lda.py

file_names = os.listdir(PATH)
texts = []
for name in file_names:
    if name.endswith(".txt"):
        with codecs.open(PATH + "/" + name, encoding = 'utf-8') as f:
            print(name)
            text = f.read()
            lst = re.findall(r'\w+', text)
            words = []
            for word in lst:
                lemma = morph.parse(word)[0]  # делаем разбор
                words.append(lemma.normal_form)

            texts.append(words)
            # each document should contain lemmatized words separated by spaces