def bad_words_filter(img_name):
base_name = img_name.split('\\')[-1]
text = ' '
with open('data/results/' + 'res_{}.txt'.format(base_name.split('.')[0]),
'r') as f:
# text = f.read().split('\n')
text = ' '
# for i in text:
# temp += i
lines = f.readlines()
for line in lines:
str_arr = line.strip().split(',')
if len(str_arr) > 4:
for word in str_arr[4:]:
text += (word + ',')
with open('data/results/' + 'res_{}.txt'.format(base_name.split('.')[0]),
'a') as f:
client = AipImageCensor(appId='18097522',
apiKey='bgeRXmpblbOMocErTxvMCiOF',
secretKey='SyRmwoFGsKXqC3LvAAYwlDlf1BckBOwG')
if client.antiSpam(text)['result']['spam'] == 0:
f.write('无敏感信息')
else:
f.write('待检测的文本里面含有侮辱、色情、暴力和政治敏感词汇。\n')
for i in client.antiSpam(text)['result']['reject']:
if (len(i['hit']) != 0):
f.write(str(i['hit']))
def test_article_filter():
try:
res = request.get_json()
article_id = res.get('article_id')
article_images = res.get('article_images')
Logging.logger.info('request_args:{0}'.format(res))
if not article_id:
return jsonify(errno=-1, errmsg='参数错误,请传入要查询的文章的article_id')
article = Article.query.get(article_id)
if not article:
return jsonify(errno=-1, errmsg='参数错误,该文章不存在')
docs = mongo_store.articles.find({'title': article.title})
doc = docs[0]
article_dict = dict()
content = doc.get('content')
title = article.title.encode("utf-8")
article_dict['title'] = title
article_dict['content'] = content
obj = SensitiveFilter()
str11 = ''.join(
[item.get('text') for item in content if item.get('text')])
text = {'content': str11}
txt_data = obj.content_check(text)
if txt_data.get('errcode') == 40001:
redis_store.delete('access_token')
txt_data = obj.content_check(text)
Logging.logger.info('res_data:{0}'.format(txt_data))
APP_ID = '15791531'
API_KEY = 'kajyVlP73XtSGBgoXDIHH5Za'
SECRET_KEY = 'u2TClEW6LaHIIpRNdFcL2HIexcgG1ovC'
client = AipImageCensor(APP_ID, API_KEY, SECRET_KEY)
txt_resp = client.antiSpam(str11)
Logging.logger.info('txt_resp:{0}'.format(txt_resp))
for img in article_images:
img_resp = client.imageCensorUserDefined(img)
print(img_resp)
Logging.logger.info('img_resp:{0}'.format(img_resp))
# img_data = obj.img_check(img)
# print img_data
return jsonify(errno=0, errmsg="OK", data=txt_data)
except Exception as e:
Logging.logger.error('errmsg:{0}'.format(e))
return jsonify(errno=-1, errmsg='文章详情查询失败')
class TextReview():
def __init__(self):
self.client = AipImageCensor(APP_ID, API_KEY, SECRET_KEY)
def recognize(self, text):
ans = self.client.antiSpam(text)
if ('error_msg' in ans.keys()):
print(ans['error_msg'])
return 0
ans = ans['result']
if (ans['reject'] == [] and ans['review'] == []): #pass
return 0
return 1
'''
class BaiduAntiSpam(object):
def __init__(self, api):
"""
:param api: ['15052846', 'SiU9AAGaZn2Zja7d8iSVqce5', 'P6NZ07ROvKTXFnSlDMmH4hf1smOxbfAA']
"""
self.api = api
self.client = AipImageCensor(*self.api)
@retry(tries=3, delay=2)
def anti_spam(self, text):
_ = self.client.antiSpam(text)
_['log_id'] = text
return _
def get_result(self, file, corpus):
with open(file, 'a') as f:
for s in tqdm_notebook(corpus):
res = self.anti_spam(s)
f.writelines("%s\n" % res)
from aip import AipImageCensor
""" 你的 APPID AK SK """
APP_ID = '11156578'
API_KEY = '3K73kH6H4aGoZbUrE1N0oTO5'
SECRET_KEY = 'YoL5g6BCnWG4mQvEo0TjyDPozlySdDRp'
client = AipImageCensor(APP_ID, API_KEY, SECRET_KEY)
result = client.antiSpam('民警提醒说，此类骗局中，通常骗子要求汇款的理由包括“发红包”、“买礼物”、“生病就医”、“凑路费”、“生意需要资金”、“见面需要彩礼”等各种理由。因此，微信交友遇到这样情形要小心防骗')
print(result)
def get_check_json(cls, content):
client = AipImageCensor(cls.APP_ID, cls.API_KEY, cls.SECRET_KEY)
return client.antiSpam(content=content)
class ContentVerify:
def __init__(self):
self.client = AipImageCensor(
current_app.config['APP_ID'],
current_app.config['API_KEY'],
current_app.config['SECRET_KEY']
)
def verify_uploaded_images(self, img):
if isinstance(img, str):
_, suffix = os.path.splitext(img)
if not suffix == '.gif':
result = self.client.imageCensorUserDefined(img)
else:
result = self.client.antiPornGif(img)
else:
filename = img.filename
_, suffix = os.path.splitext(filename)
if not suffix == '.gif':
result = self.client.imageCensorUserDefined(img.read())
else:
result = self.client.antiPornGif(img.read())
msg = self.extract_msg(result, suffix)
ok_or_not = self.is_ok(msg, suffix)
return msg, ok_or_not
def verify_uploaded_avatar(self, avatar):
filename = avatar.filename
_, suffix = os.path.splitext(filename)
result = self.client.faceAudit(avatar.read())
msg = result['result'][0]['data']['antiporn']
if "conclusion" in msg:
if not msg['conclusion'] == "色情":
return '', True
else:
return msg['conclusion'] + "，上传失败！", False
else:
return '审核失败，请重试！', False
def verify_text(self, text):
result = self.client.antiSpam(text)
spam_code = result['result']['spam']
if spam_code == 0:
return '', spam_code, True
elif spam_code == 1:
return "涉及敏感词，审核未通过，提交失败！", spam_code, False
else:
return "待审核，您将在12小时内收到审核结果", spam_code, False
def extract_msg(self, result, suffix):
if "error_msg" in result:
return [{"error_msg": result['error_msg']}]
if not suffix == '.gif':
if 'data' not in result.keys():
return [""]
data = result['data']
msg = [d["msg"] for d in data]
return msg
else:
data = result['conclusion']
msg = [data]
return msg
def is_ok(self, msg, suffix):
if isinstance(msg[0], dict):
return False
if not suffix == ".gif":
return ("存在色情内容" not in msg) and ("存在政治敏感内容" not in msg)
else:
return "色情" not in msg
