What is KijiREST?
KijiREST provides a REST (Representational State Transfer) interface for Kiji, allowing applications to interact with a Kiji cluster over HTTP by issuing the four standard actions: GET, POST, PUT, and DELETE.
Use this interface to pull data from Kiji tables into other applications, such as web pages making use of predictive modeling data to customize user interactions.
There's an example of a KijiREST client as a RubyGem in the GitHub project kiji-rest-client which can be installed by executing:
gem install kijirest-client
Every REST request is parametrized by a resource path that uniquely identifies a server-side resource. Other optional parameters relevant to the action are specified in the query and body sections of the request. For example, to retrieve row data from the 'users' table, a GET request might look like the following:
All returned data is modeled as JSON. The appropriate MIME types to accept are application/json for lists, metadata, and all non-row-data.
To make use of the interface, you need to know how to identify the KijiREST resources and what operations can be applied to the resources. This guide describes these items as well as how to setup and run KijiREST in your development and production environments.
KijiREST Open Source Development Project
The kiji-rest project will evolve over time to ensure maximum user and developer benefit.
The Kiji project encourages the use of Java annotations for indicating the stability of API components. The levels of stability are as follows:
- By default, unlabeled classes should be assumed to be Experimental.
- Stable APIs are guaranteed to change only in binary-compatible ways within a major version (e.g., all 1.x.x versions)
- Evolving APIs may change in binary-incompatible ways between minor versions (e.g., from 1.4.1 to 1.5.0) of the software. In particular, new features may be added in one minor version and altered or removed in the next.
- Experimental APIs may change at any time.
- Private APIs should all be considered Experimental.