1.3.1 is our final feature release of Smart Mastering in the Smart Mastering Core repository. As of Data Hub 5.0.0, Smart Mastering is fully integrated into MarkLogic Data Hub as a built-in capability, and the recommended way to use the Smart Mastering capability is by configuring a mastering step in Data Hub. Existing users should migrate their Smart Mastering configuration to MarkLogic Data Hub (see Import Your Smart Mastering Core Projects for instructions). The integration of Smart Mastering into Data Hub offers a variety of benefits, including:
- Built-in support for orchestrating matching and merging across documents.
- QuickStart UI for configuration of matching and merging
MarkLogic will continue to invest in Smart Mastering as a built-in capability of Data Hub.
This repo contains the libraries and services for a Smart Mastering capability built on top of MarkLogic. Smart Mastering consists of matching entities in an operational data hub, then auto-merging for high-scoring matches and recording a notification for a human reviewer for cases where the score indicates a possible, but not definite match. Match scoring rules and merging algorithms, thresholds and actions are configuration driven. APIs are available either through a set of XQuery libraries or a REST service layer.
This capability is experimental. Be prepared for the interface and implementation to change significantly. We welcome your input to guide this development process.
- MarkLogic 9.0-5 or higher
- Java 8 or higher
- Gradle is optional - this project has the Gradle wrapper included, and the instructions below reference it so that you don't need to install Gradle
- Schemas Database attached to your content DB. This is required for PROV-O to work.
Documentation on how Smart Mastering Core works and how to use it is available at GitHub.
To use the Smart Mastering Core in your own project, follow these instructions. This assumes that you're using ml-gradle in your project.
Note: be advised that this project is in its very early stages. The APIs presented here may change significantly before stabilizing.
To view smart-mastering-core in use, see the Smart Mastering project examples.
If you've found a bug or would like to ask for a new capability, please file an issue here on GitHub. If you are having trouble using smart-mastering-core, you can file a question issue here on GitHub or ask a question on Stack Overflow with the "marklogic" tag. If you'd like to discuss this project with Product Management, contact community-requests@marklogic.com.
Smart Mastering Core is a community-supported project. Help is available by filing issues here on GitHub and by asking questions on Stack Overflow; however, we can’t promise a specific resolution or timeframe for any request.
Assuming you're using ml-gradle, you can easily integrate Smart Mastering into your application.
As this project hasn't been published to the jcenter repository yet, you'll first need to publish a copy of this project to your local Maven repository, which defaults to ~/.m2/repository.
To do so, clone this repository and run the following command in the project's root directory:
./gradlew publishToMavenLocal
You can verify that the artifacts were published successfully by looking in the ~/.m2/repository/com/marklogic/community/smart-mastering-core directory.
Now that you've published Smart Mastering locally, you can add it to your own application. The minimal example project provides a simple example of doing this. You just need to add the following to your build.gradle (again, this depends on using ml-gradle).
First, in the repositories block, make sure you have your local Maven repository listed:
repositories {
mavenLocal()
}
And then just add the following to your dependencies block:
dependencies {
mlRestApi "com.marklogic.community:smart-mastering-core:0.1.DEV"
}
This assumes that the version of the artifacts you published above is 0.1.DEV. You can find the version number by looking at the version property in gradle.properties in your cloned copy of smart-mastering-core.
And that's it! Now, when you run mlDeploy, the modules in Smart Mastering will be automatically loaded into your modules database. To verify that the modules exist, you can either browse your modules database via qConsole, or you can go to your application's REST server and see that the Smart Mastering services have been installed:
http://localhost:(your REST port)/v1/config/resources
You can then run ml-gradle tasks such as mlLoadModules and mlReloadModules,
and the Smart Mastering modules will again be loaded into your modules database.
If necessary, create a gradle-local.properties file and override properties in
gradle.properties as needed.
Run ./gradlew mlDeploy
After running ./gradlew mlDeploy, point a browser to http://localhost:8042/test.
Click the Run Tests button.
./gradlew mlUnitTest
- update gradle.properties with the new version number
- update the gradle.properties for the examples with the new version number
- update the UPGRADE.md file's version reference to the new version number. Add other upgrade notes as needed.
- generate the CHANGELOG:
github_changelog_generator --token $my-github-token --future-release v1.0.0- if you don't have a "/tmp/" directory (Windows), add
--cache-file ~\tmp\github-changelog-http-cache --cache-log ~\tmp\github-changelog-logger.log
- if you don't have a "/tmp/" directory (Windows), add
- commit the CHANGELOG
- add these properties to your
gradle-local.properties- bintray_user
- bintray_key
- change the
versionproperty ingradle.propertiesto the new version number ./gradlew bintrayUpload- log in to bintray and push the publish button
- smoke test the examples to make sure they work with the latest published version
- merge the develop branch into master
- merge the docs-next branch into docs
- tag the release vx.xx
- push the tags
- add the changelog to the release page on github
You must be part of the marklogic-community organization on bintray in order to publish.
./gradlew mlUndeploy -Pconfirm=true