A mature platform for information applications
Haplo is a toolkit developed over the last six years by Haplo Services to build information applications for their clients. It’s a generally applicable platform, used in industries as diverse as healthcare, property, finance, publishing, and higher education.
We’re open sourcing the core platform so that it can become even more widely used, and to show that our interests are completely aligned with those of our clients.
In the coming months, we plan to release more of our toolkit, building on top of the core platform. This includes pluggable permission systems and extensive workflow support.
- A linked-data style object store, with pervasive multi-values and extensive hierarchy support
- An expressive label-based permissions model
- Web interface providing an exceptional user experience
- Customisable full-text search combined with object store graph queries
- File handling, including conversion, previewing and version control
- Lightweight “humane” workflow support
- Designed to evolve as user needs change, enabling agile application development
- Proven in production for over six years, used by a diverse range of clients
- Run on the Java Virtual Machine
- Primarily written in JRuby
- Object store is built on the PostgreSQL database and Xapian text indexing library
- Linked data / semantic web inspired object store
- Mix between graph database, object store, and search engine, with queries expressed as set operations
- Supports multi-values on every field
- Hierarchical relationships
- Permissions system labels every object with one or more labels, with access control written as boolean statements matching labels.
Key platform services
- Application user interface for editing and using a collection of “information”
- Management UI for configuration of the application, including the schema of the object store
- Plugin API
- Object store
- Relational database
- File store
- File processing and conversion
- Web application development framework
- HTML forms system for editing JSON document (includes repeating sections, separately open sourced)
- Access control using a label model
- Built in user database with users, groups, admin UI, etc
- Scriptable external authentication
- Audit trail
- Background tasks
- Low level HTTP access to allow custom handling of requests, as well a high level integration into the main platform UI.
- Standard UI components
- Plugin debugging and testing
The platform API encourages a programming style where:
- Plugins are small, and cooperate with other plugins to build up the applications
- Plugins share objects, operating on common types (eg “People”)
- Applications observe changes to the objects to react to changes made by users or other plugins
The Haplo server can be run in a special mode which adds plugin development tools. The Plugin Tool runs on your computer, observing changes to the plugin source on your local computer. When it notices a change, it automatically uploads the file and updates the server, for immediate testing without a server restart.
When you want to release the plugin, you use the plugin tool to ‘pack’ the plugin. This generates a production version with minified client side resources, and adds a manifest and version for use by the management systems.
For the open version, plugins are deployed by moving the packed plugin to the configured plugins directory. Plugins can be updated within a running server without a restart, allowing zero downtime for plugin updates.