dark_web_forums/Forums/DWForums/parser.py

__author__ = 'DarkWeb'

# Here, we are importing the auxiliary functions to clean or convert data
from Forums.Utilities.utilities import *
from datetime import date
from datetime import timedelta
import re

# Here, we are importing BeautifulSoup to search through the HTML tree
from bs4 import BeautifulSoup

# This is the method to parse the Description Pages (one page to each topic in the Listing Pages)

def dwForums_description_parser(soup):

    # Fields to be parsed

    topic = "-1"           # topic name
    user = []              # all users of each post
    addDate = []           # all dated of each post
    feedback = []          # all feedbacks of each vendor (this was found in just one Forum and with a number format)
    status = []            # all user's authority in each post such as (adm, member, dangerous)
    reputation = []        # all users's karma in each post (usually found as a number)
    sign = []              # all user's signature in each post (usually a standard message after the content of the post)
    post = []              # all messages of each post
    interest = []          # all user's interest in each post

    # Finding the topic (should be just one coming from the Listing Page)

    li = soup.find("h1", {"class": "p-title-value"})

    topic = li.text
    topic = topic.replace(u'\xa0', ' ')
    topic = topic.replace(",","")
    topic = topic.replace("\n","")
    topic = cleanString(topic.strip())
    # print(topic)
    # Finding the repeated tag that corresponds to the listing of posts

    # posts = soup.find("form", {"name": "quickModForm"}).findAll('div', {"class": "windowbg"}) + \
    #         soup.find("form", {"name": "quickModForm"}).findAll('div', {"class": "windowbg2"})

    try:
        posts = soup.find('div', {"class": "js-replyNewMessageContainer"}).find_all(
            'article', {"class": "js-post"}, recursive=False)
        # print(len(posts))

        # For each message (post), get all the fields we are interested to:

        for ipost in posts:

            # Finding a first level of the HTML page

            # post_wrapper = ipost.find('div', {"class": "post_wrapper"}).find('div', {"class": "poster"})
            post_wrapper = ipost.find('h4', {"class": "message-name"})
            # Finding the author (user) of the post

            # author = post_wrapper.find('h4')
            author = post_wrapper.text.strip()
            # print("author " + author)
            user.append(cleanString(author))  # Remember to clean the problematic characters

            # Finding the status of the author



            # Testing here two possibilities to find this status and combine them
            # if ipost.find('h5', {"class": "deleted_post_author"}):
            #     status.append(-1)
            #     interest.append(-1)
            #     reputation.append(-1)
            #     addDate.append(-1)
            #     post.append("THIS POST HAS BEEN REMOVED!")
            #     sign.append(-1)
            #     feedback.append(-1)
            #     continue

            # CryptBB does have membergroup and postgroup

            membergroup = ipost.find('h5', {"class": "userTitle"})
            # DWForums doesnt have postgroups
            postgroup = None
            if membergroup != None:
                membergroup = membergroup.text.strip()
                if postgroup != None:
                    postgroup = postgroup.text.strip()
                    membergroup = membergroup + " - " + postgroup
            else:
                if postgroup != None:
                    membergroup = postgroup.text.strip()
                else:
                    membergroup = "-1"

            status.append(cleanString(membergroup))
            # print("status " + cleanString(membergroup))
            # Finding the interest of the author
            # DWForums does not have blurb
            blurb = ipost.find('li', {"class": "blurb"})
            if blurb != None:
                blurb = blurb.text.strip()
            else:
                blurb = "-1"
            interest.append(cleanString(blurb))

            # Finding the reputation of the user
            # CryptBB does have reputation
            author_stats = ipost.find('div', {"class": "message-userExtras"})
            if author_stats != None:
                karma = author_stats.find_all('dl', {"class": "pairs"})[2]
            else:
                karma = None
            if karma != None:
                karma = karma.text
                karma = karma.replace("Reaction score","")
                karma = karma.replace(":", "")
                karma = karma.strip()
            else:
                karma = "-1"
            reputation.append(cleanString(karma))
            # print("karma " + cleanString(karma))
            # Getting here another good tag to find the post date, post content and users' signature

            postarea = ipost.find('div', {"class": "message-attribution-main"})

            dt = postarea.find('time', {"class": "u-dt"})['datetime']
            # dt = dt.strip().split()
            dt = dt.strip()[:16]
            dt = dt.replace("T",", ")
            day=date.today()
            if "Yesterday" in dt:
                yesterday = day - timedelta(days=1)
                yesterday = yesterday.strftime('%m-%d-%Y')
                stime = dt.replace('Yesterday,','').strip()
                date_time_obj = yesterday+ ', '+stime
                date_time_obj = datetime.strptime(date_time_obj,'%m-%d-%Y, %H:%M')
            elif "hours ago" in dt:
                day = day.strftime('%m-%d-%Y')
                date_time_obj = postarea.find('span', {"class": "post_date"}).find('span')['title']
                date_time_obj = datetime.strptime(date_time_obj, '%m-%d-%Y, %H:%M')
            else:
                date_time_obj = datetime.strptime(dt, '%Y-%m-%d, %H:%M')
                stime = date_time_obj.strftime('%b %d, %Y')
                sdate = date_time_obj.strftime('%I:%M %p')


            addDate.append(date_time_obj)
            # print("date " + str(date_time_obj))
            # Finding the date of the post
            # date_time_obj = datetime.strptime(dt, '%a %b %d, %Y %I:%M %p')
            # smalltext = postarea.find('div', {"class": "flow_hidden"}).find('div', {"class": "keyinfo"})\
            #     .find('div', {"class": "smalltext"})
            # sdatetime = smalltext.text
            # sdatetime = sdatetime.replace(u"\xab","") # Removing unnecessary characters
            # sdatetime = sdatetime.replace(u"\xbb","") # Removing unnecessary characters
            # sdatetime = sdatetime.split("on: ")       # Removing unnecessary characters
            # sdatetime = sdatetime[1].strip()
            # stime = sdatetime[:-12:-1]                # Finding the time of the post
            # stime = stime[::-1]
            # sdate = sdatetime.replace(stime,"")       # Finding the date of the post
            # sdate = sdate.replace(",","")
            # sdate = sdate.strip()

            # Covert the date of the post that can be informed as: "12 February 2016", "today", "yesterday". We need
            # a date format here as "mm/dd/yyyy"

            # addDate.append(convertDate(sdate,"english", crawlerDate) + " " + stime)

            # Finding the post

            inner = ipost.find('article', {"class": "message-body"})
            inner = inner.text.strip()
            # print(inner)
            post.append(cleanString(inner))

            # Finding the users's signature

            # signature = ipost.find('div', {"class": "post_wrapper"}).find('div', {"class": "moderatorbar"}).find('div', {"class": "signature"})
            signature = ipost.find('aside', {"class": "message-signature"})
            if signature != None:
                signature = signature.text.strip()
                # print(signature)
            else:
                signature = "-1"
            sign.append(cleanString(signature))

            # As no information about users's feedback was found, just assign "-1" to the variable

            feedback.append("-1")
    except:
        print("error when parsing posts")
        # if soup.find('td', {"class": "trow1"}).text == " You do not have permission to access this page. ":
        #     user.append("-1")
        #     status.append(-1)
        #     interest.append(-1)
        #     reputation.append(-1)
        #     addDate.append(-1)
        #     post.append("NO ACCESS TO THIS PAGE!")
        #     sign.append(-1)
        #     feedback.append(-1)


    # Populate the final variable (this should be a list with all fields scraped)

    row = (topic, post, user, addDate, feedback, status, reputation, sign, interest)

    # Sending the results

    return row

# This is the method to parse the Listing Pages (one page with many posts)

def dwForums_listing_parser(soup):

    board = "-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)

    nm = 0             # this variable should receive the number of topics
    topic = []         # all topics
    user = []          # all users of each topic
    post = []          # number of posts of each topic
    view = []          # number of views of each topic
    addDate = []       # when the topic was created (difficult to find)
    href = []          # this variable should receive all cleaned urls (we will use this to do the marge between
                       # Listing and Description pages)

    # Finding the board (should be just one)

    board = soup.find('h1', {"class": "p-title-value"}).text
    board = cleanString(board.strip())

    # Finding the repeated tag that corresponds to the listing of topics

    regex = re.compile('.*structItem--thread.*')
    itopics = soup.find_all("div", {"class": regex})
    index = 0
    for itopic in itopics:

        # For each topic found, the structure to get the rest of the information can be of two types. Testing all of them
        # to don't miss any topic


        # tds = itopic.findAll('td', {"class": "subject stickybg2"})
        #
        # if len(tds) > 0:
        #    tag.append("strong")
        #    tag.append("subject stickybg2")
        #    tag.append("stats stickybg")
        # else:
        #    tds = itopic.findAll('td', {"class": "subject windowbg2"})
        #    if len(tds) > 0:
        #       tag.append("span")
        #       tag.append("subject windowbg2")
        #       tag.append("stats windowbg")

        # Adding the topic to the topic list
        topics = itopic.find("div", {"class": "structItem-title"}).text
        topics = topics.replace(",", "")
        topics = topics.replace("\n", "")
        topic.append(cleanString(topics.strip()))

        # Counting how many topics we have found so far

        nm = len(topic)

        # Adding the url to the list of urls
        link = itopic.select_one('a[href^="/threads/"]')
        link = link['href']
        link = cleanLink(link)
        href.append(link)

        # Finding the author of the topic
        ps = itopic.find('a', {"class": "username"}).text
        author = ps.strip()
        user.append(cleanString(author))

        # Finding the number of replies
        meta = itopic.find("div", {"class": "structItem-cell--meta"})
        meta = meta.find_all("dl")
        posts = meta[0].find("dd").text
        post.append(cleanString(posts))

        # Finding the number of Views
        tview = meta[1].find("dd").text
        view.append(cleanString(tview))



        # If no information about when the topic was added, just assign "-1" to the variable
        minor = itopic.find("div", {"class": "structItem-minor"})
        dt = minor.find('time')['datetime']
        dt = dt.strip()[:16]
        dt = dt.replace("T", ", ")
        day = date.today()
        if "Yesterday" in dt:
            yesterday = day - timedelta(days=1)
            yesterday = yesterday.strftime('%m-%d-%Y')
            stime = dt.replace('Yesterday,', '').strip()
            date_time_obj = yesterday + ', ' + stime
            date_time_obj = datetime.strptime(date_time_obj, '%m-%d-%Y, %H:%M')
        else:
            date_time_obj = datetime.strptime(dt, '%Y-%m-%d, %H:%M')
            stime = date_time_obj.strftime('%b %d, %Y')
            sdate = date_time_obj.strftime('%I:%M %p')
        addDate.append(date_time_obj)




        index += 1
    return organizeTopics("DWForums", nm, topic, board, view, post, user, addDate, href)



def dwForums_links_parser(soup):

    # Returning all links that should be visited by the Crawler

    href = []
    #print(soup.find('table', {"class": "tborder clear"}).find(
     #   'tbody').find_all('tr', {"class": "inline_row"}))
    regex = re.compile('.*structItem--thread.*')
    listing = soup.find_all("div", {"class": regex})

    for a in listing:
        link = a.select_one('a[href^="/threads/"]')
        link = link['href']

        href.append(link)

    return href