Option to allow pretty print without alpha sorting

[patcon]

Right now, the options are colors, format, and all, but often the order of the json returned is intentional and for clarify. It would be nice to be able to pretty print the json without having it's keys sorted :)

Awesome tool, and thanks very much for everything!

[kracekumar]

Results from my introspection.

➜  httpie git:(issue-128) http http://headers.jsontest.com/
HTTP/1.1 200 OK
Access-Control-Allow-Origin: *
Cache-Control: private
Content-Type: application/json; charset=ISO-8859-1
Date: Sat, 27 Jul 2013 07:01:24 GMT
Server: Google Frontend
Transfer-Encoding: chunked

{
    "Accept": "*/*",
    "Host": "headers.jsontest.com",
    "User-Agent": "HTTPie/0.6.0"
}

➜  httpie git:(issue-128) curl http://headers.jsontest.com/
{
   "Host": "headers.jsontest.com",
   "User-Agent": "curl/7.24.0 (x86_64-apple-darwin12.0) libcurl/7.24.0 OpenSSL/0.9.8x zlib/1.2.5",
   "Accept": "*/*"
}

Since httpie uses requests guessing r.json() is responsible for this, trying requests.get.

In [36]: r = requests.get("http://headers.jsontest.com/")

In [37]: r.json()
Out[37]:
{u'Accept': u'*/*',
 u'Host': u'headers.jsontest.com',
 u'User-Agent': u'python-requests/1.2.3 CPython/2.7.2 Darwin/12.4.0'}

Same result. Now looking intor.json code.

In [38]: r.json??
Type:       instancemethod
String Form:<bound method Response.json of <Response [200]>>
File:       /Library/Python/2.7/site-packages/requests/models.py
Definition: r.json(self, **kwargs)
Source:
    def json(self, **kwargs):
        """Returns the json-encoded content of a response, if any.

        :param \*\*kwargs: Optional arguments that ``json.loads`` takes.
        """

        if not self.encoding and len(self.content) > 3:
            # No encoding set. JSON RFC 4627 section 3 states we should expect
            # UTF-8, -16 or -32. Detect which one to use; If the detection or
            # decoding fails, fall back to `self.text` (using chardet to make
            # a best guess).
            encoding = guess_json_utf(self.content)
            if encoding is not None:
                return json.loads(self.content.decode(encoding), **kwargs)
        return json.loads(self.text or self.content, **kwargs)

Look into requests code json is imported from compat.py, the exact code is

try:
    import simplejson as json
except ImportError:
    import json

Trying other library like ujson provides same result

n [43]: ujson.loads("""{
   "Host": "headers.jsontest.com",
   "User-Agent": "curl/7.24.0 (x86_64-apple-darwin12.0) libcurl/7.24.0 OpenSSL/0.9.8x zlib/1.2.5",
   "Accept": "*/*"
}""")
Out[43]:
{u'Accept': u'*/*',
 u'Host': u'headers.jsontest.com',
 u'User-Agent': u'curl/7.24.0 (x86_64-apple-darwin12.0) libcurl/7.24.0 OpenSSL/0.9.8x zlib/1.2.5'}

Since the the returned object is dictionary it isn't possible to maintain the order due to hashing. If it is possible to maintain the order it would be great.

[kracekumar]

Seems like I made a mistake. There is a JSONProcessor but still it might not solve the problem.

In [48]: r.content
Out[48]: '{\n   "Host": "headers.jsontest.com",\n   "User-Agent": "python-requests/1.2.3 CPython/2.7.2 Darwin/12.4.0",\n   "Accept": "*/*"\n}\n'

In [49]: json.loads(r.content)
Out[49]:
{u'Accept': u'*/*',
 u'Host': u'headers.jsontest.com',
 u'User-Agent': u'python-requests/1.2.3 CPython/2.7.2 Darwin/12.4.0'}

Trying with JSONProcessor

In [61]: j = JSONProcessor()

In [62]: j.process_body(r.content, r.headers['Content-Type'], subtype='json', encoding='utf8')
Out[62]: u'{\n    "Accept": "*/*", \n    "Host": "headers.jsontest.com", \n    "User-Agent": "python-requests/1.2.3 CPython/2.7.2 Darwin/12.4.0"\n}'

[kracekumar]

One way to get around the problem is use PygmentsProcessor.

In [64]: p = PygmentsProcessor()

In [65]: p.process_body(r.content, r.headers['Content-Type'], subtype='json', encoding='utf8')
Out[65]: u'\x1b[38;5;245m{\x1b[39m\n\x1b[38;5;245m   \x1b[39m\x1b[38;5;33m"Host"\x1b[39m\x1b[38;5;245m:\x1b[39m\x1b[38;5;245m \x1b[39m\x1b[38;5;37m"headers.jsontest.com"\x1b[39m\x1b[38;5;245m,\x1b[39m\n\x1b[38;5;245m   \x1b[39m\x1b[38;5;33m"User-Agent"\x1b[39m\x1b[38;5;245m:\x1b[39m\x1b[38;5;245m \x1b[39m\x1b[38;5;37m"python-requests/1.2.3 CPython/2.7.2 Darwin/12.4.0"\x1b[39m\x1b[38;5;245m,\x1b[39m\n\x1b[38;5;245m   \x1b[39m\x1b[38;5;33m"Accept"\x1b[39m\x1b[38;5;245m:\x1b[39m\x1b[38;5;245m \x1b[39m\x1b[38;5;37m"*/*"\x1b[39m\n\x1b[38;5;245m}\x1b[39m'

In [66]: print(p.process_body(r.content, r.headers['Content-Type'], subtype='json', encoding='utf8'))
{
   "Host": "headers.jsontest.com",
   "User-Agent": "python-requests/1.2.3 CPython/2.7.2 Darwin/12.4.0",
   "Accept": "*/*"
}

[patcon]

Nice. I recently used something called an OrderedHash is ruby that solved a similar problem. Thinking that OrderedDict might be the solution in python...
http://stackoverflow.com/questions/6921699/can-i-get-json-to-load-into-an-ordereddict-in-python

[kracekumar]

@patcon That was easier solution than mine. I will modify the code accordingly and update the pull request.

[patcon]

👍 Closing this one :)

[kracekumar]

@patcon Why?, pull request isn't merged though.

[patcon]

Ah. Oops. Misunderstood the context of the referring issue above.

[carueda]

👍 veeery useful option, please merge/fix at the next convenient opportunity. If not "reasonably" soon, any idea of by when approximately?

[pontuspalmenas]

+1
Please add this option. Sorting is a deal breaker for me.

[altfatterz]

+1
Yes, please add this option. I would like to have pretty print but without sorting the keys

[rawaludin]

Any news about this?