评论接入百度云 文本审核 API

官方地址

https://cloud.baidu.com/doc/ANTIPORN/s/pjwvypelq

接口文档下载

下载文档（更新时间：2019-11-14）

KEY 申请入口

https://console.bce.baidu.com/ai/?fromai=1#/ai/antiporn/overview/index (申请一个 APP)

接口描述

运用业界领先的深度学习技术，判断一段文本内容是否符合网络发文规范，实现自动化、智能化的文本审核。审核内容分为6大类型：色情文本、政治敏感、恶意推广、网络暴恐、低俗辱骂、低质灌水。 定制功能可支持在百度云控制台自由选择审核类型（一个或多个），支持自定义配置文本审核黑名单，且可通过调整阈值控制审核的松紧度标准，大幅度满足文本审核的个性化需求。

接口限制(免费版)

接口使用

1. 先安装 baidu-aip

pip install baidu-aip

2. 我的代码

# baidu_check.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

"""
@author: yinzhuoqun
@site: http://zhuoqun.info/
@email: yin@zhuoqun.info
@time: 2019/11/20 1:21 AM
"""

from aip import AipImageCensor
from utils.json_to_object import json_to_object

BAIDU_APP_ID = '填自己申请的'
BAIDU_API_KEY = '填自己申请的'
BAIDU_SECRET_KEY = '填自己申请的'

labels_text = {
    1: "暴恐违禁",
    2: "色情内容",
    3: "政治敏感",
    4: "恶意推广",
    5: "低俗辱骂",
    6: "低质灌水"
}

spam_text = {
    0: "不违禁",
    1: "违禁",
    2: "人工复审"
}


class CheckText:
    def __init__(self, text):
        self.text = text
        self.app_id = BAIDU_APP_ID
        self.api_key = BAIDU_API_KEY
        self.secret_key = BAIDU_SECRET_KEY
        self.client = AipImageCensor(self.app_id, self.api_key, self.secret_key)

    def request_api(self):
        # print(self.text)
        return self.client.antiSpam(self.text)

    def response(self):
        res = self.request_api()
        if "error_msg" in res.keys():
            return {"status": 203, "msg": res["error_msg"]}
        else:
            check_status = res["result"]["spam"]
            if check_status == 0:
                return {"status": 201, "msg": "审核通过"}
            elif check_status == 1:
                labels = res["result"]["reject"]
                labels_id = [(review["label"]) for review in labels]
                reason = ",".join([labels_text[label_id] for label_id in labels_id]).strip(",")
                # print(reason)
                return {"status": 403, "msg": "审核不通过", "reason": reason}
            elif check_status == 2:
                labels = res["result"]["review"]
                labels_id = [(review["label"]) for review in labels]
                reason = ",".join([labels_text[label_id] for label_id in labels_id]).strip(",")
                # print(reason)
                return {"status": 403, "msg": "人工复审", "reason": reason}
            else:
                return {"status": 200, "msg": "人工复审", "reason": "百度异常"}

    def response_format(self):
        return json_to_object(self.response())


if __name__ == "__main__":
    text = "今晚弹药枪支吗傻x"
    # text = "傻x"
    c = CheckText(text)
    print(c.response_format())

# json_to_object.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

"""
@author: yinzhuoqun
@site: http://zhuoqun.info/
@email: yin@zhuoqun.info
@time: 2019/11/20 12:02 AM
"""

from argparse import Namespace
import json


def json_to_object(data):
    """
    json 或 字典 转成 object
    :param data:
    :return:
    """
    if isinstance(data, dict):
        return json.loads(json.dumps(data), object_hook=lambda d: Namespace(**d))
    else:
        return json.loads(data, object_hook=lambda d: Namespace(**d))

Django 中去掉 HTML 标签

>>> from django.utils.html import strip_tags
>>> 
>>> html = "<a>xieboke.net</a>"
>>> strip_tags(html)
'xieboke.net'
>>>