Holy crap.

gistfile1.py

# Copyright (c) Microsoft Corporation.  All rights reserved.

"""Authorization helper functions.
"""

from hashlib import sha256
import hmac

import pydocumentdb.http_constants as http_constants


def GetAuthorizationHeader(document_client,
                           verb,
                           path,
                           resource_id,
                           resource_type,
                           headers):
    """Gets the authorization header.

    :Parameters:
        - `document_client`: document_client.DocumentClient
        - `verb`: str
        - `path`: str
        - `resource_id`: str
        - `resource_type`: str
        - `headers`: dict

    :Returns:
        dict, the authorization headers.

    """
    if document_client.master_key:
        return __GetAuthorizationTokenUsingMasterKey(verb,
                                                    resource_id,
                                                    resource_type,
                                                    headers,
                                                    document_client.master_key)
    elif document_client.resource_tokens:
        return __GetAuthorizationTokenUsingResourceTokens(
            document_client.resource_tokens, path, resource_id)


def __GetAuthorizationTokenUsingMasterKey(verb,
                                         resource_id,
                                         resource_type,
                                         headers,
                                         master_key):
    """Gets the authorization token using `master_key.

    :Parameters:
        - `verb`: str
        - `resource_id`: str
        - `resource_type`: str
        - `headers`: dict
        - `master_key`: st

    :Returns:
        dict

    """
    key = master_key.decode('base64')

    text = ((verb or '') + '\n' +
            (resource_type or '') + '\n' +
            (resource_id or '') + '\n' +
            (headers[http_constants.HttpHeaders.XDate]
             if http_constants.HttpHeaders.XDate in headers else '') + '\n' +
            (headers[http_constants.HttpHeaders.HttpDate]
             if http_constants.HttpHeaders.HttpDate in headers else '') + '\n')

    body = text.lower().decode('utf8')

    hm = hmac.new(key, body, sha256)
    signature = hm.digest().encode('base64')

    master_token = 'master'
    token_version = '1.0'
    return 'type=' + master_token +'&ver=' + token_version + '&sig=' + signature[:-1]


def __GetAuthorizationTokenUsingResourceTokens(resource_tokens,
                                              path,
                                              resource_id):
    """Get the authorization token using `resource_tokens`.

    :Parameters:
        - `resource_tokens`: dict
        - `path`: str
        - `resource_id`: str

    :Returns:
        dict

    """
    if resource_id in resource_tokens and resource_tokens[resource_id]:
        return resource_tokens[resource_id]
    else:
        path_parts = path.split('/')
        resource_types = ['dbs', 'colls', 'docs', 'sprocs', 'udfs', 'triggers',
                          'users', 'permissions', 'attachments', 'media',
                          'conflicts']
        for i in range(len(path_parts) - 1, -1, -1):
            if not path_parts[i] in resource_types and path_parts[i] in resource_tokens:
                return resource_tokens[path_parts[i]]
        return None

Hi whalesalad, I was the author of this SDK. It’s built in short period of time so not good enough yet. Any feedback on how I can improvement it?

@shipunyc Do you come from Java or C#?

@shipunyc It's probably that the code isn't "pythonic". Basically, it's written as if the language is some other language. It often revolves around using the good parts of Python. For example, instead of

path_parts = path.split('/')
for i in range(path_parts):
    do_something(path_parts[i])

a much smoother way of doing it would be:

for part in path.split('/'):
    do_something(part)

@shipunyc, please consider following the official style guide: http://legacy.python.org/dev/peps/pep-0008/.

You can automate style checking with these tools:
https://pypi.python.org/pypi/pep8
https://pypi.python.org/pypi/pylint

@peterbe, @smilliken, thanks for the feedback. Good to know there are style checking tools.
The code was translated from C# SDK and I was more of a C++ developer.

this gist is iconic... using some editors like sublime text or ... and let it do the linting and formatting