Java HTML Shell
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
archetype rename some method parameters in archetype May 10, 2018
neoemf-io add missing `.gitkeep` files May 4, 2018
neoemf-tests use `ServiceLoader` for `Context`s discovery in tests May 11, 2018
neoemf-utils update Eclipse P2 repository from `luna` to `oxygen` May 11, 2018
plugins/eclipse update Eclipse P2 repository from `luna` to `oxygen` May 11, 2018
.codecov.yml update CodeCov coverage range Feb 12, 2018
.dockerignore some refactoring in `Config`s Oct 17, 2017
.travis.yml update `` Apr 30, 2018
overview.html update the Javadoc and add missing `` files Nov 21, 2017


Build Status CodeCov Codacy Maven Central Javadoc Plugin Licence

Check out our wiki for further information on the installation, usages, supported backends, code examples, developer resources, etc.

What Is NeoEMF?

NeoEMF is an AtlanMod research team project that aims at handling large EMF models in an efficient and scalable way.

It is based on our work on Neo4EMF, a scalable and graph-based backend for persisting EMF models.

NeoEMF is a multi-backend framework developped to allow an easy integration of custom backends depending on user needs.

What Are the Features?

NeoEMF comes with some features depending on the backend used:

  • Lazy-loading mechanism: A model is loaded part by part while needed
  • Caching: NeoEMF relies on database caches to retrieve EObjects But in some situation where time performance is important this is not enough and it is possible to use application-level caches to speed up information loading
  • Auto-commit: For the backend having restrictions on transaction sizes it is possible to use the auto-commit feature to split large transaction into several small ones
  • Dirty saving: Handle large models that haven't been persisted to avoid memory overhead

Why Should I Use NeoEMF?

NeoEMF is fully EMF compliant, that means there is no modification to do to use it with existing EMF based applications.

Additionally NeoEMF can be plugged with several persistence solutions and is configurable to fit the best your needs (custom caches, in-memory resources, auto-commit,...).

In addition, NeoEMF provides a query API to perform optimized OCL queries over models and reify the results as navigable models.


Latest Release

The most recent release is NeoEMF 1.0.2, released May 21, 2017.

To add a dependency on NeoEMF using Maven, use the following:



Supported Implementations

All native implementations are located under the neoemf-data artifact. To add the dependency of the specific implementation you want to use, simply use:


Where {name} is any of:

  • Blueprints:
    • TinkerGraph : blueprints-core
    • Neo4j : blueprints-neo4j (requires blueprints-core)
  • MapDB : mapdb
  • BerkeleyDB: berkeleydb
  • HBase: hbase


Snapshots are automatically build from the master and are available throught Maven using 1.0.3-SNAPSHOT.


Dedicated pages are available in our wiki for the installation and the build.

Update Site

The simplest way to install NeoEMF is to use the update-site that is available here. Install the NeoEMF Persistence Framework component, which provides NeoEMF core classes and utils. Select the backend specific implementation(s) you want to plug in NeoEMF.

Local Build

You can build locally NeoEMF by following the instructions in the Build section. It is then possible to install NeoEMF from the local built update-site, or by importing the generated plugins.


A full benchmarks description is available at here.


If you experience issues installing or using NeoEMF, you can submit an issue on github or contact us at

Known issues:

  • Only Oxygen plugins are fetched (the compatibility is not ensured for previous Eclipse versions)
  • The Eclipse plugins can not be built if Maven can not access internet (it is needed to fetch Oxygen p2 repositories)
  • Sometimes the plugin build crashes and/or freezes during p2 index fetching from Oxygen repositories. It is generally sufficient to cancel the build (ctrl-c) and to relaunch it.


Performance problems and memory leaks are diagnosed with Java Profiler