Skip to content

plone/plone.rest

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Github Actions Status Coveralls github Egg Status Latest Version License

Plone REST

Purpose

plone.rest allows you to use HTTP verbs such as GET, POST, PUT, DELETE, etc. in Plone.

REST stands for Representational State Transfer. It is a software architectural principle to create loosely coupled web APIs.

plone.rest provides the basic infrastructure that allows us to build RESTful endpoints in Plone.

The reason for separating this infrastructure into a separate package from the 'main' full Plone REST API is so you can create alternative endpoints tailored to specific usecases. A number of these specific endpoints are already in active use.

Audience

plone.rest is for experienced web developers who want to build their own HTTP/REST endpoints on top of Plone.

If you want to use a ready-made full RESTful Plone API, you should use plone.restapi. That package uses, and depends upon, this one.

Features

  • Registering RESTful service endpoints for the following HTTP verbs:
    • GET
    • POST
    • PUT
    • DELETE
    • PATCH
    • OPTIONS
  • Support for Dexterity and Archetypes-based content objects
  • Content negotiation: Services can be registered for arbitrary media types (e.g. 'application/json').
  • Named services allows to register service endpoints for custom URLs

Registering RESTful Service Endpoints

plone.rest allows you to register HTTP verbs for Plone content with ZCML.

This is how you would register a PATCH request on Dexterity content:

<plone:service
  method="PATCH"
  accept="application/json"
  for="plone.dexterity.interfaces.IDexterityContent"
  factory=".service.Patch"
  permission="cmf.ModifyPortalContent"
  />

You have to specify the HTTP verb (GET, POST, PUT, DELETE, HEAD, OPTIONS), the media type used for content negotiation, the interface for the content objects, the factory class that actually returns the content and the permission required to access the service.

The factory class needs to inherit from the plone.rest 'Service' class and to implement a render method that returns the body of the response:

from plone.rest import Service

class Patch(Service):

    def render(self):
        return '{"message": "PATCH: Hello World!"}'

Content Negotiation

To access the service endpoint we created we have to send a GET request to a Dexterity object:

  1. by setting the 'Accept' header to 'application/json':

    PATCH /Plone/doc1 HTTP/1.1
    Host: localhost:8080
    Accept: application/json
    
  2. or by adding an explicit traversal step to the path like so:

    PATCH /Plone/++api++/doc1 HTTP/1.1
    Host: localhost:8080
    

The server then will respond with '200 OK':

HTTP/1.1 200 OK
Content-Type: application/json

{
  "message": "PATCH: Hello World!"
}

Why two methods? Using the 'Accept' header is the intended way of RESTful APIs to get different responses from the same URL. However, if it comes to caching the response in an web accelerator like Varnish or Cloudflare, additional challenges are added. Setting the Vary header to 'Vary: Accept' helps to a certain degree in Varnish. But cache pollution may happen, because different browsers send different headers on normal HTML requests. Hosted services like Cloudflare just do not support the 'Vary' usage and can not be used for sites with REST calls. Thus a second option with different URLs is needed.

You can try this out on the command line:

$ http --auth admin:admin PATCH localhost:8080/Plone/doc1 Accept:application/json

Note

You have to install httpie (pip install httpie) to make this example work.

Here is a list of examples for all supported HTTP verbs:

GET:

$ http --auth admin:admin GET localhost:8080/Plone/doc1 Accept:application/json
$ http --auth admin:admin GET localhost:8080/Plone/++api++/doc1

POST:

$ http --auth admin:admin POST localhost:8080/Plone/doc1 Accept:application/json
$ http --auth admin:admin POST localhost:8080/Plone/++api++/doc1

PUT:

$ http --auth admin:admin PUT localhost:8080/Plone/doc1 Accept:application/json
$ http --auth admin:admin PUT localhost:8080/Plone/++api++/doc1

DELETE:

$ http --auth admin:admin DELETE localhost:8080/Plone/doc1 Accept:application/json
$ http --auth admin:admin DELETE localhost:8080/Plone/++api++/doc1

PATCH:

$ http --auth admin:admin PATCH localhost:8080/Plone/doc1 Accept:application/json
$ http --auth admin:admin PATCH localhost:8080/Plone/++api++/doc1

OPTIONS:

$ http --auth admin:admin OPTIONS localhost:8080/Plone/doc1 Accept:application/json
$ http --auth admin:admin OPTIONS localhost:8080/Plone/++api++/doc1

Named Services

Named services can be registered by providing a 'name' attribute in the service directive:

<plone:service
  method="GET"
  accept="application/json"
  for="Products.CMFPlone.interfaces.IPloneSiteRoot"
  factory=".service.Search"
  name="search"
  permission="zope2.View"
  />

This registers a service endpoint accessible at the site root using the following request:

GET /Plone/search HTTP/1.1
Host: localhost:8080
Accept: application/json

Service endpoints can also be accessed using the traverser:

GET /Plone/++api++/search HTTP/1.1
Host: localhost:8080

Additional Path Segments

To handle additional path segments after the service url like /Plone/myservice/1/2 a service has to implement IPublishTraverse. The following example simply stores all path segments in an array in self.params.

from plone.rest import Service
from zope.interface import implementer
from zope.publisher.interfaces import IPublishTraverse

@implementer(IPublishTraverse)
class MyService(Service):

    def __init__(self, context, request):
        super(MyService, self).__init__(context, request)
        self.params = []

    def publishTraverse(self, request, name):
        self.params.append(name)
        return self

    def render(self):
        return {'service': 'named get', 'params': self.params}

See also the implementation of the workflow transition endpoint in plone.restapi for an other example.

CORS

plone.rest allows you to define CORS policies for services in ZCML. The following example defines a policy for all services.

<plone:CORSPolicy
  allow_origin="http://example.net"
  allow_methods="DELETE,GET,OPTIONS,PATCH,POST,PUT"
  allow_credentials="true"
  expose_headers="Content-Length,X-My-Header"
  allow_headers="Accept,Authorization,Content-Type,X-Custom-Header"
  max_age="3600"
  />

CORS policies can be bound to specific interfaces of content objects and to specific browser layers. This allows us to define different policies for different content types or to override existing policies. The following example defines a policy for the site root.

<plone:CORSPolicy
  for="Products.CMFPlone.interfaces.IPloneSiteRoot"
  layer="myproduct.interfaces.IMyBrowserLayer"
  allow_origin="*"
  allow_methods="GET"
  />

The CORSPolicy directive supports the following options:

allow_origin
Origins that are allowed to access the resource. Either a comma separated list of origins, e.g. "http://example.net,http://mydomain.com" or "*".
allow_methods
A comma separated list of HTTP method names that are allowed by this CORS policy, e.g. "DELETE,GET,OPTIONS,PATCH,POST,PUT". If not specified, all methods for which there's a service registered are allowed.
allow_credentials
Indicates whether the resource supports user credentials in the request.
allow_headers
A comma separated list of request headers allowed to be sent by the client, e.g. "X-My-Header"
expose_headers
A comma separated list of response headers clients can access, e.g. "Content-Length,X-My-Header".
max_age
Indicates how long the results of a preflight request can be cached.
for
Specifies the interface for which the CORS policy is registered. If this attribute is not specified, the CORS policy applies to all objects.
layer
A browser layer for which this CORS policy is registered. Useful for overriding existing policies or for making them available only if a specific add-on has been installed.

Installation

Install plone.rest by adding it to your buildout:

[buildout]

 ...

 eggs =
     plone.rest

and then running "bin/buildout"

Plone/Python Support

plone.rest 4.x.x supports Plone 5.2 and 6.x on Python 3.8 and newer.

plone.rest 3.x.x supports Plone 5.2 on Python 2.7 and 3.6 to 3.8 and Plone 6.0 on Python 3.8 to 3.11.

If you need to use Plone 4.3, 5.0, or 5.1 on Python 2.7, check out plone.rest 2.x.x or 1.x.x.

Redirects

plone.rest will handle redirects created by plone.app.redirector pretty much the same way as regular Plone.

If a redirect exists for a given URL, a GET request will be answered with 302, and the new location for the resource is indicated in the Location header:

HTTP/1.1 302 Moved Temporarily

Content-Type: application/json
Location: http://localhost:8080/Plone/my-folder-new-location

Any other request method than GET (POST, PATCH, ...) will be answered with 307 Temporary Redirect. This status code instructs the client that it should NOT switch the method, but retry (if desired) the request with the same method at the new location.

In practice, both the Python requests library a well as Postman seem to honour this behavior by default.

Contribute

Support

This package is maintained by Timo Stollenwerk <[email protected]>.

If you are having issues, please let us know.

Credits

plone.rest has been written by Timo Stollenwerk (kitconcept GmbH) and Ramon Navarro Bosch (Iskra).

plone.rest was added as a Plone core package with Plone 5.2 (see plone/Products.CMFPlone#2177).

License

The project is licensed under the GPLv2.