|
| 1 | +from selenium.webdriver.support import expected_conditions as EC |
| 2 | +from selenium.webdriver.common.action_chains import ActionChains |
| 3 | +from webdriver_manager.chrome import ChromeDriverManager |
| 4 | +from webdriver_manager.firefox import GeckoDriverManager |
| 5 | +from selenium.webdriver.support.ui import WebDriverWait |
| 6 | +from selenium.webdriver.common.keys import Keys |
| 7 | +from selenium.webdriver.common.by import By |
| 8 | +from selenium import webdriver |
| 9 | +from bs4 import BeautifulSoup |
| 10 | + |
| 11 | + |
| 12 | +chrome_options = webdriver.ChromeOptions() |
| 13 | +chrome_options.add_argument("--incognito") |
| 14 | +chrome_options.add_argument("--start-maximized") |
| 15 | +chrome_options.add_argument("--enable-automation") |
| 16 | +chrome_options.add_argument("--disable-extensions") |
| 17 | +chrome_options.add_argument("--disable-popup-blocking") |
| 18 | +self.driver = webdriver.Chrome(ChromeDriverManager().install(),options=chrome_options) |
| 19 | + |
| 20 | + |
| 21 | + |
| 22 | +webdriver.ActionChains(self.driver).send_keys(self.keyword).send_keys(Keys.ENTER).perform() |
| 23 | +webdriver.ActionChains(self.driver).send_keys(Keys.ESCAPE).perform() |
| 24 | +webdriver.ActionChains(self.driver).send_keys(Keys.ENTER).perform() |
| 25 | + |
| 26 | +html = self.driver.page_source |
| 27 | +soup = BeautifulSoup(html,'html5lib') |
| 28 | + |
| 29 | +searchList = soup.select("div[class*="+ "'" + some_name + "'" +"]") |
| 30 | +searchList = soup.find_all('div',attrs={'class': 'some_name'}) |
| 31 | +for s in searchList: |
| 32 | + a = s.find('div',attrs={'class': 'some_name'}) |
| 33 | + if a : |
| 34 | + a = a.get_text() |
| 35 | + |
| 36 | + |
| 37 | +WebDriverWait(self.driver, 60).until(EC.presence_of_element_located((By.CLASS_NAME, self.elem['chatlist']))) |
| 38 | +javaScriptHeight = "return document.getElementById('pane-side').scrollHeight;" |
| 39 | +# Scroll of Div elem |
| 40 | +height = int(self.driver.execute_script(javaScriptHeight)) |
| 41 | +for i in range(0,height,1000): |
| 42 | + javaScriptScroll = "document.getElementById('pane-side').scrollTo(0,"+i+");" |
| 43 | + self.driver.execute_script(javaScriptScroll) |
| 44 | + |
| 45 | + |
| 46 | +# for dynamic nature |
| 47 | +thing = 'cool' |
| 48 | +some = self.driver.find_element_by_xpath("//div[@class="+"'"+thing+"'"+"]").text |
| 49 | +self.driver.find_element_by_xpath("div[@class="+"'"+ok+"'"+"]").click() |
| 50 | +self.driver.find_element_by_xpath("//input[@accept='*']").send_keys(doc) |
| 51 | +infoBar = self.driver.find_element_by_xpath("//a[@interactionclass=""'" +info+"'" "]") |
| 52 | + |
| 53 | +# for classes with same name in first half |
| 54 | +# e.g ZZZ a , ZZZ b ZZZ c |
| 55 | +variable = 'ZZZ ' |
| 56 | +selected = soup.find_all('div', |
| 57 | + { |
| 58 | + 'class': lambda x: x |
| 59 | + and variable in x.split() |
| 60 | + } |
| 61 | +) |
| 62 | + |
| 63 | + |
| 64 | +regex = re.compile('^.*copyable-text.*$') |
| 65 | +text = s.find('div',attrs={'class': regex}) |
| 66 | + |
| 67 | + |
| 68 | +# For scrolling up an div elem window |
| 69 | +scrollUp = "document.getElementsByClassName("+"'"+scroll+"'"+")[0].scroll(0,-500);" |
| 70 | +last_height = self.driver.execute_script(javaScript) |
| 71 | +while True: |
| 72 | + try: |
| 73 | + self.driver.execute_script(up) |
| 74 | + time.sleep(SCROLL_PAUSE_TIME) |
| 75 | + new_height = self.driver.execute_script(javaScript) |
| 76 | + if new_height == last_height : |
| 77 | + break |
| 78 | + last_height = new_height |
| 79 | + except : |
| 80 | + break |
| 81 | + |
| 82 | + |
| 83 | +self.driver.execute_script("window.onbeforeunload = function() {};") |
| 84 | +self.driver.execute_script("window.open('"+url+"', '_self')") |
| 85 | +self.driver.implicitly_wait(2) |
| 86 | + |
| 87 | +videosXpath = self.driver.find_elements_by_xpath("//*[@class='abc']") |
| 88 | +for v in videosXpath: |
| 89 | + video = v.get_attribute("src") |
| 90 | + fileName = 'abc.mp4' |
| 91 | + r = requests.get(url) |
| 92 | + if r.status_code == 200: |
| 93 | + with open(fileName, 'wb') as f: |
| 94 | + f.write(r.content) |
| 95 | + |
| 96 | +height = str(self.driver.execute_script("return document.body.scrollHeight")) |
| 97 | +height = "-"+height |
| 98 | +js = "window.scrollTo(0, "+height+");" |
| 99 | +self.driver.execute_script(js) |
| 100 | + |
| 101 | + |
| 102 | +Photo = soup.find('img',attrs = {'class' : 'PhotoImg photo img'}) |
| 103 | +if Photo: |
| 104 | + Photo = Photo.get('src') |
| 105 | + path = self.filePathPhotos + os.sep + self.name + '_Photo.jpg' |
| 106 | + urllib.urlretrieve(Photo,path ) |
| 107 | + print('File saved at = {}'.format(path)) |
| 108 | + |
| 109 | + |
| 110 | +# Helps to scroll down Fast |
| 111 | +def scollEnd(self): |
| 112 | + SCROLL_PAUSE_TIME = 4 |
| 113 | + last_height = self.driver.execute_script("return document.body.scrollHeight") |
| 114 | + while True: |
| 115 | + self.driver.execute_script("window.scrollTo(0, document.body.scrollHeight);") # Scroll down to bottom |
| 116 | + time.sleep(SCROLL_PAUSE_TIME) # Wait to load page |
| 117 | + new_height = self.driver.execute_script("return document.body.scrollHeight") |
| 118 | + print('Old height = {} , New Height = {}'.format(last_height,new_height)) |
| 119 | + if new_height == last_height or self.terminate : |
| 120 | + return |
| 121 | + last_height = new_height |
| 122 | + |
| 123 | + |
| 124 | +# Helps to scroll down smoothly |
| 125 | +def scrollEnd(self): |
| 126 | + SCROLL_PAUSE_TIME = 4 |
| 127 | + # Get scroll height |
| 128 | + # last_height = self.driver.execute_script("return document.body.scrollHeight") |
| 129 | + last_height = self.driver.execute_script("return window.pageYOffset") |
| 130 | + height = 350 |
| 131 | + while True: |
| 132 | + down = str(height) |
| 133 | + js = "window.scrollTo(0, "+down+");" |
| 134 | + self.driver.execute_script(js) |
| 135 | + height += 350 |
| 136 | + time.sleep(SCROLL_PAUSE_TIME) |
| 137 | + new_height = self.driver.execute_script("return window.pageYOffset") |
| 138 | + if new_height == last_height : |
| 139 | + break |
| 140 | + last_height = new_height |
| 141 | + |
0 commit comments