Pdf2wordcloud.py

# Pdf2wordcloud  
  
# 1) pdf to text object  
import os  
import PyPDF2  
from PyPDF2 import PdfFileWriter  
  
party="grüne" #  and others: Linke, Grüne, SPD, FDP, CDU/CSU, AfD   
  
pa= "C:/Users/Dome/Desktop/nu/Wahlprogramme 2017/"  
os.chdir(pa)  
pdoc= pa+party+".pdf"  
  
pdfFileObj = open(pdoc,'rb')  
pdfReader = PyPDF2.PdfFileReader(pdfFileObj)  
num_pages = pdfReader.numPages  
  
writer = PdfFileWriter()  
count = 0  
text = ""  
li=""  
  
# loop for each page  
while count < num_pages:  
    pageObj = pdfReader.getPage(count)  
    text = pageObj.extractText()  
    count+=1  
    li += " "+text  
      
###################################  
##### 2) TEXT CLEANING ############  
  
from stop_words import get_stop_words  
import re  
sw = get_stop_words('de') # German stopwords. Not as useful as EN version  
  
re.sub(r'[0-9]+', '', li) # remove numbers  
  
# replace most common noise  
liste=li.replace('\n', ' ').replace('œ', ' ').replace('Œ', ' ').replace(  
        "-","").replace("KAPITEL","").replace("Kapitel","").split(" ")  
  
stz=[word for word in liste if word.isalpha()]# remove marks  
  
# extend for most common meaningless words and word fragments  
sw.extend(("macht","schen","schon","Kapitel","setzt","weitere","lehnt","viel",  
           "stellen","gehen","chen","geht","gilt","lehnen", "viele","gibt",  
           "darf","halten","dürfen","neben","gehört","vielen","jedoch",  
           "braucht", "nehmen","rung","seit","sollten","deren","etwa","beim",  
           "außerdem","stehen", "sitzen", "mehr","sollen","müssen","sowie",  
           "deshalb","daher","dafür","dabei","brauchen","zudem","setzen",  
           "besser","neue","neu","neuen","immer","gute"))  
  
# list comprehension, only longer than 4 characters and not in stopwords  
tes=[word for word in stz if word.lower() not in sw if len(word) >= 4]  
  
###################################  
##### 2) WORD CLOUD ###############  
  
from wordcloud import WordCloud, STOPWORDS, ImageColorGenerator  
from PIL import Image  
import numpy as np  
import matplotlib.pyplot as plt  
  
words = ' '.join(tes) # list to text  
mask = np.array(Image.open(  
        "C:/Users/Dome/Desktop/nu/Wahlprogramme 2017/germany-flag-medium.png"))  
image_colors = ImageColorGenerator(mask)  
# function takes in your text and your mask and generates a wordcloud   
def generate_wordcloud(words, mask):  
    word_cloud = WordCloud(width = 512, height = 512,max_words=100,  
                           background_color='white', stopwords=STOPWORDS,  
                           mask=mask).generate(words)  
    plt.figure(figsize=(10,8),facecolor = 'white', edgecolor='blue')  
    plt.imshow(word_cloud.recolor(color_func=image_colors), #colors  
               interpolation="bilinear")  
    plt.axis('off')  
    plt.tight_layout(pad=0)  
    plt.savefig(pa+"Wordclouds/"+party+".png", format="png") # save  
    plt.show()  
      
# run function to create wordcloud  
generate_wordcloud(words, mask)