = dataset.connect('mysql://choozmo:pAssw0rd@db.ptt.cx:3306/seo?charset=utf8mb4')


def re_get_webdriver():
    global driver
    result=[]
    if driver is not None:
        print('closing....')
        driver.quit()
        os.system('killall chrome')
        print('quit....')
        driver=None
    try:
        ua = userAgents()

        user_agent = ua.random()        

        options = webdriver.ChromeOptions()
        options.add_argument("--no-sandbox")
        options.add_argument("--disable-dev-shm-usage")
        options.add_argument("--headless")

        print(user_agent)
#        options.add_argument("--user-agent=" +user_agent)
        options.add_argument("--incognito")
        driver=None
        try:
            driver = webdriver.Chrome(options=options)
        except:
#            driver.quit()
#            os.system('pkill -f ')
            os.system('kill %d' % os.getpid())
            sys.exit()
            return
        driver.set_window_size(1400,1000)
        return
    except:
        import traceback
        traceback.print_exc()
        driver=None
        return None



def from_shopping(kw):
    global driver
    driver.get('https://shopping.google.com')
    time.sleep(5)
    elmt = driver.find_element(By.XPATH, "//input[@id='REsRA']")
    elmt.send_keys('幸福空間') 
    elmt.send_keys(Keys.ENTER) #hits space
    time.sleep(7)
    elmt = driver.find_element(By.XPATH, "//div[@class='hdtb-mitem']/a[contains(text(),'全部') or contains(text(),'All')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(5)

def from_book(kw):
    global driver
    driver.get('https://books.google.com/')
    time.sleep(5)
    elmt = driver.find_element(By.XPATH, "//input[@id='oc-search-input']")
    elmt.send_keys('幸福空間') 
    elmt.send_keys(Keys.ENTER) #hits space
    time.sleep(7)
    elmt = driver.find_element(By.XPATH, "//div[@class='hdtb-mitem']/a[contains(text(),'全部') or contains(text(),'All')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(5)


def from_wiki(kw):
    global driver
    driver.get('https://en.wikipedia.org/wiki/Google_Search')
    time.sleep(4)
    elmt = driver.find_element(By.XPATH, "//a[contains(@href,'https://www.google.com/')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(5)
    elmt = driver.find_element(By.XPATH, "//input[@class='gLFyf gsfi']")
    time.sleep(1)
    elmt.send_keys(kw)
    elmt.send_keys(Keys.ENTER)
    time.sleep(6)

def from_bing(kw):
    global driver
    driver.get('https://www.bing.com/search?q=google')
    time.sleep(4)
    elmt = driver.find_element(By.XPATH, "//a[contains(@href,'https://www.google.com/')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(5)
    elmt = driver.find_element(By.XPATH, "//input[@class='gLFyf gsfi']")
    time.sleep(1)
    elmt.send_keys(kw)
    elmt.send_keys(Keys.ENTER)
    time.sleep(6)

def from_ecosia(kw):
    global driver
    driver.get('https://www.ecosia.org/search?method=index&q=GOOGLE')
    time.sleep(4)
    elmt = driver.find_element(By.XPATH, "//a[contains(@href,'https://www.google.com/')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(5)
    elmt = driver.find_element(By.XPATH, "//input[@class='gLFyf gsfi']")
    time.sleep(1)
    elmt.send_keys(kw)
    elmt.send_keys(Keys.ENTER)
    time.sleep(6)

def from_brave(kw):
    global driver
    driver.get('https://search.brave.com/search?q=google&source=web')
    time.sleep(4)
    elmt = driver.find_element(By.XPATH, "//a[contains(@href,'https://www.google.com/')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(5)
    elmt = driver.find_element(By.XPATH, "//input[@class='gLFyf gsfi']")
    time.sleep(1)
    elmt.send_keys(kw)
    elmt.send_keys(Keys.ENTER)
    time.sleep(6)





def from_duckduckgo(kw):
    global driver
    driver.get('https://duckduckgo.com/?q=google')
    time.sleep(4)
    elmt = driver.find_element(By.XPATH, "//a[contains(@href,'https://www.google.com/')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(5)
#    time.sleep(9999)
    elmt = driver.find_element(By.XPATH, "//input[@class='gLFyf gsfi']")
    time.sleep(1)
    elmt.send_keys(kw)
    elmt.send_keys(Keys.ENTER)
    time.sleep(6)


def from_ekoru(kw):
    global driver
    driver.get('https://www.ekoru.org/?q=google')
    time.sleep(4)
    elmt = driver.find_element(By.XPATH, "//a[contains(@href,'https://www.google.com/')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(5)
#    time.sleep(9999)
    elmt = driver.find_element(By.XPATH, "//input[@class='gLFyf gsfi']")
    time.sleep(1)
    elmt.send_keys(kw)
    elmt.send_keys(Keys.ENTER)
    time.sleep(6)





def from_yahoo(kw):
    global driver
    driver.get('https://search.yahoo.com/search?p=google')
    time.sleep(4)
    elmt = driver.find_element(By.XPATH, "//a[contains(@href,'https://www.google.com/')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(2)
    driver.switch_to.window(driver.window_handles[1])
    time.sleep(3)
    print(driver.current_url)
    elmt = driver.find_element(By.XPATH, "//input[@name='q']")
    time.sleep(1)
    elmt.send_keys(kw)
    elmt.send_keys(Keys.ENTER)
    time.sleep(6)

def from_gibiru(kw):
    global driver
    driver.get('https://gibiru.com/results.html?q=google')
    time.sleep(4)
    elmt = driver.find_element(By.XPATH, "//a[contains(@href,'https://www.google.com/')]")
    webdriver.ActionChains(driver).move_to_element(elmt).perform()
    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
    time.sleep(2)
    driver.switch_to.window(driver.window_handles[1])
    time.sleep(3)
    print(driver.current_url)
    elmt = driver.find_element(By.XPATH, "//input[@name='q']")
    time.sleep(1)
    elmt.send_keys(kw)
    elmt.send_keys(Keys.ENTER)
    time.sleep(6)



def run_once(jsobj):

    table=db['rank_detection']
    print(jsobj)
    global driver


#    i=random.randint(0,9)
    i=100
    if driver is None:
        time.sleep(8)
        re_get_webdriver()
    if driver is None:
        return
    try:
        kw=jsobj['kw']
        domain=jsobj['domain']
#        kw='leo 娛樂城 484'
#        kw='leo 娛樂城 3011'

#        domain='tha484.com'
#        domain='tha3011.com'


        if i==0:
            from_book(kw)
        elif i==1:
            from_shopping(kw)
        elif i==2:
            from_wiki(kw)
        elif i==3:
            from_bing(kw)
        elif i==4:
            from_duckduckgo(kw)
        elif i==5:
            from_yahoo(kw)
        elif i==6:
            from_gibiru(kw)
        elif i==7:
            from_ekoru(kw)
        elif i==8:
            from_ecosia(kw)
        elif i==9:
            from_brave(kw)
#        time.sleep(9999)

        driver.get('https://www.google.com?num=100')
        time.sleep(3)
        print(driver.current_url)
        elmt = driver.find_element(By.XPATH, "//input[@name='q']")
        time.sleep(1)
        elmt.send_keys(kw)
        elmt.send_keys(Keys.ENTER)
        time.sleep(6)

        elmts = driver.find_elements(By.XPATH, "//div[@class='yuRUbf']/a")

        numresults=len(elmts)
        print('搜尋結果數量',numresults)

        idx=1
        found=False
        for elmt in elmts:
            href=elmt.get_attribute('href')
            txt=elmt.text
            if len(txt)>10:
                if domain in href:
                    print('found....')
                    print('clicked....')
                    print(href)
                    print(txt)
                    print("ranking", idx)
                    table.insert({'ranking':idx,'kw':kw,'results':numresults,'url':href,'title':txt})
                    found=True
                    webdriver.ActionChains(driver).move_to_element(elmt).perform()
                    webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
                    time.sleep(5)
                    break
            idx+=1
        if not found:
            table.insert({'ranking':-1,'kw':kw,'results':numresults,'url':'','title':'未收錄'})

    except:
        print('exception')
        traceback.print_exc()

    driver.quit()
    sys.exit()



r = redis.Redis(host='db.ptt.cx', port=6379, db=2,password='choozmo9')

##data=r.get('personal_seo')
#jstext=data.decode('utf-8')
#jsobj=json.loads(jstext)

#js=random.choice(jsobj)
js=['seo','台北','新北','新竹','竹北','台灣','最強','集仕多','新聞','是什麼','搜尋','優化','如何','元宇宙','加速','排名','查詢','關鍵字','計劃','曝光','推薦','工具','google','排行','排序','公司','提升','收費','行情','網站','網頁','youtube','計畫','AI','人工智慧','deep learning','深度學習','評分','研究','價格','工具','論壇','自然','規則','流量','建議','寫作','技巧','課程','測試','因素','改善','購買','谷歌','成本','推廣','人員','方式','行銷','外貿','企業','電商','電子商務','商務','改版','分析','老師','講師','顧問','提高','影片','主播']

elmt1=random.choice(js)
elmt2=random.choice(js)

run_once({'domain':'choozmo.com','kw':elmt1+" "+elmt2+" choozmo"})