dark_web_forums/Forums/Dread/parser.py

__author__ = 'DarkWeb'

# Here, we are importing the auxiliary functions to clean or convert data
import datetime

from Forums.Utilities.utilities import *
from datetime import date
from datetime import timedelta
import re
import traceback
# 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 dread_description_parser(soup):

    # Fields to be parsed

    topic = "-1"              # 0 *topic name
    user = []                 # 1 *all users of each post
    status = []               # 2 all user's authority in each post such as (adm, member, dangerous)
    reputation = []           # 3 all user's karma in each post (usually found as a number)
    interest = []             # 4 all user's interest in each post
    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 dates of each post

    # Finding the topic (should be just one coming from the Listing Page)
    container = soup.find('div', {"class": "content"})
    li = container.find("a", {"class": "title"})
    if li == None:
        return None
    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"})

    # putting the initial post data since it is separated from comments
    # author name
    init_post = container.find('div', {"class": "item"})
    author = init_post.find('div', {"class": "author"}).select_one('a[href^="/u/"]').text
    flair = init_post.find('div', {"class": "author"}).find("span", {"class": "flair"})
    try:
        flair = flair.text.strip()
        author = author.replace(flair, '')
    except:
        pass
    author = author.strip()
    user.append(cleanString(author))
    # status
    flair = init_post.find("span", {"class": "flair"})
    if flair != None:
        flair = flair.text.strip()
    else:
        flair = "-1"
    status.append(cleanString(flair))
    # no blurb
    interest.append(-1)
    # points for post
    karma = init_post.find("div", {"class": "voteCount"})
    if karma != None:
        karma = karma.text
        karma = karma.replace("points", "")
        karma = karma.replace(":", "")
        karma = karma.strip()
    else:
        karma = "-1"
    reputation.append(cleanString(karma))
    # date
    spans = init_post.find('div', {"class": "author"}).find('span', recursive=False)
    dt = spans['title']
    month = find_month(dt)
    split_text = dt.split()
    day = int(re.search(r'\d+', split_text[0]).group())
    year = int(split_text[2])
    hm = re.findall(r'\d+', split_text[-1])
    hm[0] = int(hm[0])
    hm[1] = int(hm[1])
    date_time_obj = datetime(year, month, day, hour=hm[0], minute=hm[1])
    addDate.append(date_time_obj)

    # content
    inner = init_post.find("div", {"class": "postContent"})
    inner = inner.text.strip()
    post.append(cleanString(inner))

    # no signature
    sign.append(-1)
    # no feedback
    feedback.append(-1)


    comments = soup.find('div', {"class": "postComments"})
    if comments == None:
        row = (topic, post, user, addDate, feedback, status, reputation, sign, interest)
        return row
    else:
        comments = soup.find('div', {"class": "postComments"}).find_all('div', "comment")
    # print(len(posts))

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

    for ipost in comments:

        # Finding a first level of the HTML page

        # post_wrapper = ipost.find('div', {"class": "post_wrapper"}).find('div', {"class": "poster"})
        cc = ipost.find('div', {"class": "commentContent"})

        post_wrapper = cc.find('a', {"class": "username"}).text
        flair = cc.find("span", {"class": "flair"})
        try:
            flair = flair.text.strip()
            post_wrapper = post_wrapper.replace(flair, '')
        except:
            pass
        author = post_wrapper.strip()
        user.append(cleanString(author))


        # Finding the status of the author

        # Dread does not have membergroup and postgroup, but it has flair, similar enough


        postgroup = None
        if flair != None:
            if postgroup != None:
                postgroup = postgroup.text.strip()
                flair = flair + " - " + postgroup
        else:
            if postgroup != None:
                flair = postgroup.text.strip()
            else:
                flair = "-1"

        status.append(cleanString(flair))
        # print("status " + cleanString(membergroup))
        # Finding the interest of the author
        # Dread does not have blurb

        interest.append(-1)

        # Finding the reputation of the user
        # Dread doesn't have reputation per user, but instead each post has its own point system
        karma = cc.find('div', {"class": "votes"})

        if karma != None:
            karma = karma.text
            karma = karma.replace("points","")
            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": "timestamp"}).find('span', recursive=False)
        dt = postarea['title']
        month = find_month(dt)
        split_text = dt.split()
        day = int(re.search(r'\d+', split_text[0]).group())
        year = int(split_text[2])
        hm = re.findall(r'\d+', split_text[-1])
        hm[0] = int(hm[0])
        hm[1] = int(hm[1])
        date_time_obj = datetime(year, month, day, hour=hm[0], minute=hm[1])
        addDate.append(date_time_obj)

        # Finding the post

        inner = ipost.find('div', {"class": "commentBody"})
        inner = inner.text.strip()
        # print(inner)
        post.append(cleanString(inner))

        # No signature for Dread

        sign.append(-1)

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

        feedback.append("-1")

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

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

    # Sending the results

    return row

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

def dread_listing_parser(soup):

    nm = 0              # *this variable should receive the number of topics
    forum = "Dread"  # 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)
    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
                        # Listing and Description pages)
    addDate = []        # 7 when the topic was created (difficult to find)

    # Finding the board (should be just one)

    board = soup.find('a', {"class": "banner-top"}).text
    board = cleanString(board.strip())

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

    itopics = soup.find("div", {"class": "postBoard"}).find_all("div", {"class": "item"}, recursive=False)

    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

        # Adding the topic to the topic list
        topic_title = itopic.find("a", {"class": "title"})
        title_flair = topic_title.find('span', {"class": "flair"})
        topics = topic_title.text
        try:
            title_flair = title_flair.text.strip()
            topics = topics.replace(title_flair, '')
        except:
            pass
        topics = topics.replace(u'\xa0', ' ')
        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 = topic_title['href']
        link = cleanLink(link)
        href.append(link)

        # Finding the author of the topic
        ps = itopic.find('div', {"class": "author"})
        post_wrapper = ps.select_one('a[href^="/u/"]').text
        flair = ps.find("span", {"class": "flair"})
        try:
            flair = flair.text.strip()
            post_wrapper = post_wrapper.replace(flair, '')
        except:
            pass
        user = post_wrapper.strip()
        author.append(cleanString(user))

        # Finding the number of replies
        meta = itopic.find("div", {"class": "postMain"})
        post = meta.find("a").text
        post = post.replace("comments", '').strip()
        posts.append(cleanString(post))

        # Finding the number of Views - not shown in Dread
        views.append("-1")

        # If no information about when the topic was added, just assign "-1" to the variable
        spans = itopic.find('div', {"class": "author"}).find('span', recursive=False)
        dt = spans['title']
        month = find_month(dt)
        split_text = dt.split()
        day = int(re.search(r'\d+', split_text[0]).group())
        year = int(split_text[2])
        hm = re.findall(r'\d+', split_text[-1])
        hm[0] = int(hm[0])
        hm[1] = int(hm[1])
        date_time_obj = datetime(year, month, day, hour=hm[0], minute=hm[1])
        addDate.append(date_time_obj)

    return organizeTopics(forum, nm, board, author, topic, views, posts, href, addDate)


def dread_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"}))

    listing = soup.find("div", {"class": "postBoard"}).find_all("div",{"class": "item"}, recursive=False)

    for a in listing:
        link = a.find("a", {"class": "title"})
        link = link['href']

        href.append(link)

    return href

def find_month(s):
    if 'January' in s:
        return 1
    elif 'February' in s:
        return 2
    elif 'March' in s:
        return 3
    elif 'April' in s:
        return 4
    elif 'May' in s:
        return 5
    elif 'June' in s:
        return 6
    elif 'July' in s:
        return 7
    elif 'August' in s:
        return 8
    elif 'September' in s:
        return 9
    elif 'October' in s:
        return 10
    elif 'November' in s:
        return 11
    elif 'December' in s:
        return 12