Difference between revisions of "Moroccan Parliament Web Crawler"

From edegan.com
Jump to navigation Jump to search
Line 4: Line 4:
 
|Has start date=9/27/2016
 
|Has start date=9/27/2016
 
|Image=morocco_flag.jpg
 
|Image=morocco_flag.jpg
|Has project status=Active
 
 
|Has notes=
 
|Has notes=
|Has project status=
 
 
|Is dependent on=
 
|Is dependent on=
 
|Depends upon it=
 
|Depends upon it=

Revision as of 19:03, 2 March 2017


McNair Project
Moroccan Parliament Web Crawler
Project logo 02.png
Project Information
Project Title Moroccan Parliament Web Driver
Owner Peter Jalbert
Start Date 9/27/2016
Deadline
Primary Billing
Notes
Has project status
Copyright © 2016 edegan.com. All Rights Reserved.


Overview

This web driver is designed to save information from the Moroccan Legislature website as a pdf before the information is removed from the website due to lack of space.

Navigating the Site

On the right hand side of the website, the final bullet is a link to archived bills.

On the right hand side of the website, one bullet above the last header, is a link to links of all the proposed bills. When clicked, the user is directed to the most recent batch of links of proposed bills. The movement button on the bottom left in arabic means, "go to last page", and the second from the left means "previous page". When "go to last page" is clicked, the batch of links to proposed bills that are about to be removed from the website are listed with the oldest link at the bottom of the list.

When any proposed bill link is clicked, it opens up a page with information about the bill, and about discussions in parliament. There are up to two hyperlinks on this page, the first of which is a pdf of the original bill. If there is a second link further below, it contains a pdf of the Committee Report for the bill (if it was sent to a Committee).

The two pdfs, as well as the webpage of the proposed bill may be subject to record keeping.


Main Page

Driving Opportunities

Five websites within the Moroccan Legislature site have data that needs to be recorded.

Moroccan Monarchy Proposed Bills

Monarchy Proposed Bills

The data that needs to be extracted from this site includes the pdfs of all the bill pages, as well as any interior pdfs on each page. The bill pages should be named by their url, and the interior pdfs should be named by their respective bill numbers.

Moroccan House of Representatives Proposed Bills

House Proposed Bills

See Monarchy proposed bills for instructions.

Moroccan Legislature Ratified Bills

Ratified Bills

See Monarchy proposed bills for instructions.

Moroccan Legislature Oral Questions

Oral Questions

The embedded pdfs on this site are not useful. The useful data elements are the dates and questions listed on the main site. To get this data, I built a web crawler that scraped the date, question, and all relevant information about the question from the site. Then, I used Selenium to download pdfs of the questions.

Moroccan Legislature Written Questions

Written Questions

See Oral Questions section.


Web Driver Code

   #General Bill Download
   from selenium import webdriver
   from selenium.webdriver.common.action_chains import ActionChains
   from selenium.webdriver.common.keys import Keys
   import time
   import urllib
   import string
   import re
   ##########################
   #launch Google Chrome Browser
   driver = webdriver.Chrome()
   ##########################
   def switch_window():
       handles = driver.window_handles
       driver.switch_to_window(handles[-1])
   ##########################
   #Visit desired website
   driver.get('http://www.chambredesrepresentants.ma/ar/%D8%A7%D9%84%D8%AA%D8%B4%D8%B1%D9%8A%D8%B9/%D9%84%D8%A7%D8%A6%D8%AD%D8%A9-            %D9%85%D9%82%D8%AA%D8%B1%D8%AD%D8%A7%D8%AA-%D8%A7%D9%84%D9%82%D9%88%D8%A7%D9%86%D9%8A%D9%86?body_value=&field_og_commission_target_id=All')
   ##########################
   bills_list = driver.find_elements_by_xpath("//li/h3/a")
   for i in range(len(bills_list)):
       ActionChains(driver).key_down(Keys.SHIFT).perform()
       bills_list[i].click()
       ActionChains(driver).key_up(Keys.SHIFT).perform()
       switch_window()
       url = driver.current_url
       unicode_url =  urllib.unquote(str(url)).decode('utf8') 
       url_parts = string.split(unicode_url, "/")
       i = len(url_parts)
      ##########################
       #Build arabic tag backwards, accounting for backwards spelling
       tag = ""
       while i > 4:
           tag += url_parts[i - 1]
           i -= 1
      ##########################
       #Navigate to pdf of website
       change_button  = driver.find_elements_by_xpath("//a [@class='pdf' and @rel='nofollow']")[0]
       ActionChains(driver).key_down(Keys.SHIFT).perform()
       change_button.click()
       ActionChains(driver).key_up(Keys.SHIFT).perform()
       switch_window()
      #########################
       #Gets current window's URL
       url = driver.current_url
      ########################
       #Saves file at URL to current directory 
       urllib.urlretrieve(url, tag)
       driver.close()
       switch_window()
     ######################### 
       pdfs_on_page = driver.find_elements_by_xpath("//div/div/div/article/div/ul/li/a")
       #finds interior pdfs on the page
       if pdfs_on_page:
           for j in range(len(pdfs_on_page)):
               element = pdfs_on_page[j]
              #######################
               #click on pdf
               ActionChains(driver).key_down(Keys.SHIFT).perform()
               element.click()
               ActionChains(driver).key_up(Keys.SHIFT).perform()
               switch_window()
               url = driver.current_url
               pdf_tag = string.split(str(url), "/")[-1]
              ######################
               #leaves link if it is not a pdf
               if re.findall(".pdf", pdf_tag):
                   #saves interior pdf
                   urllib.urlretrieve(url, pdf_tag)                
               driver.close()
               switch_window()
       ########################
       driver.close()
       switch_window()
  ######################### 
   print "download complete"  
   #########################
   #close browser
   driver.quit()


Web Crawler Code

I used a python library called Scrapy.

Once downloaded and installed properly, a new project is created by typing into the command line:

scrapy startproject projectname

This will create a project for you with all of the necessary components you need.

From this window, scrapy provides a helpful tool to test any web scraping lines of code you would like to try out. Enter:

scrapy shell 'webaddress'

From here, you can type selector statements and print them to see if your statements are getting the data you desire.

To actually build your webcrawler, open a new python script, and save it in the spiders folder that was created automatically for you under your projectname folder. Some example code for a spider is shown below; this was my spider for the oral questions portion of the Moroccan site.

import scrapy
import string



class MySpider(scrapy.Spider):
    name = "oral"
    page_range = 375
    start_urls = (["http://www.chambredesrepresentants.ma/ar/%D9%85%D8%B1%D8%A7%D9%82%D8%A8%D8%A9-%D8%A7%D9%84%D8%B9%D9%85%D9%84-  
%D8%A7%D9%84%D8%AD%D9%83%D9%88%D9%85%D9%8A/%D8%A7%D9%84%D8%A3%D8%B3%D9%80%D8%A6%D9%84%D8%A9%D8%A7%D9%84%D8%B4
%D9%81%D9%88%D9%8A%D8%A9? field_ministeres_tid=All&field_groupe_concerne_target_id=All&field_parlementaires_associes_target_id=All&body_value
=&field_transfere_ou_non_value=All"] +
    ["http://www.chambredesrepresentants.ma/ar/%D9%85%D8%B1%D8%A7%D9%82%D8%A8%D8%A9-%D8%A7%D9%84%D8%B9%D9%85%D9%84
%D8%A7%D9%84%D8%AD%D9%83%D9%88%D9%85%D9%8A/%D8%A7%D9%84%D8%A3%D8%B3%D9%80%D8%A6%D9%84%D8%A9-     
%D8%A7%D9%84%D8%B4%D9%81%D9%88%D9%8A%D8%A9field_ministeres_tid=All&field_groupe_concerne_target_id=All&field_parlementaires
_associes_target_id=All&body_value=&field_transfere_ou_non_value=All&page=" + str(num) for num in range(page_range)])

    def parse(self, response):
        a = response.css('ul.listing_questions')
        for header in a.css('li'):
            date = header.css('h3.sorting_date::text').extract_first()
            if date != None:
                placeholder = date
            if date == None:
                date = placeholder
            question = header.css('a::text').extract_first()
            info = header.css('div.questionss_group::text').extract()
            info = .join(elem for elem in info)
            info_split = string.split(info, "\n")
            info1 = info_split[2]
            info1 = ' '.join(info1.split())
            info2 = info_split[4]
            info2 = ' '.join(info2.split())
            info3 = info_split[5]
            info3 = ' '.join(info3.split())
            
            yield {
                'date': date,
                'info1':info1,
                'info2': info2,
                'info3': info3,
                'question': question,
                'url': response.url
            }



Further Inquiries

Further inquiries have been requested for the Kuwait, Tunisian, and Algerian Parliaments. In addition, a way to link the downloaded pdfs to pertinent lines in the csv files would be very useful.

Kuwait Parliament

Kuwait Site

Data to download (Tables and PDF files): Monarchy - Proposed Bills (مشروع بقانون); For 13th and 14th terms. Parliament - Proposed Bills (اقتراح بقانون); for 14th term: Proposals (اقتراح بــرغــبـــة): For 13th term and 14th terms Meeting Agendas/Minutes (جدول اعمال الجلسه); for 14th term

UPDATE: Kuwait Bills Site

The site generates dynamic content. The desired pages for scraping are found by selecting the first option of the first dropdown menu, and hitting enter. """The website's URL never changes.""" The HTML of the site will change as objects are selected on the site. I used Selenium to interact with the elements on the page, and the .text method of Web Elements in Selenium to have Selenium act as a psuedo web crawler. The code is found in the following directory:

E:\McNair\Projects\Middle East Studies Web Drivers\Kuwait

The next step in the project requires the data found when selecting the links on each page. This data has specifics on each of the bills. The page appears as a dynamically generated frame which presents some issues. There is a close button on the window that is opened up that cannot be accessed from the interior frame, but also cannot be accessed from the original framework. I experimented with executing javascript code, and switching frames, but nothing worked. My workaround solution was to reload the page every time I opened up the interior frame. This is a separate script saved in the same directory listed above and titled kuwait_retrieve.py.


Tunisian Parliament

Tunisian Site

Data to Download(save web pages as PDFs and documents as PDFs):

Bills proposed to parliament(مشروع قانون معروض على المجلس): click here Bills proposed to legislative committees (مشروع قانون معروض على اللجان): click here General Assembly Deliberations (الجلسات العامة): click here


Algerian Parliament

Algerian Site

Data to download: Bills that have been voted on (القوانين المصوت عليها الفترة التشريعية السابعة : 2012 - 2017): click here Parliamentary initiatives/questions (المبادرات البرلمانية): click here