Script to generate a coma seporated file(.csv) from a given input file with normal english words and unicode characters...

english-words.py

from nltk.corpus import wordnet #Using Natural language toolkit (http://www.nltk.org/)
import codecs
def check(str):
    if not wordnet.synsets(str):
        return 1
    else:
        return 0

a=0
if __name__ == '__main__':
    out = codecs.open('csvout.txt','w',"utf8")
    with codecs.open("Dictionary.txt", "r" , "utf8") as f: #Input file 
        for line in f:
            words = line.split(" ")
            flag=0
            for word in words:
                test = check(word)
                if test == 0:
                    if flag == 0:
                        flag = 1
                        out.write('"')
                    out.write(word)
                    a+=1
                    print("Word found:"+str(a))
                if test == 1:
                    if flag == 1:
                        flag = 2
                        out.write('","')
                    out.write(word.replace("\n",""))
            out.write('"\n')
    print("Done....")

Usage instructions

• Download and install nltk using pip

pip install nltk

• Download the corpus: http://www.nltk.org/data.html#interactive-installer
• Execute the script. Don't forget to change the filename of input & output files

python english-words.py