Last active
December 9, 2021 13:11
-
-
Save rloredo/e8fd44ef5f80c98820eb3b2990e7097e to your computer and use it in GitHub Desktop.
Use Stanza library to tokenize and lemmatize texts
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import stanza | |
import pandas as pd | |
#Load a dataframe with text in one column | |
df = pd.DataFrame({'label':[1], 'text' : ['Hi Juan Carlos'] }) | |
#Initialize the engine. In this case in Portuguese | |
nlp_pt = stanza.Pipeline(lang='pt', processors='tokenize,mwt,pos,lemma') | |
#Tokenize, lemmatize and POS | |
doc_list = [] | |
error_elements = [] | |
print('Processing') | |
for i, e in tqdm(enumerate(df.text.values)): | |
try: | |
doc_list.append((nlp_pt(e), i)) | |
except: | |
print('Error in element {}'.format(i)) | |
error_elements.append((e,i)) | |
#Extract all lemmas | |
def get_lemma(tuple_list): | |
''' | |
Get list of pos elements [pos, text] | |
Use extract_pos to get all pos of one type | |
''' | |
upos = [] | |
for doc, i in tqdm(tuple_list): | |
upos.append (([[word.upos, word.lemma] for sent in doc.sentences for word in sent.words], i)) | |
return upos | |
pos = get_lemma(doc_list) | |
df_temp = pd.DataFrame(pos, columns=['user_pos', 'ind_original']) | |
#Extract an specific part of speech | |
def extract_pos(list_of_pos, list_pos_names, to_set = False): | |
''' | |
Extract all the lemmas from a list of pos | |
Use to_set to get unique values | |
''' | |
if to_set == True: | |
return set([x[1] for x in list_of_pos if x[0] in list_pos_names]) | |
else: | |
return [x[1] for x in list_of_pos if x[0] in list_pos_names] | |
df_temp['user_noun_verbs'] = [extract_pos(x, ['NOUN', 'VERB', 'PROPN']) for x in df_temp['user_pos']] | |
df_temp['user_adjs'] = [extract_pos(x, ['ADJ']) for x in df_temp['user_pos']] | |
df_temp['user_advs'] = [extract_pos(x, ['ADV']) for x in df_temp['user_pos']] |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment