Skip to content

Commit f65f687

Browse files
tianshoulong1988tianshoulong1988
committed
Update p01_interact_with_http_services_as_client.rst
1 parent f003a5f commit f65f687

File tree

1 file changed

+151
-161
lines changed

1 file changed

+151
-161
lines changed

‎source/c11/p01_interact_with_http_services_as_client.rst

Lines changed: 151 additions & 161 deletions
Original file line numberDiff line numberDiff line change
@@ -5,210 +5,200 @@
55
----------
66
问题
77
----------
8-
You need to access various services via HTTP as a client. For example, downloading
9-
data or interacting with a REST-based API.
10-
8+
你需要以客户端的方式通过HTTP协议方位多种服务。例如，下载数据或者与基于REST的API进行交互。
119
|
1210
1311
----------
1412
解决方案
1513
----------
16-
For simple things, it’s usually easy enough to use the urllib.request module. For
17-
example, to send a simple HTTP GET request to a remote service, do something like this:
14+
对于简单的事情来说，通常使用 ``urllib.request`` 模块就够了。例如，发送一个简单的HTTP GET请求到远程的服务上，可以这样做：
1815

19-
from urllib import request, parse
16+
.. code-block:: python
2017
21-
# Base URL being accessed
22-
url = 'http://httpbin.org/get'
18+
from urllib import request, parse
2319
24-
# Dictionary of query parameters (if any)
25-
parms = {
26-
'name1' : 'value1',
27-
'name2' : 'value2'
28-
}
20+
# Base URL being accessed
21+
url = 'http://httpbin.org/get'
2922
30-
# Encode the query string
31-
querystring = parse.urlencode(parms)
23+
# Dictionary of query parameters (if any)
24+
parms = {
25+
'name1' : 'value1',
26+
'name2' : 'value2'
27+
}
3228
33-
# Make a GET request and read the response
34-
u = request.urlopen(url+'?' + querystring)
35-
resp = u.read()
29+
# Encode the query string
30+
querystring = parse.urlencode(parms)
3631
37-
If you need to send the query parameters in the request body using a POST method,
38-
encode them and supply them as an optional argument to urlopen() like this:
32+
# Make a GET request and read the response
33+
u = request.urlopen(url+'?' + querystring)
34+
resp = u.read()
3935
40-
from urllib import request, parse
36+
如果你需要使用POST方法在请求主体中发送查询参数，可以将参数编码后作为可选参数提供给 ``URLopen()`` 函数，就像这样：
4137

42-
# Base URL being accessed
43-
url = 'http://httpbin.org/post'
38+
.. code-block:: python
4439
45-
# Dictionary of query parameters (if any)
46-
parms = {
47-
'name1' : 'value1',
48-
'name2' : 'value2'
49-
}
40+
from urllib import request, parse
5041
51-
# Encode the query string
52-
querystring = parse.urlencode(parms)
42+
# Base URL being accessed
43+
url = 'http://httpbin.org/post'
5344
54-
# Make a POST request and read the response
55-
u = request.urlopen(url, querystring.encode('ascii'))
56-
resp = u.read()
45+
# Dictionary of query parameters (if any)
46+
parms = {
47+
'name1' : 'value1',
48+
'name2' : 'value2'
49+
}
5750
58-
If you need to supply some custom HTTP headers in the outgoing request such as a
59-
change to the user-agent field, make a dictionary containing their value and create a
60-
Request instance and pass it to urlopen() like this:
51+
# Encode the query string
52+
querystring = parse.urlencode(parms)
6153
62-
from urllib import request, parse
63-
...
54+
# Make a POST request and read the response
55+
u = request.urlopen(url, querystring.encode('ascii'))
56+
resp = u.read()
6457
65-
# Extra headers
66-
headers = {
67-
'User-agent' : 'none/ofyourbusiness',
68-
'Spam' : 'Eggs'
69-
}
58+
如果你需要在发出的请求中提供一些自定义的HTTP头，例如修改 ``user-agent`` 字段,可以创建一个包含字段值的字典，并创建一个Request实例然后将其传给 ``urlopen()`` ，如下：
7059

71-
req = request.Request(url, querystring.encode('ascii'), headers=headers)
60+
.. code-block:: python
7261
73-
# Make a request and read the response
74-
u = request.urlopen(req)
75-
resp = u.read()
62+
from urllib import request, parse
63+
...
7664
77-
If your interaction with a service is more complicated than this, you should probably
78-
look at the requests library. For example, here is equivalent requests code for the
79-
preceding operations:
65+
# Extra headers
66+
headers = {
67+
'User-agent' : 'none/ofyourbusiness',
68+
'Spam' : 'Eggs'
69+
}
8070
81-
import requests
71+
req = request.Request(url, querystring.encode('ascii'), headers=headers)
8272
83-
# Base URL being accessed
84-
url = 'http://httpbin.org/post'
73+
# Make a request and read the response
74+
u = request.urlopen(req)
75+
resp = u.read()
8576
86-
# Dictionary of query parameters (if any)
87-
parms = {
88-
'name1' : 'value1',
89-
'name2' : 'value2'
90-
}
77+
如果需要交互的服务比上面的例子都要复杂，也许应该去看看 requests 库（https://pypi.python.org/pypi/requests）。例如，下面这个示例采用requests库重新实现了上面的操作：
9178

92-
# Extra headers
93-
headers = {
94-
'User-agent' : 'none/ofyourbusiness',
95-
'Spam' : 'Eggs'
96-
}
79+
.. code-block:: python
9780
98-
resp = requests.post(url, data=parms, headers=headers)
81+
import requests
9982
100-
# Decoded text returned by the request
101-
text = resp.text
83+
# Base URL being accessed
84+
url = 'http://httpbin.org/post'
10285
103-
A notable feature of requests is how it returns the resulting response content from a
104-
request. As shown, the resp.text attribute gives you the Unicode decoded text of a
105-
request. However, if you access resp.content, you get the raw binary content instead.
106-
On the other hand, if you access resp.json, then you get the response content inter‐
107-
preted as JSON.
108-
Here is an example of using requests to make a HEAD request and extract a few fields
109-
of header data from the response:
86+
# Dictionary of query parameters (if any)
87+
parms = {
88+
'name1' : 'value1',
89+
'name2' : 'value2'
90+
}
11091
111-
import requests
92+
# Extra headers
93+
headers = {
94+
'User-agent' : 'none/ofyourbusiness',
95+
'Spam' : 'Eggs'
96+
}
11297
113-
resp = requests.head('http://www.python.org/index.html')
98+
resp = requests.post(url, data=parms, headers=headers)
11499
115-
status = resp.status_code
116-
last_modified = resp.headers['last-modified']
117-
content_type = resp.headers['content-type']
118-
content_length = resp.headers['content-length']
100+
# Decoded text returned by the request
101+
text = resp.text
119102
120-
Here is a requests example that executes a login into the Python Package index using
121-
basic authentication:
122-
import requests
103+
关于requests库，一个值得一提的特性就是它能以多种凡是从请求中返回响应结果的内容。从上面的代码来看， ``resp.text`` 带给我们的是以Unicode解码的响应文本。但是，如果去访问 ``resp.content`` ，就会得到原始的二进制数据。另一方面，如果访问 ``resp.json``` ，那么就会得到JSON格式的响应内容。
123104

124-
resp = requests.get('http://pypi.python.org/pypi?:action=login',
125-
auth=('user','password'))
105+
下面这个示例利用 ``requests`` 库发起一个HEAD请求，并从响应中提取出一些HTTP头数据的字段：
126106

127-
Here is an example of using requests to pass HTTP cookies from one request to the
128-
next:
107+
.. code-block:: python
129108
130-
import requests
109+
import requests
131110
132-
# First request
133-
resp1 = requests.get(url)
134-
...
111+
resp = requests.head('http://www.python.org/index.html')
135112
136-
# Second requests with cookies received on first requests
137-
resp2 = requests.get(url, cookies=resp1.cookies)
113+
status = resp.status_code
114+
last_modified = resp.headers['last-modified']
115+
content_type = resp.headers['content-type']
116+
content_length = resp.headers['content-length']
138117
139-
Last, but not least, here is an example of using requests to upload content:
118+
Here is a requests example that executes a login into the Python Package index using
119+
basic authentication:
120+
import requests
140121
141-
import requests
142-
url = 'http://httpbin.org/post'
143-
files = { 'file': ('data.csv', open('data.csv', 'rb')) }
122+
resp = requests.get('http://pypi.python.org/pypi?:action=login',
123+
auth=('user','password'))
144124
145-
r = requests.post(url, files=files)
125+
Here is an example of using requests to pass HTTP cookies from one request to the
126+
next:
127+
128+
import requests
129+
130+
# First request
131+
resp1 = requests.get(url)
132+
...
133+
134+
# Second requests with cookies received on first requests
135+
resp2 = requests.get(url, cookies=resp1.cookies)
136+
137+
Last, but not least, here is an example of using requests to upload content:
138+
139+
import requests
140+
url = 'http://httpbin.org/post'
141+
files = { 'file': ('data.csv', open('data.csv', 'rb')) }
142+
143+
r = requests.post(url, files=files)
146144
147-
|
148145
149146
----------
150147
讨论
151148
----------
152-
For really simple HTTP client code, using the built-in urllib module is usually fine.
153-
However, if you have to do anything other than simple GET or POST requests, you really
154-
can’t rely on its functionality. This is where a third-party module, such as requests,
155-
comes in handy.
156-
For example, if you decided to stick entirely with the standard library instead of a library
157-
like requests, you might have to implement your code using the low-level http.cli
158-
ent module instead. For example, this code shows how to execute a HEAD request:
159-
160-
from http.client import HTTPConnection
161-
from urllib import parse
162-
163-
c = HTTPConnection('www.python.org', 80)
164-
c.request('HEAD', '/index.html')
165-
resp = c.getresponse()
166-
167-
print('Status', resp.status)
168-
for name, value in resp.getheaders():
169-
print(name, value)
170-
171-
Similarly, if you have to write code involving proxies, authentication, cookies, and other
172-
details, using urllib is awkward and verbose. For example, here is a sample of code that
173-
authenticates to the Python package index:
174-
175-
import urllib.request
176-
177-
auth = urllib.request.HTTPBasicAuthHandler()
178-
auth.add_password('pypi','http://pypi.python.org','username','password')
179-
opener = urllib.request.build_opener(auth)
180-
181-
r = urllib.request.Request('http://pypi.python.org/pypi?:action=login')
182-
u = opener.open(r)
183-
resp = u.read()
184-
185-
# From here. You can access more pages using opener
186-
...
187-
188-
Frankly, all of this is much easier in requests.
189-
Testing HTTP client code during development can often be frustrating because of all
190-
the tricky details you need to worry about (e.g., cookies, authentication, headers, en‐
191-
codings, etc.). To do this, consider using the httpbin service. This site receives requests
192-
and then echoes information back to you in the form a JSON response. Here is an
193-
interactive example:
194-
195-
>>> import requests
196-
>>> r = requests.get('http://httpbin.org/get?name=Dave&n=37',
197-
... headers = { 'User-agent': 'goaway/1.0' })
198-
>>> resp = r.json
199-
>>> resp['headers']
200-
{'User-Agent': 'goaway/1.0', 'Content-Length': '', 'Content-Type': '',
201-
'Accept-Encoding': 'gzip, deflate, compress', 'Connection':
202-
'keep-alive', 'Host': 'httpbin.org', 'Accept': '*/*'}
203-
>>> resp['args']
204-
{'name': 'Dave', 'n': '37'}
205-
>>>
206-
207-
Working with a site such as httpbin.org is often preferable to experimenting with a real
208-
site—especially if there’s a risk it might shut down your account after three failed login
209-
attempts (i.e., don’t try to learn how to write an HTTP authentication client by logging
210-
into your bank).
211-
Although it’s not discussed here, requests provides support for many more advanced
212-
HTTP-client protocols, such as OAuth. The requests documentation is excellent (and
213-
frankly better than anything that could be provided in this short space). Go there for
214-
more information.
149+
对于真的很简单HTTP客户端代码，用内置的 ``urllib`` 模块通常就足够了。但是，如果你要做的不仅仅只是简单的GET或POST请求，那就真的不能在依赖它的功能了。这时候就是第三方模块比如 ``requests`` 大显身手的时候了。
150+
151+
例如，如果你决定坚持使用标准的程序库而不考虑像 ``requests`` 这样的第三方库，那么也许就不得不使用底层的 ``http.client`` 模块来实现自己的代码。比方说，下面的代码展示了如何执行一个HEAD请求：
152+
153+
.. code-block:: python
154+
155+
from http.client import HTTPConnection
156+
from urllib import parse
157+
158+
c = HTTPConnection('www.python.org', 80)
159+
c.request('HEAD', '/index.html')
160+
resp = c.getresponse()
161+
162+
print('Status', resp.status)
163+
for name, value in resp.getheaders():
164+
print(name, value)
165+
166+
167+
同样地，如果必须编写涉及代理、认证、cookies以及其他一些细节方面的代码，那么使用 ``urllib`` 就显得特别别扭和啰嗦。比方说，下面这个示例实现在Python包索引上的认证：
168+
169+
.. code-block:: python
170+
171+
import urllib.request
172+
173+
auth = urllib.request.HTTPBasicAuthHandler()
174+
auth.add_password('pypi','http://pypi.python.org','username','password')
175+
opener = urllib.request.build_opener(auth)
176+
177+
r = urllib.request.Request('http://pypi.python.org/pypi?:action=login')
178+
u = opener.open(r)
179+
resp = u.read()
180+
181+
# From here. You can access more pages using opener
182+
...
183+
184+
坦白说，所有的这些操作在 ``equests`` 库中都变得简单的多。
185+
186+
在开发过程中测试HTTP客户端代码常常是很令人沮丧的，因为所有棘手的细节问题都需要考虑（例如cookies、认证、HTTP头、编码方式等）。要完成这些任务，考虑使用httpbin服务（http://httpbin.org）。这个站点会接收发出的请求，然后以JSON的形式将相应信息回传回来。下面是一个交互式的例子：
187+
188+
.. code-block:: python
189+
190+
>>> import requests
191+
>>> r = requests.get('http://httpbin.org/get?name=Dave&n=37',
192+
... headers = { 'User-agent': 'goaway/1.0' })
193+
>>> resp = r.json
194+
>>> resp['headers']
195+
{'User-Agent': 'goaway/1.0', 'Content-Length': '', 'Content-Type': '',
196+
'Accept-Encoding': 'gzip, deflate, compress', 'Connection':
197+
'keep-alive', 'Host': 'httpbin.org', 'Accept': '*/*'}
198+
>>> resp['args']
199+
{'name': 'Dave', 'n': '37'}
200+
>>>
201+
202+
在要同一个真正的站点进行交互前，先在 httpbin.org 这样的万展上做实验常常是可取的办法。尤其是当我们面对3次登录失败就会关闭账户这样的风险时尤为有用（不要尝试自己编写HTTP认证客户端来登录你的银行账户）。
203+
204+
尽管本节没有涉及， ``request`` 库还对许多高级的HTTP客户端协议提供了支持，比如OAuth。 ``requests`` 模块的文档（http://docs.python-requests.org)质量很高（坦白说比在这短短的一节的篇幅中所提供的任何信息都好），可以参考文档以获得更多地信息。

0 commit comments

Comments
 (0)