如何使用编码的unicode字符对一个python字典的值进行urlencode

Question

I'm trying to make a url-encoded web request in python 2.7 where I want to send a list of python dictionaries that would be on the server decoded as a list of JSON objects. In essence I'm making:

>>>urllib.urlencode({"param":"val", "items":[item1, item2] }, True)

where item1 can be something like { "a": u"š".encode("utf8") } (simplified for the example)

The problem arises because of the unicode characters.

If an item1 is by itself encoded, you get something meaningful:

>>>urllib.urlencode(item1)
'a=%C5%A1'

however, if I call urllib.urlencode({"test": item1}) I get a mess:

'test=%7B%27a%27%3A+%27%5Cxc5%5Cxa1%27%7D'

In this case, the unicode character is no longer encoded as %C5%A1 but as a longer sequence that is then incorrectly decoded on the server side.

Does anybody have a suggestion how to properly transform complex dictionary values (ie item1 ) before calling urlencode to avoid this issue?

Answer 1

One way or another you need to decode anything that was encoded before re-encoding Here is one approach:

dictionary = {"test": item1}
urllib.urlencode(dict([(k, decode_operation(v)) for k, v in dictionary.iteritems()]))

Answer 2

I solved my problem by first calling json.dumps(item) for each item in item list and then calling urllib.urlencode.

In short:

>>>urllib.urlencode({"param":"val", "items":[json.dumps(item) for item in items] }, True)

This solved the problem because the unicode characters are then encode using the \\uXXXX which can be on the server side properly decoded with a json parser.