lio_and_central_flores / processing.py
holylovenia's picture
Upload processing.py with huggingface_hub
ff92401 verified
raw
history blame contribute delete
No virus
2.29 kB
import re
import re
def parse_text(data):
"""
Parses a list of text data into English and Lio language lists.
Args:
data (list): A list of strings containing text data.
Returns:
tuple: A tuple containing two lists: English language data and Lio language data.
"""
# Clean the data by removing unwanted characters and formatting
cleaned_data = [re.sub(r'^\ufeff?\d+\t|\b\d+\)\s*', '', re.sub(r'\t+', '\t', text)).replace('\t\n', '').replace('\ufeff', '') for text in data]
# Split the data into English and Lio lists
eng_data, lio_data = [], []
sublist = []
for item in cleaned_data:
# Replace tabs with spaces
item = item.replace('\t', ' ')
if item != '':
sublist.append(item)
else:
if sublist:
# Append sublist to result if not empty
eng_data.append(sublist[-1])
lio_data.append(sublist[0])
sublist = []
# Append the remaining sublist if any
if sublist:
eng_data.append(sublist[-1])
lio_data.append(sublist[0])
return eng_data, lio_data
def parse_wordlist(data):
"""
Parses a word list for translation into English, Indonesian, and Nage languages.
Args:
data (list): A list of strings containing word list data.
Returns:
tuple: A tuple containing three lists: English words, Indonesian words, and Nage words.
"""
# Clean the data by removing unwanted characters and formatting
cleaned_data = [re.sub(r'^\ufeff?\d+\t|\b\d+\)\s*', '', re.sub(r'\t+', '\t', text)).replace('\t\n', '').replace('\ufeff', '') for text in data]
# Remove empty elements
cleaned_data = [elem for elem in cleaned_data if elem != '']
eng_word = []
nage_word = []
ind_word = []
for item in cleaned_data:
split_data = item.split()
# make sure that we only get [eng, ind, nage] because the actual wordlist have
# added 'comment' column and not all the words have their comments
if len(split_data) == 3:
eng_word.append(split_data[0])
ind_word.append(split_data[1])
nage_word.append(split_data[2])
return eng_word, ind_word, nage_word