roelfsche / jelly

A flexible ORM for Kohana 3.1+

Jelly is a nice little ORM for Kohana 3.1+. The project was originally started by Jonathan Geiger and co-developed by Paul Banks.

IMPORTANT

Critical to know:

• use the 3.x/master branches for production as the 3.x/develop branches are subject to frequent and major changes
• userguide is being updated

Requirements

Jelly requires the following Kohana versions per Git branch:

• 3.1/develop and 3.1/master branches: Kohana 3.1.3+
• 3.2/develop and 3.2/master branches: Kohana 3.2
• 3.3/master branch: Kohana 3.3+

Useful stuff:

• Jelly driver for Kohana Auth
• Report an issue or feature request
• Leave feedback in the forum

Get involved in Jelly's developement

As Jelly has always been a community project it's development and future depends on people who are willing to put some time into it. The easiest way to contribute is to fork the project.

Remember:

• you can directly edit files on GitHub (look for the Edit this file button), there's no need to get familiar with Git if you don't want to
• please follow the Kohana conventions for coding
• read the introduction to the unit tests in the guide and run them if you make changes to Jelly to minimalize the chances of introducing new bugs
• and thanks for helping Jelly become better!

Notable Features

• Standard support for all of the common relationships — This includes belongs_to, has_many, and many_to_many. Pretty much standard these days.

• Top-to-bottom table column aliasing – All references to database columns and tables are made via their aliased names and converted transparently, on the fly.

• Active testing on MySQL and SQLite — All of the Jelly unit tests work 100% correctly on both MySQL, SQLite and PostgresSQL databases.

• A built-in query builder — This features is a near direct port from Kohana's native ORM. I find its usage much simpler than Sprig's.

• Extensible field architecture — All fields in a model are represented by a Field_* class, which can be easily overridden and created for custom needs. Additionally, fields can implement behaviors that let the model know it has special ways of doing things.

• No circular references — Fields are well-designed to prevent the infinite loop problem that sometimes plagues Sprig. It's even possible to have same-table child/parent references out of the box without intermediate models.