naufalnashif commited on
Commit
a71e00a
·
1 Parent(s): 98dbc9e

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +10 -25
app.py CHANGED
@@ -67,41 +67,26 @@ def scrape_klikindomaret(nama_barang, num_items):
67
  @st.cache_data
68
  def scrape_shopee(nama_barang, num_items):
69
  products = []
70
-
71
-
72
- #path = ''
73
-
74
- #Customize chrome display
75
- #chrome_options = Options()
76
- #chrome_options.add_argument('--no-sandbox')
77
- #chrome_options.add_argument('--headless')
78
- #chrome_options.add_argument('--disable-notifications')
79
- #chrome_options.add_argument('--disable-infobars')
80
 
81
- #Customize chrome display huggingface
82
- options = webdriver.ChromeOptions()
83
  options.add_argument('--no-sandbox')
84
  options.add_argument('--headless')
85
  options.add_argument('--disable-notifications')
86
  options.add_argument('--disable-infobars')
87
  options.add_argument('--disable-dev-shm-usage')
 
 
88
 
89
 
90
  try :
91
- page = 1
92
- query = quote(nama_barang)
93
- #driver = webdriver.Chrome(executable_path = path, options = chrome_options)
94
- driver = webdriver.Chrome(options = options)
95
  url = f'https://shopee.co.id/search?keyword={query}&page={page}'
96
  driver.get(url)
97
 
98
-
99
- # Cari elemen berdasarkan tagname HTML
100
- html_element = driver.find_element(By.TAG_NAME, "html")
101
 
102
- # Dapatkan HTML dari elemen
103
- html = html_element.get_attribute("innerHTML")
104
- #html = driver.execute_script("return document.getElementsByTagName('html')[0].innerHTML")
105
  soup = BeautifulSoup(html, "html.parser")
106
 
107
  for i in soup.find_all('div', class_ = "ie3A+n bM+7UW Cve6sh" ):
@@ -206,15 +191,15 @@ def scrape_tokped_with_selenium(nama_barang, num_items):
206
  products = []
207
  page = 1
208
  query = quote(nama_barang)
209
- driver = webdriver.Chrome(options = options)
210
 
211
-
212
- options = webdriver.ChromeOptions()
213
  options.add_argument('--no-sandbox')
214
  options.add_argument('--headless')
215
  options.add_argument('--disable-notifications')
216
  options.add_argument('--disable-infobars')
217
  options.add_argument('--disable-dev-shm-usage')
 
 
218
 
219
  while len(products) < num_items :
220
  try :
 
67
  @st.cache_data
68
  def scrape_shopee(nama_barang, num_items):
69
  products = []
70
+ page = 1
71
+ query = quote(nama_barang)
 
 
 
 
 
 
 
 
72
 
73
+ options = webdriver.FirefoxOptions()
 
74
  options.add_argument('--no-sandbox')
75
  options.add_argument('--headless')
76
  options.add_argument('--disable-notifications')
77
  options.add_argument('--disable-infobars')
78
  options.add_argument('--disable-dev-shm-usage')
79
+
80
+ driver = webdriver.Firefox(options=options)
81
 
82
 
83
  try :
84
+
 
 
 
85
  url = f'https://shopee.co.id/search?keyword={query}&page={page}'
86
  driver.get(url)
87
 
88
+ html = driver.page_source
 
 
89
 
 
 
 
90
  soup = BeautifulSoup(html, "html.parser")
91
 
92
  for i in soup.find_all('div', class_ = "ie3A+n bM+7UW Cve6sh" ):
 
191
  products = []
192
  page = 1
193
  query = quote(nama_barang)
 
194
 
195
+ options = webdriver.FirefoxOptions()
 
196
  options.add_argument('--no-sandbox')
197
  options.add_argument('--headless')
198
  options.add_argument('--disable-notifications')
199
  options.add_argument('--disable-infobars')
200
  options.add_argument('--disable-dev-shm-usage')
201
+
202
+ driver = webdriver.Firefox(options=options)
203
 
204
  while len(products) < num_items :
205
  try :