Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP

A dumb little wrapper around RESTful interfaces

tag: v0.8.0

Fetching latest commit…

Cannot retrieve the latest commit at this time

README.markdown

Dolt

A dumb little wrapper around RESTful interfaces

What is Dolt?

Dolt is a minimalist wrapper around RESTful interfaces, specifically, JSON RESTful interfaces. Instead of adding another layer on top of the calls, this uses Httplib2 and some Python magic to allow truly simple wrappers on top of already well thought out (at least sometimes) REST APIs.

For example, let's look at the Twitter API call for grabbing a user. We'll use my user, tswicegood

http://twitter.com/users/show.json?screen_name=tswicegood

python-twitter maps that to:

twitter = twitter.Api()
twitter.GetUser("tswicegood")

So where'd this come from? The creators of python-twitter. You have to read their documentation to figure out what call to make, because the api docs don't mention GetUser anywhere.

Here's what that same call in the Twitter Dolt object as shown in the examples/ directory:

twitter = Twitter()
twitter.users.show(screen_name="tswicegood")

Notice the similarities? This way, you can use the official API docs to figure out how your'e supposed to interact with the object that represents it. No extra documentation needed.

How to handle non-GET methods

Dolt can handle all sorts of methods in its requests, not just GET method requests. To do that, you have to muddy up the API call just a bit.

Here's an example making a post to update the status. First, the Twitter API:

http://twitter.com/statuses/update.json

Now the Dolt version:

twitter = Twitter()
twitter.statuses.update.with_body(status="Hello from Dolt!").POST()

Notice that all you need to add to it is the method you want to call. If you're feeling very Pythonic and want to be explicit in every call, you can add .GET as the final method call, though that's always assumed.

Sometimes having that POST or PUT at the end seems weird. You can stick the method wherever you want in the call string of properties, it just has to be in all uppercase. For example, this works just the same as the previous code:

twitter = Twitter()
twitter.POST.with_body(status="Hello from Dolt!").statuses.update()

This works for other HTTP methods as well, such as PUT, DELETE, and HEAD.

JSON Handling

Dolt will automatically decode JSON if the response uses one of the JSON content-types and return a dict.

Dolt can also send JSON:

twitter = Twitter()
twitter.statuses.update.with_json(status="Hello from Dolt!").POST()

Sending Headers

Dolt can send headers with the request:

api = Dolt()
api.foo.with_headers(Accept='text/html').GET()

Handling authentication

Dolt relies on the httplib2 project for its underlying HTTP requests. Httplib2 has an add_credentials method that allows you to add credentials to it. Dolt takes an http parameter in its __init__ method which allows you to pass in an Http object with credentials. For example:

http = Http()
http.add_credentials("some_user", "secret")
some_api = Dolt(http=http)

You can also use the add_basic_auth helper:

from dolt.helpers import add_basic_auth
some_api = Dolt()
some_api = add_basic_auth(some_api, username, password)

Using dictionary-style lookups

Not all sections of a path can be represented directly as part of a string. You can use a dict-like syntax to add segments to the URL. For example, you can use the following:

from dolt.apis.couchdb import CouchDB
couch = CouchDB("awesome")
couch["_design/posts"]["_list/all"]()

That is equivalent to:

couch._design.posts._list.all()

Re-using requests

When chainging attributes or using with_* functions, Dolt returns a clone of the current state. This allows you to safely re-use Dolt requests for batch processing.

http = Http()
http.add_credentials("some_user", "secret")
some_api = Dolt(http=http)

update = some_api.collection.with_params(api_key=API_KEY).PUT
update.with_json(name='Foo')(id=123)
update.with_json(name='Bar')(id=345)

The Http connection is re-used for each request along with the path parts and params.

item = some_api.admin.item.with_params(api_key=API_KEY)
item[uid1].DELETE()
# DELETE /admin/item/<uid1>?api_key=<API_KEY>
item[uid2].comments.DELETE()
# DELETE /admin/item/<uid2>/comments?api_key=<API_KEY>

Included APIs

Dolt's primary purpose is to make it easy to wrap other APIs, but there are a few extras that are included as part of the distribution.

In alphabetical order:

  • bit.ly
  • CouchDB
  • Disqus
  • GitHub
  • Mosso/Rackspace Cloud
  • Statsmix
  • Twitter

Contributing

All development happens on GitHub. Follow these simple steps to get your code added:

  • Create something awesome -- make the code better, add a new API, whatever (this is the hardest part).
  • Fork Dolt
  • Create a topic branch to house your changes
  • Get all of your commits in the new topic branch
  • Submit a pull request.

Optional Steps

  • Offer to buy me a beer. :-)
  • Or Make sure your pull request contains all of the relevant unit tests to verify your code with.

Running the test suite

  • Install the requirements/development.txt file:

    pip install -r requirements/development.txt

  • run the tests:

    cd Dolt/tests python unit_tests.py

License

As of v0.3, this work is licensed under the BSD.

Of course, I'm not a lawyer, so if you have questions about the licensing and how it effects your use, please consult a professional lawyer, not some random README file.

Something went wrong with that request. Please try again.