Browse Source

demo run edits

main
westernmeadow 1 year ago
parent
commit
c4706047bb
3 changed files with 602 additions and 1583 deletions
  1. +2
    -2
      Forums/CryptBB/crawler_selenium.py
  2. +3
    -3
      Forums/CryptBB/parser.py
  3. +597
    -1578
      Forums/Initialization/geckodriver.log

+ 2
- 2
Forums/CryptBB/crawler_selenium.py View File

@ -32,7 +32,7 @@ baseURL = 'http://cryptbbtg65gibadeeo2awe3j7s6evg7eklserehqr4w4e2bis5tebid.onion
# Opens Tor Browser, crawls the website
def startCrawling():
opentor()
# forumName = getForumName()
forumName = getForumName()
driver = getAccess()
if driver != 'down':
@ -43,7 +43,7 @@ def startCrawling():
print(driver.current_url, e)
closetor(driver)
# new_parse(forumName, baseURL, False)
new_parse(forumName, baseURL, False)
# Opens Tor Browser


+ 3
- 3
Forums/CryptBB/parser.py View File

@ -23,7 +23,7 @@ def cryptBB_description_parser(soup):
sign = [] # 5 all user's signature in each post (usually a standard message after the content of the post)
post = [] # 6 all messages of each post
feedback = [] # 7 all feedbacks of each vendor (this was found in just one Forum and with a number format)
addDate = [] # 8 all dated of each post
addDate = [] # 8 *all dated of each post
# Finding the topic (should be just one coming from the Listing Page)
@ -170,8 +170,8 @@ def cryptBB_listing_parser(soup):
forum = "CryptBB" # 0 *forum name
board = "-1" # 1 *board name (the previous level of the topic in the Forum categorization tree.
# For instance: Security/Malware/Tools to hack Facebook. The board here should be Malware)
topic = [] # 2 *all topics
author = [] # 3 *all authors of each topic
author = [] # 2 *all authors of each topic
topic = [] # 3 *all topics
views = [] # 4 number of views of each topic
posts = [] # 5 number of posts of each topic
href = [] # 6 this variable should receive all cleaned urls (we will use this to do the marge between


+ 597
- 1578
Forums/Initialization/geckodriver.log
File diff suppressed because it is too large
View File


Loading…
Cancel
Save