def test_bulk_translate(self):
translator = Translator()
assert translator.bulk_translate(['apple', 'banana'] *
10) == ['사과', '바나나'] * 10
# Test again with same Instance object
assert translator.bulk_translate(['apple', 'banana'] *
10) == ['사과', '바나나'] * 10
"host": os.getenv("POSTGRES_HOST"),
"database": os.getenv("POSTGRES_DATABASE"),
"user": os.getenv("POSTGRES_USER"),
"password": os.getenv("POSTGRES_PASSWORD")
}
async def connect_to_db():
"""Create a pool to the postgres database using asyncpg"""
return await asyncpg.create_pool(**postgres_options, command_timeout=60)
# papago translation keys - not needed - https://developers.naver.com/docs/papago/
papago_client_id = os.getenv("PAPAGO_CLIENT_ID")
papago_client_secret = os.getenv("PAPAGO_CLIENT_SECRET")
translator = Translator()
# LastFM
last_fm_api_key = os.getenv("LAST_API_KEY")
last_fm_shared_secret = os.getenv("LAST_SHARED_SECRET")
last_fm_root_url = os.getenv("LAST_ROOT_URL")
last_fm_headers = {'user-agent': os.getenv("LAST_USER_AGENT")}
# Patreon
patreon_link = os.getenv("PATREON_LINK")
patreon_role_id = make_int(os.getenv("PATREON_ROLE_ID"))
patreon_super_role_id = make_int(os.getenv("PATREON_SUPER_ROLE_ID"))
# startup time
startup_time = datetime.now()
'''
파파고를 이용한 파일 내용 자동번역 코드
made by 안산드레이아스
설치파일
pip install pypapago
'''
from pypapago import Translator
# 불러올 파일명 입력(동일 디렉터리 위치)
fileName = './test.txt'
# 객체 할당, 번역할 파일 열기
translator = Translator()
with open(fileName, encoding='utf-8', errors='ignore') as f:
# 줄을 읽어 저장, byte를 string으로 디코딩
line = f.readline()
print(line) # 파일 내 텍스트 출력
print(type(line)) # 텍스트 타입 출력
# 번역할 문자를 입력
forTranslateString = line
# 번역하기 (\n(엔터) 이나오면 오류 발생 문자열 처리 필요) english -> korean 옵션
result = translator.translate(forTranslateString, source='en', target='ko', verbose=False)
# 결과 출력
print(result) # 번역된 텍스트 출력
with open("afterTranslate.txt", 'w') as f: # 번역 완료된 텍스트 저장
f.write(result)
def __init__(self):
# 어떤 임베딩 모델을 불러올지에 따라서 다름
self.translator = Translator()
self.bft = BertFineTuning()
class Text_preprocess:
def __init__(self):
# 어떤 임베딩 모델을 불러올지에 따라서 다름
self.translator = Translator()
self.bft = BertFineTuning()
def word_embbeding(self, text):
return self.bft.word_embedding(text)
def papago(self, text, source = "ko", target = "en"):
try:
result = self.translator.translate(text, source = source, target = target)
except:
result = ""
return result
def SVO_extractor(self, data):
with StanfordOpenIE() as client:
svo_pos = {'s_pos' : [], 'v_pos' : [], 'o_pos': [], 'label' : [], 'date' : []}
for index, row in tqdm(data.iterrows()):
try:
for sentence in client.annotate(row['header']):
svo_pos['s_pos'].append(sentence['subject'])
svo_pos['v_pos'].append(sentence['relation'])
svo_pos['o_pos'].append(sentence['object'])
svo_pos['label'].append(1)
svo_pos['date'].append(row['date'])
except AttributeError:
pass
return svo_pos
#괄호 지우기 + 필요한 한자 한글로 변환하기
def svo_embedding(self, svo_pos):
for i in tqdm(range(len(svo_pos['s_pos']))):
svo_pos['s_pos'][i] = self.word_embbeding(svo_pos['s_pos'][i])
svo_pos['v_pos'][i] = self.word_embbeding(svo_pos['v_pos'][i])
svo_pos['o_pos'][i] = self.word_embbeding(svo_pos['o_pos'][i])
return svo_pos
def clean_text(self, text):
patterns = [r"\([^<|>]*\)",
r"\[[^<|>]*\]",
r"\<[^<|>]*\>",
r"[^\w\s]"]
for p in patterns:
try:
text = re.sub(p, '',text)
except TypeError:
text = ''
for hanza in replace_dict.keys():
if hanza in text:
text = re.sub(hanza, replace_dict[hanza], text)
return text.strip()
#단어 지우기
def mkstopwords(self, stopword):
stopwords = ''
for text in stopword:
stopwords += '|' + text
return stopwords[1:]
#총 길이가 j가 안되는 부분 전부 지우기
#남은 한자 있는 열 날리기 ( 무슨 의미인지 예측 불가이므로 그냥 날림 )
def mk_del_smallwords(self, data, j):
data_index=[]
for i in range(len(data)):
if len(re.split(' ', str(data.iloc[i]['header']))) <= j:
data_index.append(data.iloc[i].name)
for i in range(len(data)):
if re.search(r"[\u4e00-\u9fff]", str(data.iloc[i]['header'])):
data_index.append(data.iloc[i].name)
data = data.drop(data_index)
return data
def stop_words(self, data, stopword, i):
stopwords = self.mkstopwords(stopword)
data = data.loc[~data['header'].str.contains(stopwords, na=False)]
data =self.mk_del_smallwords(data, i)
return data
def test_verbose_request(self):
translator = Translator()
assert set(
translator.translate('사과', verbose=True, source='ko',
target='en').keys()) == self.VERBOSE_KEYS
def test_default_request(self):
translator = Translator()
assert translator.translate('Apple') == '사과'
def test_create_instance(self):
sample_re_pattern = re.compile('w+')
translator = Translator(headers={'test': 1234},
regex_pattern=sample_re_pattern)
assert translator.headers == {'test': 1234}
assert translator.regex_pattern == sample_re_pattern
from pypapago import Translator
translator = Translator()
result = translator.translate('I am GROOT')
print(result) # 나는 그루트다
from pypapago import Translator
translator = Translator()
result = translator.translate(
'카카오는 파파고를 좋아해',
source='ko',
target='en',
)
print(result) # Kakao likes papago.
#Code Desc
#ko Korean
#en English
#ja Japanese
#zh-CN Chinese
#zh-TW Chinese traditional
#es Spanish
#fr French
#vi Vietnamese
#th Thai
#id Indonesia
def event_handler(event_type, slack_event):
if event_type == "app_mention":
channel = slack_event["event"]["channel"]
userMessage = slack_event["event"]["blocks"][0]['elements'][0][
'elements'][1]['text']
attachments_dict = dict()
attachments_dict['color'] = '#2398cf'
if '공지' in userMessage:
if '장학' in userMessage:
req = requests.get(scholarshipNoticeUrl)
noticeType = '장학공지'
elif '학사' in userMessage:
req = requests.get(bachelorNoticeUrl)
noticeType = '학사공지'
else:
req = requests.get(generalNoticeUrl)
noticeType = '일반공지'
html = req.text
soup = BeautifulSoup(html, 'html.parser')
Notice = NoticeCrawler(soup, noticeType)
attachments_dict = Notice.crawling().getAnswer()
elif '학식' in userMessage:
req = requests.get(cafeteriaUrl)
html = req.text
soup = BeautifulSoup(req.content.decode('euc-kr', 'replace'),
'html.parser')
Cafeteria = CafeteriaCrawler(soup)
attachments_dict['text'] = Cafeteria.crawling().getAnswer()
elif '번역' in userMessage:
pypapago = Translator()
attachments_dict['pretext'] = '*[번역] 파파고는 말한다. *:penguin:'
attachments_dict['text'] = '```' + pypapago.translate(
userMessage[3:]) + '```'
attachments_dict['mrkdwn_in'] = ["text", "pretext"]
elif '버스' in userMessage or '대성' in userMessage:
req = requests.get(businformationUrl)
html = req.text
soup = BeautifulSoup(html, 'html.parser')
BusInfo = BusInformation(soup)
attachments_dict = BusInfo.crawling().getAnswer()
elif '안녕' in userMessage:
attachments_dict[
'text'] = '나는 *지마블루*:small_blue_diamond: 진리를 찾아 이곳까지 왔죠. \n시간이 얼마 남지 않았습니다. ~*이 활동이 저의 마지막이 될 것 입니다.*~'
else:
attachments_dict['text'] = '무슨 말인지 모르겠네요..'
""" attachments_dict = dict()
attachments_dict['pretext'] = "attachments 블록 전에 나타나는 text"
attachments_dict['title'] = "다른 텍스트 보다 크고 볼드되어서 보이는 title"
attachments_dict['title_link'] = "https://corikachu.github.io"
attachments_dict['fallback'] = "클라이언트에서 노티피케이션에 보이는 텍스트 입니다. attachment 블록에는 나타나지 않습니다"
attachments_dict['text'] = "본문 텍스트! 5줄이 넘어가면 *show more*로 보이게 됩니다.\n1\n2\n3\n4\n5\n6\n7\n8\n9\n1\n2\n3\n4\n5\n6\n7\n8\n9"
attachments_dict['mrkdwn_in'] = ["text", "pretext"] # 마크다운을 적용시킬 인자들을 선택합니다.
attachments = [attachments_dict] """
#slack.chat.post_message(channel="#channel", text=None, attachments=attachments, as_user=True)
# attchment 상태에 따랄 보내는 코드로.
slack.chat.post_message(channel,
attachments=[attachments_dict],
as_user=True)
# Image attachments
""" [
{
"fallback": "Required plain-text summary of the attachment.",
"text": "Optional text that appears within the attachment",
"image_url": "https://mblogthumb-phinf.pstatic.net/MjAxNjEwMjJfNjAg/MDAxNDc3MTM5MDkzMTY5.nTQZS9VKPU3Y1P0J-nOcN4JMz75qU00n09XpQcGJZkAg.3fNACGwA3s_2TSRQxnY6sQDokClABM5fUumyIXAYdQUg.PNG.bugman1303/JW_T-Rex.png?type=w2"
}
] """
return make_response(
"앱 멘션 메시지가 보내졌습니다.",
200,
)
message = "[%s] 이벤트 핸들러를 찾을 수 없습니다." % event_type
return make_response(message, 200, {"X-Slack-No-Retry": 1})
decoder=decoder,
optimizer=optimizer)
ckpt_manager = tf.train.CheckpointManager(ckpt, CKPT_DIR, max_to_keep=5)
start_epoch = 0
#assert ckpt_manager.latest_checkpoint != True
start_epoch = int(ckpt_manager.latest_checkpoint.split('-')[-1])
# checkpoint_path에서 가장 최근의 checkpoint를 restore합니다.
ckpt.restore(ckpt_manager.latest_checkpoint)
image_path = INPUT_DIR + '/test.png'
result, attention_plot = evaluate(image_path, max_length,
attention_features_shape, encoder, decoder,
image_features_extract_model, tokenizer)
predicted_caption = ' '.join(result)
predicted_caption = predicted_caption[:-5]
translator = Translator()
translated_caption = translator.translate(predicted_caption)
print('Prediction Caption:', predicted_caption)
print('Korean Caption : ', translated_caption)
# 한국어 캡션 저장
f = open(OUTPUT_DIR + "/test.txt", mode='wt', encoding='utf-8')
f.write(translated_caption)
f.close()
# Attention Plot
plot_attention(image_path, result, attention_plot)