Skip to main content

Django API layer

Project description

about

Django-wapiti is a generic API layer that can be added to any django project to allow other applications to get access to its models and methods very easily through JSON.

It is being developped by Ecometrica and made open source under a BSD license.

status

Django-wapiti is still under development. What works currently:

  • enabling the api for a given model

  • accessing objects by id

  • searching for objects through any of its fields, using any of the django queryset field matching operators (exact, iexact, contains, …)

  • auto-complete call allows frontend code to easily auto-complete user-entered data; it will search for objects through different fields

  • the ModelApi class, like ModelAdmin class for the django admin, is how you define the behavior an object should have through the API

  • instance methods are supported: simply decorate the method to make it available

  • class methods are also supported in the same manner

  • API keys

  • limiting requests per-{ip,session,user,apikey} per-{hour,day,month,year,ever} per-{http method, object, type, method}

  • slicing of results - add slice_left and slice_right parameters to your request to slice the results * the default max slice size is 100, override with settings.WAPITI_MAX_SLICE_SIZE

short-term plan

Because we need the following functionality, it likely will be incorporated soonish:

  • better method for defining the permissions for a given API key

  • user authentication (OAuth or otherwise) and row-level permissions

  • call logging (for quota and throttling enforcement)

  • api introspection: /api/VER/ should return a json (or XML) representation of the currently registered models and calls

known bugs

  • the JSON decoder will interpret both 2010-12-31 and “2010-12-31” as a date; it should interpret the latter as a string

usage

  1. install django-wapiti. the easiest way might be to use:

    pip install git+http://git@github.com/ecometrica/django-wapiti.git#egg=django-wapiti
  2. wapiti requires the decorator module to function:

    pip install decorator
  3. add wapiti to your INSTALLED_APPS in the settings.py file of your django project

  4. call wapiti.helpers.register_models from somewhere, usually your urls.py, to register your api models and methods

  5. to make the model class Foo from the app fooapp available through the API, create a file fooapp/api.py with this content:

    from wapiti.helpers import register
    from wapiti.modelapi import ModelApi
    
    from fooapp.models import Foo
    
    class FooApi(ModelApi):
        auto_complete_fields = ('name', 'description')
        auto_complete_order_by = 'name'
        model = Foo
    
    register('foo', FooApi)

    We’ve made the auto_complete method work by automatically searching for Foo objects with a case-insensitive search on the fields name and description (which therefore must exist in the Foo model definition).

    See the file wapiti/modelapi.py for a list of all options available in the ModelApi class for a given model.

  6. At this point, all objects of type Foo are available through the API at /api/1.0/foo/. To make an instance method available for calling through the API, apply the wapiti.helpers.api_method decorator to it. To call foo_method on the Foo object with ID 3, you would GET /api/1.0/foo/3/foo_method

  7. Similarly, for a class method, just decorate it with that same api_method, but apply the python classmethod decorator above it as well. To call the foo_cls_method on Foo through the API you would then GET /api/1.0/foo/foo_cls_method

  8. To make calls through the API, you’ll need API keys. From the main django admin interface, click on Add APIKey, add a name, and add a permission with resource_regex .* on method GET. All API calls need to have a k=THEAPIKEY parameter with a valid and active API key.

  9. All calls through JSON must have all their parameters be proper JSON! This means that to pass a string argument to a method, it needs to be surrounded by double-quotes! See the JSON spec for details.

  10. In order to use per-IP limiting, we need the REMOTE_ADDR variable. If you’re using nginx, this needs to be in the proper location section:

    fastcgi_param REMOTE_ADDR $remote_addr;

Release Notes

v0.0.1

  • Initial version

v0.0.2

  • Slicing of results

  • Compatible with Django 1.4 to 1.6

v0.0.3

  • post-South migrations

  • Compatible with Django 1.7 to 1.9

v1.0.0

  • Compatible with Django 1.8 and 1.9 (no longer compatible with Django =< 1.7)

v1.0.1

  • File handler now only takes field to better support non-FileSystemStorage files. Old way is still supported, but deprecated.

v1.0.2

  • Previous version forgot to handle the case of no file handler being supplied. This release fixes that.

v1.0.3

  • Allow serialising pre-1900 dates.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

django-wapiti-1.0.3.tar.gz (19.8 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page