Spaces:
Sleeping
Sleeping
Upload scrape_fake_app.py
Browse files- scrape_fake_app.py +37 -29
scrape_fake_app.py
CHANGED
@@ -1,6 +1,7 @@
|
|
1 |
from playwright.sync_api import sync_playwright
|
2 |
import os
|
3 |
from dotenv import load_dotenv
|
|
|
4 |
|
5 |
def load_credentials()-> dict:
|
6 |
load_dotenv()
|
@@ -45,39 +46,46 @@ def extract_timetable_text(page):
|
|
45 |
|
46 |
# print(URL,USERNAME,PASSWORD)
|
47 |
def get_homework() -> str:
|
48 |
-
|
49 |
-
|
50 |
-
with sync_playwright() as playwright:
|
51 |
-
browser = playwright.firefox.launch(headless=True)
|
52 |
-
page = browser.new_page()
|
53 |
-
page.goto(URL,wait_until="domcontentloaded")
|
54 |
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
63 |
|
64 |
def get_timetable() -> str:
|
65 |
-
|
66 |
-
|
67 |
-
with sync_playwright() as playwright:
|
68 |
-
browser = playwright.firefox.launch(headless=True)
|
69 |
-
page = browser.new_page()
|
70 |
-
page.goto(URL,wait_until="domcontentloaded")
|
71 |
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
81 |
|
82 |
if __name__=="__main__":
|
83 |
print(get_homework())
|
|
|
1 |
from playwright.sync_api import sync_playwright
|
2 |
import os
|
3 |
from dotenv import load_dotenv
|
4 |
+
import json
|
5 |
|
6 |
def load_credentials()-> dict:
|
7 |
load_dotenv()
|
|
|
46 |
|
47 |
# print(URL,USERNAME,PASSWORD)
|
48 |
def get_homework() -> str:
|
49 |
+
try:
|
50 |
+
URL,USERNAME,PASSWORD=load_credentials()
|
|
|
|
|
|
|
|
|
51 |
|
52 |
+
with sync_playwright() as playwright:
|
53 |
+
browser = playwright.firefox.launch(headless=True)
|
54 |
+
page = browser.new_page()
|
55 |
+
page.goto(URL,wait_until="domcontentloaded")
|
56 |
+
|
57 |
+
page.get_by_role('textbox',name='username').fill(USERNAME)
|
58 |
+
page.get_by_role('textbox',name='password').fill(PASSWORD)
|
59 |
+
page.get_by_role('button',name='login').click()
|
60 |
+
page.wait_for_url("**/dashboard")
|
61 |
+
# page.wait_for_timeout(1000)
|
62 |
+
homework = extract_homework_text(page)
|
63 |
+
browser.close()
|
64 |
+
return homework
|
65 |
+
|
66 |
+
except Exception as e:
|
67 |
+
return json.dumps(e)
|
68 |
|
69 |
def get_timetable() -> str:
|
70 |
+
try:
|
71 |
+
URL,USERNAME,PASSWORD=load_credentials()
|
|
|
|
|
|
|
|
|
72 |
|
73 |
+
with sync_playwright() as playwright:
|
74 |
+
browser = playwright.firefox.launch(headless=True)
|
75 |
+
page = browser.new_page()
|
76 |
+
page.goto(URL,wait_until="domcontentloaded")
|
77 |
+
|
78 |
+
page.get_by_role('textbox',name='username').fill(USERNAME)
|
79 |
+
page.get_by_role('textbox',name='password').fill(PASSWORD)
|
80 |
+
page.get_by_role('button',name='login').click()
|
81 |
+
page.wait_for_url("**/dashboard")
|
82 |
+
# page.wait_for_timeout(1000)
|
83 |
+
timetable = extract_timetable_text(page)
|
84 |
+
browser.close()
|
85 |
+
return timetable
|
86 |
+
|
87 |
+
except Exception as e:
|
88 |
+
return json.dumps(e)
|
89 |
|
90 |
if __name__=="__main__":
|
91 |
print(get_homework())
|