def load(self, device: str):
"""
Load user-selected task-specific model
Args:
device (str): device information
Returns:
object: User-selected task-specific model
"""
from word2word import Word2word
model = Word2word(self.config.lang, self._tgt)
return PororoWord2Word(model, self.config)
def main(folder_path, lang):
os.chdir(folder_path)
files = os.listdir(folder_path)
for index, file in enumerate(files):
if file.endswith(".csv") and lang in file:
print(file)
df = pd.read_csv(file, dtype=str, sep='\t')
i = 0
with open(lang, 'a') as f:
l = convert_iso_code(lang)
with open(lang + '.en', 'a') as e:
with open(lang + '.' + l, 'a') as t:
for i, row in df.iterrows():
if isinstance(row['source_text'],
str) and isinstance(
row['target_text'], str
) and row['item_type'] != 'RESPONSE':
source = remove_punctuation_and_lower_case(
row['source_text'])
target = remove_punctuation_and_lower_case(
row['target_text'])
f.write(str(i) + '|' + source + '|' + target)
i += 1
f.write("\n")
e.write(str(i) + '|' + source)
e.write("\n")
t.write(str(i) + '|' + target)
t.write("\n")
f.close()
e.close()
t.close()
print(folder_path + "/" + lang + "_dict")
mcsq_dict = Word2word.make("en",
l,
folder_path + "/" + lang,
savedir=folder_path)
#!/usr/bin/env python
# - *- coding: utf- 8 - *-
from word2word import Word2word
dict_path = '/Volumes/Data/dataset/word2word'
with open('word2word/supporting_languages.txt') as f:
lines = f.readlines()
for dict_pair in lines:
codes = dict_pair.strip('\n').split('-')
print("getting dictionary %s-%s" % (codes[0], codes[1]))
w2w = Word2word(codes[0], codes[1], dict_path=dict_path)
def main():
parser = argparse.ArgumentParser()
parser.add_argument('--lang1',
type=str,
required=True,
help="ISO 639-1 code of language. "
"See `http://opus.nlpl.eu/OpenSubtitles2018.php`")
parser.add_argument('--lang2',
type=str,
required=True,
help="ISO 639-1 code of language. "
"See `http://opus.nlpl.eu/OpenSubtitles2018.php`")
parser.add_argument('--datapref',
type=str,
default=None,
help="data prefix to a custom parallel corpus. "
"builds a bilingual lexicon using OpenSubtitles2018 "
"unless this option is provided.")
parser.add_argument('--n_lines',
type=int,
default=100000000,
help="number of parallel sentences used")
parser.add_argument(
'--cutoff',
type=int,
default=5000,
help=
"number of words that are used in calculating collocates within each language"
)
parser.add_argument(
'--rerank_width',
default=100,
type=int,
help="maximum number of target-side collocates considered for reranking"
)
parser.add_argument(
'--rerank_impl',
default="multiprocessing",
type=str,
help=
"choice of reranking implementation: simple, multiprocessing (default)"
)
parser.add_argument('--cased',
dest="cased",
action="store_true",
help="Keep the case.")
parser.add_argument('--n_translations',
type=int,
default=10,
help="number of final word2word translations kept")
parser.add_argument('--save_cooccurrence',
dest="save_cooccurrence",
action="store_true",
help="Save the cooccurrence results")
parser.add_argument('--save_pmi',
dest="save_pmi",
action="store_true",
help="Save the pmi results")
parser.add_argument('--savedir',
type=str,
default=None,
help="location to store bilingual lexicons."
"make sure to use this input when loading from "
"a custom-bulit lexicon.")
parser.add_argument('--num_workers',
default=16,
type=int,
help="number of workers used for multiprocessing")
args = parser.parse_args()
Word2word.make(**vars(args))
from word2word import Word2word
en2fr = Word2word("en","fr") #French
en2es = Word2word("en","es") #Spanish
en2it = Word2word("en","it") #Italian
en2de = Word2word("en","de") #German
en2ta = Word2word("en","ta")
en2te = Word2word("en","te")
def translate(english_word,caseTranslate):
print("Yaha Translate.py ka ilaka start hota hai")
list=english_word.split(" ")
print(list)
try:
if(caseTranslate==1):
for word in list:
trans=en2fr(word)[0]
english_word=english_word.replace(word,trans)
return english_word
elif(caseTranslate==2):
for word in list:
trans=en2es(word)[0]
english_word=english_word.replace(word,trans)
return english_word
elif(caseTranslate==3):
for word in list:
trans=en2it(word)[0]
english_word=english_word.replace(word,trans)
#!/usr/bin/env python
# - *- coding: utf- 8 - *-
from word2word import Word2word
dict_path = '/Volumes/Data/dataset/word2word'
en2fr = Word2word("en", "fr", dict_path=dict_path)
# out: ['pomme', 'pommes', 'pommier', 'tartes', 'fleurs']
print '%s' % ','.join(map(str, en2fr("apple")))
fr2en = Word2word("fr", "en", dict_path=dict_path)
print '%s' % ','.join(map(str, fr2en("pomme", n_best=2)))
# out: ['travaillé', 'travaillait']
print '%s' % ','.join(map(str, en2fr("worked", n_best=2)))
en2zh = Word2word("en", "zh_cn", dict_path=dict_path)
# out: ['老师', '教师', '学生', '导师', '墨盒']
print '%s' % ','.join(map(str, en2zh("teacher")))
zh2en = Word2word("zh_cn", "en", dict_path=dict_path)
print '%s' % ','.join(map(str, zh2en("老师")))
hi2en = Word2word("hi", "en", dict_path=dict_path)
print '%s' % ','.join(map(str, hi2en("मिलने")))