Update fb parser
This commit is contained in:
parent
33fd3a4668
commit
4f8c054663
2 changed files with 5 additions and 3 deletions
|
@ -1,20 +1,21 @@
|
|||
import json
|
||||
import facebook_scraper
|
||||
import enum
|
||||
|
||||
FILE = 'facebook_data.json'
|
||||
|
||||
PAGES = ["amicale.deseleves", "campus.insat"]
|
||||
|
||||
|
||||
def scrape_data(page):
|
||||
post_list = []
|
||||
for post in facebook_scraper.get_posts(page, pages=3):
|
||||
for post in facebook_scraper.get_posts(page, pages=4):
|
||||
print(post)
|
||||
cleaned_post = {
|
||||
"id": post["post_id"],
|
||||
"message": post["post_text"],
|
||||
"url": post["post_url"],
|
||||
"image": post["image"],
|
||||
"images": post["images"],
|
||||
"video": post["video"],
|
||||
"link": post["link"],
|
||||
"time": post["time"].timestamp(),
|
||||
|
@ -27,6 +28,7 @@ def scrape_data(page):
|
|||
def get_all_data():
|
||||
data = {}
|
||||
for page in PAGES:
|
||||
print(" -> " + page)
|
||||
data[page] = scrape_data(page)
|
||||
return data
|
||||
|
||||
|
|
|
@ -4,7 +4,7 @@ bs4==0.0.1
|
|||
certifi==2020.6.20
|
||||
chardet==3.0.4
|
||||
cssselect==1.1.0
|
||||
facebook-scraper==0.2.9
|
||||
facebook-scraper==0.2.34
|
||||
fake-useragent==0.1.11
|
||||
html2text==2020.1.16
|
||||
idna==2.10
|
||||
|
|
Loading…
Reference in a new issue