Charles Karpati | Wordclouds

This actually turned out to be rather straight forward

import requests
 from bs4 import BeautifulSoup as bs
 from wordcloud import WordCloud, STOPWORDS 
 from PIL import Image #use to open the image
 import matplotlib.pyplot as plt
 import numpy as np
 import os 
 %matplotlib inline

A library exists to do what we want

from wordcloud import ImageColorGenerator

So we just get some text

words =""
 bs = bs(requests.get('http://www.myths.com/pub/lyrics/Bangles_.html').text, 'html.parser')
 for x in bs.find_all("pre"): words = words+x.text
 words = words.replace("\n", " ").lower() 

words

Get an image and font face

fontpath = os.path.join('EastSeaDokdo-Regular.ttf') 

https://fonts.google.com/specimen/East+Sea+Dokdo#standard-styles

Lets createa visualization function to do everything we could possibly want

def show(img, s1, s2):
     plt.figure(figsize=(s1,s2))
     plt.clf() # Clear Figure
     plt.imshow(img, interpolation='bilinear')
     plt.axis('off')
     plt.show()
     #fig.set_figwidth(14) # set width
     #fig.set_figheight(12) # set height 

create an array mask of the image

img = Image.open('510VP5aoo+L._AC_SL1402_.jpg')
 mask = np.array(img)

and.... GO!

cloud = WordCloud(stopwords=STOPWORDS, background_color='white', font_path=fontpath, max_words=300, width=300, height=200 ).generate(words)
 show(cloud, 4, 4)

show(img, 4, 4)

cloud = WordCloud(mask= mask, stopwords=STOPWORDS, background_color='white', max_words=300, width=300, height=200 ).generate(words)
 show(cloud, 4, 4)

# max_font_size=1000
 # font_path=fontpath,
 color_mask = ImageColorGenerator(mask)
 coloredMaskCloud = cloud.recolor(color_func=color_mask)
 show(coloredMaskCloud, 4, 4)

Animating the SVG would be cool. But tricky. Animate the words within the svg?

Maybe next time.

For now, here are some miscellaneous notes on how to read files and do stuff with em. Like removing stop words.

exClean = """
 ds_2_words= ds_2.split()
 print('Length before removing the stopwords: ', len(ds_2_words))
 for word in ds_2_words:
     if word in ENGLISH_STOP_WORDS:
         ds_2_words.remove(word)
     else:
         pass
     
 print('Length after removing the stopwords: ', len(ds_2_words))
 """

exeList = """
 values = df['Client'].value_counts().keys().tolist()
 counts = df['Client'].value_counts().tolist()
 """

exeRead = """
 with open('word.txt', 'r') as f: #Read the data or words
     word_text= f.read()
 words= word_text.split(',') #preprocessing before using in WordCloud module
 cloud= WordCloud().generate_from_text(' '.join(words))
 """