Das deutsche Python-Forum

Code: Alles auswählen

def nlp_textrazor():

path = 'Input/'				
	outdir = ''
	dirs = os.listdir(path)
	print dirs

	for file in dirs:
		print "Processing: " + file
		filename = path + file
		outfilename = (outdir + file)
		text = open(filename, 'r').read().split()
		s = ''.join(text)
		
		stop_words = open('stopwords.txt', 'r').read().split()	
		words = (text)

		filtered_text = [w for w in words if not w in stop_words]

		newfile = open(outfilename, "w")
		for line in filtered_text:
			s = str(line) + '\n'
			newfile.write(s)
				 
	return

Code: Alles auswählen

print os.path.abspath(path)

Code: Alles auswählen

def load_words(filename):
    with open(filename, 'r') as text_file:
        return text_file.read().split()


def nlp_textrazor():
    path = 'Input'                        
    filenames = os.listdir(path)
    print filenames
    stop_words = set(load_words('stopwords.txt'))
    for filename in filenames:
        print 'Processing:', filename
        words = load_words(os.path.join(path, filename))
        with open(filename, 'w') as filtered_file:
            for word in (w for w in words if w not in stop_words):
                filtered_file.write(word + '\n')