Question

我有一个small_file.txt文件，其中包含：

1asdfaksdjfhlaksjdhflkjashdflkjhasldkjfhlaksdfhasdf:
2asdfaksdjfhlaksjdhflkjashdflkjhasldkjfhlaksdfhasdf:
3asdfaksdjfhlaksjdhflkjashdflkjhasldkjfhlaksdfhasdf
4asdfaksdjfhlaksjdhflkjashdflkjhasldkjfhlaksdfhasdf:

注意结尾处的冒号，它们只是常规字符串。

当我尝试使用python requests发送时，它无法正常工作。出于某种原因，它等待带冒号的第一行，然后从那里开始发送所有行。例如，在上面的文件中，它仅POST：

3asdfaksdjfhlaksjdhflkjashdflkjhasldkjfhlaksdfhasdf
4asdfaksdjfhlaksjdhflkjashdflkjhasldkjfhlaksdfhasdf:

如何解决此问题？我不确定发生了什么。

以下是我的代码的简单版本：

import requests
import sys
import json
import os


token                    = 'nVQowAng0c'
url                      = "https://api.hipchat.com/v2/room/test_room/share/file"
headers                  = {'Content-type': 'multipart/related; boundary=boundary123456'}
headers['Authorization'] = "Bearer " + token


filepath = 'small_file.csv'
data     = open(filepath, 'rb').read()

payload = """\
--boundary123456
Content-Type: application/json; charset=UTF-8
Content-Disposition: attachment; name="metadata"
--boundary123456
Content-Disposition: attachment; name="file"; filename="{0}"
{1}
--boundary123456--\
""".format(os.path.basename(filepath), data)


r = requests.post(url, headers=headers, data=payload)
r.raise_for_status()

当我尝试在每行上发送带有时间戳的.csv文件之类的内容时，不会发送任何内容，因为每行都有冒号。

Answer 1

您的即时错误是您错误编码了MIME多部分元素。每个部分都有两个部分，标题和内容，两者之间有一个双换行符。你的错过第二个换行符，将其添加到：

payload = """\
--boundary123456
Content-Type: application/json; charset=UTF-8
Content-Disposition: attachment; name="metadata"

--boundary123456
Content-Disposition: attachment; name="file"; filename="{0}"

{1}
--boundary123456--\
""".format(os.path.basename(filepath), data)

我不是手动构建内容，而是重新使用requests-toolbelt project来让您以流媒体方式上传数据：

from requests_toolbelt import MultipartEncoder


class MultipartRelatedEncoder(MultipartEncoder):
    """A multipart/related encoder"""
    @property
    def content_type(self):
        return str(
            'multipart/related; boundary={0}'.format(self.boundary_value)
        )

    def _iter_fields(self):
        # change content-disposition from form-data to attachment
        for field in super(MultipartRelatedEncoder, self)._iter_fields():
            content_type = field.headers['Content-Type']
            field.make_multipart(
                content_disposition='attachment', 
                content_type=content_type)
            yield field


m = MultipartRelatedEncoder(
    fields={
        'metadata': (None, '', 'application/json; charset=UTF-8'),
        'file': (os.path.basename(filepath), open(filepath, 'rb'), 'text/csv'),
    }
)

headers['Content-type'] = m.content_type

r = requests.post(url, data=m, headers=headers)

我已调整requests_toolbelt.MultipartEncoder class以发出multipart/related数据流而不是multipart/form-data消息。

请注意，我传入了打开文件对象，而不是文件数据本身;这是因为MultipartEncoder允许您将数据流传输到远程服务器，因此不必将文件读入内存中。

您可能希望在metadata部分传递实际的JSON数据;用有效的JSON文档替换(None, '', 'application/json; charset=UTF-8'元组中的空字符串。

Answer 2

以下是来自@Martijn Pieters的合并代码：

# do this:
#     pip install requests_toolbelt

from os                import path
from sys               import exit, stderr
from requests          import post
from requests_toolbelt import MultipartEncoder


class MultipartRelatedEncoder(MultipartEncoder):
    """A multipart/related encoder"""
    @property
    def content_type(self):
        return str('multipart/related; boundary={0}'.format(self.boundary_value))

    def _iter_fields(self):
        # change content-disposition from form-data to attachment
        for field in super(MultipartRelatedEncoder, self)._iter_fields():
            content_type = field.headers['Content-Type']
            field.make_multipart(content_disposition = 'attachment',
                                 content_type        = content_type)
            yield field




def hipchat_file(token, room, filepath, host='api.hipchat.com'):

    if not path.isfile(filepath):
        raise ValueError("File '{0}' does not exist".format(filepath))


    url                      = "https://{0}/v2/room/{1}/share/file".format(host, room)
    headers                  = {'Content-type': 'multipart/related; boundary=boundary123456'}
    headers['Authorization'] = "Bearer " + token



    m = MultipartRelatedEncoder(fields={'metadata' : (None, '', 'application/json; charset=UTF-8'),
                                        'file'     : (path.basename(filepath), open(filepath, 'rb'), 'text/csv')})

    headers['Content-type'] = m.content_type

    r = post(url, data=m, headers=headers)

if __name__ == '__main__:

    my_token = <my token>   
    my_room  = <room name>    
    my_file  = <filepath>

    try:
        hipchat_file(my_token, my_room, my_file)
    except Exception as e:
        msg = "[ERROR] HipChat file failed: '{0}'".format(e)
        print(msg, file=stderr)
        exit(1)

Python：如果我的数据中有标点符号，请求将不会POST

2 个答案: