urllib.Request 删除 Content-Type 标头

Question

I want to remove the content-type header from a POST request. I have tried setting the header to ''

try:
    from urllib.request import Request, urlopen
except ImportError:
    from urllib2 import Request, urlopen

url = 'https://httpbin.org/post'
test_data = 'test'

req = Request(url, test_data.encode(), headers={'Content-Type': ''})
req.get_method = lambda: 'POST'
print(urlopen(req).read().decode())

But this sends:

{
  // ...
  "headers": {
     "Content-Type": "", // ...
  }
  // ...
}

What I would like it to do is send no Content-Type at all, instead of a blank one. By default, it is application/x-www-form-urlencoded .

This can easily be acheived with requests :

print(requests.post(url, test_data).text)

But this is a script I need to distribute, so cannot have dependencies. I need it to have no Content-Type at all, since the server is very picky, so I cannot use text/plain or application/octet-stream .

Answer 1

You can specify custom handler:

try:
    from urllib.request import Request, urlopen, build_opener, BaseHandler
except ImportError:
    from urllib2 import Request, urlopen, build_opener, BaseHandler

url = 'https://httpbin.org/post'
test_data = 'test'

class ContentTypeRemover(BaseHandler):
    def http_request(self, req):
        if req.has_header('Content-type'):
            req.remove_header('Content-type')
        return req
    https_request = http_request

opener = build_opener(ContentTypeRemover())
req = Request(url, test_data.encode())
print(opener.open(req).read().decode())

---

Another (hacky) way: monkey-patching the request object to pretend there's Content-type header already there; prevents AbstractHTTPHandler to default Content-Type header.

req = Request(url, test_data.encode())
req.has_header = lambda header_name: (header_name == 'Content-type' or
                                      Request.has_header(req, header_name))
print(urlopen(req).read().decode())

Answer 2

Just to add ontop of @falsetru's answer,

If you need to filter more headers than this, req.headers is not the thing you want, it's more like this:

class ContentTypeRemover(BaseHandler):
    def __init__(self, headers_to_filter={'Content-type'}): # set or dict works
        self.headers_to_filter = headers_to_filter

    def http_request(self, req):
        for header, value in req.header_items():
            if header in self.headers_to_filter:
                req.remove_header(header)
        return req
    https_request = http_request

And if you need to modify a header, not remove it.. It gets a bit weird (at least this is the only method I've found to be working):

req.remove_header(header)
req.add_header(header, self.modify_headers[header.lower()])