【华为云-上云之路】 连接华为云资源nlp

连接华为云资源nlp

本文是主要是记录博主连接nlp资源踩下的坑，以及分享一些注意事项.

使用的环境时ubuntu 16.04 +vscode +python 2.7

连接方法是用python构造函数，采用api相应的进行访问

准备篇

可以先看下官方的推荐文档，华为云nlp。推荐先使用插件操作之后，在进行相应的api的操作。官方推荐插件

获取项目的token

华为云提供两种方式访问nlp资源，token和ak/sk。对后一种连接方式不了解？？要使用相应的签名算法，对token也是看了文档才明白是什么, 为了占个小便宜，我容易吗？

用python构造消息头与消息体并拿到token

先使用了requests库封装一个函数，在调用该函数用来实现访问api资源的功能。

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

import requests


class HttpRequests():
    '''对requests库封装的GET、POST、PUT等方法调用类'''

    def __init__(self, url, data=None, type='GET', cookie=None, headers=None, proxies=None):
        self.url = url
        self.data = data
        self.type = type
        self.cookie = cookie
        self.headers = headers
        self.proxies = proxies
        self.send_request()

    def send_request(self):
        '''setup a request'''

        if self.url == None or self.url == '':
            raise ('The url should not empty!')
        if self.type == 'POST':
            self.req = requests.post(self.url, data=self.data, headers=self.headers, proxies=self.proxies)
        elif self.type == 'GET':
            if self.data == None:
                self.req = requests.get(self.url, headers=self.headers, proxies=self.proxies)
            else:
                self.req = requests.get(self.url, params=self.data, headers=self.headers, proxies=self.proxies)
        elif self.type == 'PUT':
            self.req = requests.put(self.url, data=self.data, headers=self.headers, proxies=self.proxies)
        else:
            self.req = None
            raise ('The http request type NOT support now!')

    def get_code(self):
        try:
            return self.req.status_code
        except:
            raise ('get code fail')

    def get_url(self):
        try:
            return self.req.url
        except:
            raise ('get url fail')

    def get_text(self):
        try:
            return self.req.text
        except:
            raise ('get text fail')

    def get_headers(self):
        try:
            return self.req.headers
        except:
            raise ('get headers fail')

    def get_cookie(self):
        headers = self.get_headers()
        if 'set-cookie' in headers:
            return headers['set-cookie']
        else:
            return None

将拿取的保存在本地的token，用来构造接下来的消息头，代码如下。记得要更改相应的参数@两个用户名@密码@保存参数

#!/usr/bin/env python
# -*- coding: UTF-8 -*-


import json
import requests
from HttpRequest import HttpRequests
take_headers = {"content-type": "application/json"}
take_url = "https://iam.cn-north-4.myhuaweicloud.com/v3/auth/tokens"
data = {
        "auth": {
                "identity": {
                    "methods": [
                        "password"
                    ],
                    "password": {
                        "user": {
                            "name": "xxxxx",
                            "password": "xxxxxx",
                            "domain": {
                                "name": "xxxxx"
                            }
                        }
                    }
                },
                "scope": {
                    "project": {
                        "name": "cn-north-4"
                    }
                }
            }
        }

req = HttpRequests(take_url, data=json.dumps(data), type="POST", headers =take_headers)
#print(req.get_code())
token_info = {
                "token": req.get_headers()["X-Subject-Token"],
                "project_id": "cn-north-4",
                "expires_at": "expires_timestamp"
            }
f = open('/home/danoao/文档/test.txt','w')
f.write(json.dumps(token_info))
f.close()
print(req.get_code())
print(req.get_headers())

如果以上代码运行的没有问题的话就可以看到，打印出来的状态码与消息头了。

将取出的token够造函数头取访问相应的api

写了一个类，用来调取nlu的那一部分资源。这里不得不吐嘈，！！！相应的url看的人脑袋晕乎乎的，{project_id}与{project_name}居然不是一个东西…代码如下
可以按需更该以下参数@url@call_data@地址

#!/usr/bin/env python
# -*- coding: UTF-8 -*-


import json
import requests
from HttpRequest import HttpRequests
import sys
reload(sys)
import time
sys.setdefaultencoding( "utf-8" )
headers = {"content-type": "application/json;charset=utf8"}
url = "https://nlp-ext.cn-north-4.myhuaweicloud.com/v1/0629bb1a768026c42fffc00e25759f39/nlu/"

class Clientnlu():
    def __init__(self):
        self.headers = {"content-type": "application/json;charset=utf8"}
        self.get_token()
    def get_token(self):
        '''
        获取保存在本地的token，配套另一python代码使用
        具体可以访问网址
        https://support.huaweicloud.com/api-iam/iam_30_0001.html
        '''
        try:
            f = open("/home/danoao/文档/test.txt", 'r')
            token_str = f.read().encode('utf-8')
            f.close()
            token_info = json.loads(token_str)
            headers["x-auth-token"] = token_info["token"]
            print('读取token成功')
        except:
            print('读取失败')
    def classification(self,content):
         '''
           nlu的分类资源，具体可参见
           https://support.huaweicloud.com/api-nlp/nlp_03_0017.html
        '''
        call_data={
            "content":content,
            "domain":1
            }
        class_url= url + "classification"
        call_req = HttpRequests(class_url, data=json.dumps(call_data),
        type="POST", headers =headers)
        if(call_req.get_code() == 200):
            a = call_req.get_text()
            b= json.loads(a)
            print("连接资源成功")
            content = b["result"]["content"]
            print(content)
            label = b["result"]["label"]
            #print(label)
            confidence=b["result"]["confidence"]
            #print(confidence)
        else:
            print(call_req.get_text())

if __name__ == "__main__":
    x = Clientnlu()
    while(1):
        i = 1000
        xx = str(i)
        y  =x.classification(xx)
        time.sleep(1)
        i = i-1
        time.sleep(1)
    #print(y)

总结

别说，只想占个便宜，哈哈哈最后的代码把额的意图露出来了，只想静静的水一水一个坐垫？？？？小的代码迭代了6次。最终版本也刷了3次？？？

哈哈哈，期待华为云啥时候能开放一下更牛逼一些的nlu资源啊，期待中。