temporary

website_clickjobs/type-1/_clickjob1023.py

@@ -0,0 +1,177 @@
+import time
+from datetime import datetime
+import json
+from selenium import webdriver
+from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
+import time
+import os
+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
+import requests
+import dataset
+import traceback
+import sys
+from selenium.webdriver.common.keys import Keys
+import timeit
+import socket
+
+add_tabs = [7,9,11,13,15,7,9,11,13,15,7,9,11,13,15,7,9,11,13,15]
+
+db = dataset.connect('mysql://choozmo:pAssw0rd@db.ptt.cx:3306/seo?charset=utf8mb4')
+driver=None
+headers = {
+        "Authorization": "Bearer " + "t35vhZtWNgvDNWHc3DJh0OKll3mcB9GvC8K2EAkBug2",
+        "Content-Type": "application/x-www-form-urlencoded"
+}
+
+sleepoffset = 0
+
+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(domain, target_domain, brands, query):
+    print(query)
+    sleepoffset = 0
+    global driver
+    if query == "艾立思" and "index" in target_domain:
+        driver.get('https://www.google.com/search?num=100&q=艾立思&rlz=1C1ONGR_zh-TWTW997TW997&ei=zjdUY_DBG9Lm-Abpgq84&start=0&sa=N&filter=0&ved=2ahUKEwjw4KeEvfT6AhVSM94KHWnBCwcQ8tMDegQIARAQ&cshid=1666463754367857&biw=1368&bih=761&dpr=2')
+        time.sleep(4)
+    else:
+        driver.get('https://www.google.com?num=100')
+        time.sleep(3)
+        print(driver.current_url)
+
+        # elmts=driver.find_elements_by_xpath("//div[@class='yuRUbf']/a")
+        # ABOVE METHOD IS DEPRECATED STARTING SELENIUM 4.3.0, USE THIS
+        #
+        elmt = driver.find_element(By.XPATH, "//input[@name='q']")
+        time.sleep(1)
+
+        elmt.send_keys(query)
+        elmt.send_keys(Keys.ENTER)
+
+    idx=1
+    ranking=-1
+    domain_in_link = 0
+
+    googleurl = driver.current_url
+    print(driver.current_url)
+
+    if "sorry" in googleurl:
+        return 444
+
+    elmts=driver.find_elements("xpath","//div[@class='yuRUbf']/a")
+
+    print (len(elmts))
+    # driver.save_screenshot('c:/tmp/test.png')
+
+    for el in elmts:
+        href=el.get_attribute('href')
+        txt=el.text
+        print(href)
+        if len(txt)>10:
+            if domain in href:
+                domain_in_link += 1
+                print('clicked....')
+                print(href)
+                print(txt)
+
+                if query == "艾立思" and "index" in target_domain and href != "https://hhh.com.tw/brand-index.php?brand_id=211":
+                    print("wrong site")
+                    continue
+                
+                webdriver.ActionChains(driver).move_to_element(el).perform()
+                webdriver.ActionChains(driver).move_to_element(el).click().perform()
+                time.sleep(15)
+
+                new_windows_count = add_tabs[random.randint(0,19)]
+                print(str(new_windows_count) + " new tabs")
+                for i in range (0,new_windows_count):
+                    print("Tab " + str(i+1))
+                    #original_window = driver.current_window_handle
+                    #driver.switch_to.new_window('window')
+                    #driver.get(href)
+                    sleepoffset += 12
+                    driver.execute_script('window.open("'+href+'","_blank");')
+                    driver.execute_script("window.scrollTo(0, 600)")
+                    time.sleep(15)
+                    #driver.close()
+                    #driver.switch_to.window(original_window)
+
+                if domain in target_domain:
+                    print("Target link found")
+                    time_stamp = datetime.fromtimestamp(time.time())
+                    time_stamp = time_stamp.strftime("%Y-%m-%d %H:%M:%S")
+                    db['click_results'].insert({"time_stamp": time_stamp, "brand": brands[domain], "domain": domain, "query": query, "url": href, "content": txt, "extra_windows": new_windows_count})
+                break
+
+    if domain in target_domain:
+        print("Target domain found")
+        time_stamp = datetime.fromtimestamp(time.time())
+        time_stamp = time_stamp.strftime("%Y-%m-%d %H:%M:%S")
+        db['query_results'].insert({"time_stamp": time_stamp, "brand": brands[domain], "domain": domain, "query": query, "googleurl": googleurl, "element_count": len(elmts), "domain_in_link_count": domain_in_link})
+            
+
+    print(domain_in_link)
+    return 200
+    
+
+def run_once(domain, target_domain, brands, query):
+    global driver
+    result=[]
+    options = webdriver.ChromeOptions()
+    options.add_argument('--headless')
+#    options.add_argument("--user-agent=" +user_agent)
+    options.add_argument("--incognito")
+    options.add_argument('--no-sandbox')
+    options.add_argument('--disable-dev-shm-usage')
+
+    driver = webdriver.Chrome(
+    options=options)
+
+    driver.delete_all_cookies()
+    driver.set_window_size(1400,1000)
+
+    statuscode = process_query(domain, target_domain, brands, query)
+    driver.quit()
+
+    return statuscode
+
+#execution starts here
+
+def execute(domain, target_domain, brands, query_list):
+    print("Ctrl+C or Ctrl+Z to stop.")
+    statuscode = 0
+    st = timeit.default_timer()
+    try:
+        statuscode = run_once(domain, target_domain, brands, random.choice(query_list))
+    except:
+        traceback.print_exc()
+    timetaken = timeit.default_timer()-st
+    print("Time taken: " + str(timetaken))
+    
+    print("Process returned with " + str(statuscode))
+    if statuscode == 444:
+        print("You have been caught!!! Program terminating.")
+        
+        #notify("Clickbot " + brands[domain] + " has been caught by Google and will terminate. IP: ")
+        return statuscode
+
+    extrasleep = 0
+    if(timetaken < 70):
+        extrasleep = 70 - timetaken
+    print("Ctrl+C or Ctrl+Z to stop now.")
+    print("You have " + str(10 + extrasleep) + " seconds.")
+    time.sleep(10 + extrasleep)
+    return statuscode

website_clickjobs/type-1/hhh_e2.py

@@ -0,0 +1,55 @@
+from _clickjob1023 import *
+import xmltodict
+
+def get_xml_data(urls):
+    data = []
+    if len(priority)!=0:
+        for p in priority:
+            link = 'https://hhh.com.tw/HHH_NEW/designers_static/' + str(p[0]) + '.html'
+            data.append('https://hhh.com.tw/HHH_NEW/designers_static/' + str(p[0]) + '.html')
+            p[0] = 'https://hhh.com.tw/HHH_NEW/designers_static/' + str(p[0]) + '.html'
+    else:
+        for u in urls:
+            page = requests.get(u)
+            designerdata = xmltodict.parse(page.content)['urlset']['url']
+            for d in designerdata:
+                data.append(d['loc']) 
+
+    if len(data)>=10:
+        print(data[:10])
+    else:
+        print(data)
+    return data
+
+sourceurls = ["https://hhh.com.tw/HHH_NEW/sitemap/sitemap.xml","https://m.hhh.com.tw/HHH_NEW/sitemap/sitemap.xml"]
+
+# PRIORITY IDs, clear list (make it become [] ) to disable priority
+priority = [['https://hhh.com.tw/brand-index.php?brand_id=216', '好萊得'],['https://hhh.com.tw/brand-product.php?brand_id=211', '艾立思'], ['https://hhh.com.tw/brand-index.php?brand_id=211', '艾立思']]
+#priority = [['https://hhh.com.tw/brand-index.php?brand_id=211', '艾立思']]
+#priority = [[54, '虞國綸', '格綸設計'], [793, '許伯争', '樂作空間設計'], 32]      # RESERVED -- LEVEL 2
+#priority = [[54, '虞國綸', '格綸設計'], [793, '許伯争', '樂作空間設計'], 32, 331] # RESERVED -- LEVEL 3
+
+domain = 'hhh.com.tw'
+#target_domain = get_xml_data(sourceurls)
+brands={domain:'hhh'}
+query_list = ('幸福空間')
+
+statuscode = 0
+
+while True:
+    if len(priority)!=0:
+        target = random.choice(priority)
+        ql = target[1:]
+        if random.randint(0, 3) % 2 == 3:
+            for q in range (len(ql)):
+                ql[q] = '幸福空間 ' + ql[q]
+        print(ql)
+        print(target[0])
+        statuscode = execute(domain, target[0], brands, ql)
+    else:
+        target = random.choice(target_domain)
+        #statuscode = execute(domain, target_domain, brands, query_list)
+        print(len(target_domain))
+    if statuscode == 444:
+        break
+time.sleep(60)