One of the guiding principles in developing the connector is simplicity and keeping the core small and efficient with as little external dependencies as possible to avoid version conflicts. We do not want to force any third-party dependencies onto our users, so we aim to avoid any of the big frameworks. Of course, if you want to use them, you still can add them to your extensions (see: [TBW]). The connector is a plain Java application built with Gradle, but it can be embedded into any form of application deployment.
Developer documentation can be found under docs/developer, where the main concepts and decisions are captured as decision records.
Some more documentation can be found at extensions, launchers and the samples repository.
For detailed information about the whole project, please take a look at our GitHub pages.
If you are not yet familiar with the EDC nomenclature, we strongly advise reading the documentation and/or watching the introductory videos.
The Samples repository provides a set of examples for getting familiar with the EDC framework and learning how to set up and use a connector. The samples begin with the very basics (e.g. learning how to write an extension or use a configuration file) and then move on to more complex scenarios (e.g. performing different data transfers). More samples will be added in the future, so be sure to check back regularly.
Official versions are available through MavenCentral
.
Please add the following instructions in your build.gradle[.kts]
file (if not already present):
repositories {
mavenCentral()
// ... other maven repos
}
We strongly recommend to use official versions and only switch to snapshots if there is a clear need to do so, or you've been instructed to do so, e.g. to verify a bugfix.
All artifacts are under the org.eclipse.edc
group id, for example:
dependencies {
implementation("org.eclipse.edc:spi:core-spi:<<version>>")
// any other dependencies
}
In addition, EDC regularly publishes snapshot versions, which are available at Sonatype's snapshot repository. In order to add them to your build configuration, simply add this:
repositories {
mavenCentral()
maven {
url = uri("https://oss.sonatype.org/content/repositories/snapshots/")
}
// any other repos
}
Then you can add snapshot dependencies by simply using the -SNAPSHOT
version suffix:
dependencies {
implementation("org.eclipse.edc:spi:core-spi:0.4.2-SNAPSHOT")
// any other dependencies
}
You may check MavenCentral for a comprehensive list of all official versions.
Please be aware of the following pitfalls:
- snapshots are by definition unstable - every new snapshot replaces an old one
- this may cause unrepeatable builds
- snapshots are created irregularly, we do not have any fixed publish schedule
We plan to have actual release versions starting some time mid 2022. Please check back soon.
For more information about versioning please refer to the release documentation
The project requires JDK 11+. To get started:
git clone [email protected]:eclipse-edc/Connector.git
cd Connector
./gradlew clean build
That will build the connector and run unit tests.
If you wish to configure your IDE/editor to automatically apply the EDC code style, please follow this guide.
Note: the style guide will be checked/enforced in GitHub Actions.
Please refer to this document.
This is the primary extension point for the connector. It contains all necessary interfaces that need to be implemented
as well as essential model classes and enums. Basically, the spi
modules defines the extent to what users can
customize and extend the code.
Contains all absolutely essential building that is necessary to run a connector such as TransferProcessManager
,
ProvisionManager
, DataFlowManager
, various model classes, the protocol engine and the policy piece. While it is
possible to build a connector with just the code from the core
module, it will have very limited capabilities to
communicate and to interact with a data space.
This contains code that extends the connector's core functionality with technology- or cloud-provider-specific code. For example a transfer process store based on Azure CosmosDB, a secure vault based on Azure KeyVault, etc. This is where technology- and cloud-specific implementations should go.
If someone were to create a configuration service based on Postgres, then the implementation should go into
the extensions/database/configuration-postgres
module.
Launchers are essentially connector packages that are runnable. What modules get included in the build (and thus: what
capabilities a connector has) is defined by the build.gradle.kts
file inside the launcher subdirectory. That's also
where a Java class containing a main
method should go. We will call that class a "runtime" and in order for the
connector to become operational the runtime
needs to perform several important tasks (="bootstrapping"). For an
example take a look at
this runtime
Contains implementations for communication protocols a connector might use, such as DSP.
GitHub releases are listed here. Please find more information about releases in our release approach.
See here for more information about project and milestone planning. Scheduled and ongoing milestones are listed here.
Available tags can be found here.
See how to contribute.