A go service broker for Dataverse that implements the Open Service Broker API.
This project is an implementation of osb-starter-pack.
You should use this project if you're interfacing a containerized application in Kubernetes that will utilize data stored on Dataverse.
You'll need:
goprogramming language- A running Kubernetes (or openshift) cluster
- The service-catalog installed in that cluster
If you're using Helm to deploy this project, you'll need to have it installed in the cluster. Make sure RBAC is correctly configured for helm.
You can go get or git clone this repo.
$ go get github.com/dataverse-broker/dataverse-broker/cmd/dataverse-brokerOr clone the repo:
$ cd $GOPATH/src && mkdir -p github.com/dataverse-broker && cd github.com/dataverse-broker && git clone git://github.com/dataverse-broker/dataverse-brokerChange into the project directory:
$ cd $GOPATH/src/github.com/dataverse-broker/dataverse-broker$ make deploy-helm$ make deploy-openshiftRunning either of these flavors of deploy targets will build the dataverse-broker binary,
build the image, deploy the broker into your Kubernetes, and add a
ClusterServiceBroker to the service-catalog.
When logging in, if you are not automatically directed to the service catalog, you can do so manually by using the dropdown menu labelled "Add to Project" and selecting "Browse Catalog." There, you will see dataverse subtree icons among the list of services supported by the catalog.
To begin the process of provisioning and binding a dataverse subtree service, click on a dataverse subtree icon on the service catalog to generate a dialog window. The dialog window contains the following information in the order presented:
Provides a description of the corresponding dataverse subtree, including plans, if more than one. This tab is purely educational, and has no bearing on the actual provisioning/binding phase of the service.
Configure service to be provisioned/binded. Along with prompts to create a new project, you will be prompted to enter your API-token for this subtree (optional). The broker will check that your token has the necessary credentials to access that dataverse. During the Results tab, the provision step will fail if a provided token is invalid.
Allows for the option to bind the service and store the necessary information in a secret, or to create the binding at a later time inside a project.
At this page, the broker will attempt to provision and bind the service. Upon successful provision, the bind step will create a secret with the dataverse coordiates and your credentials. Use this secret with your created project to connect to the Dataverse Service.
For this section we'll asume that you're using the sample-dataverse-app as your application.
On your project page, you'll see your provisioned service. Expand the service by clicking on the arrow to the left of the name to see the secret you've created. Click "View Secret" to see the contents of the secret.
The values of the secret parameters are hidden by default. You can view them by selecting "Reveal Secret".
The secret can be used inside your application. To add this secret to your application, click on the button on the top right labelled "Add to Application". This will open a window as illustrated in the figure below.
You'll have the option to add the secret in the form of environment variables or as a volume. For the sample application, select "Environment variables" as illustrated in the figure above. This will allow the application to have access to the coordinates and credentials in the secret. Click the blue "Save" button on the lower right of the window.
In order for a dataverse to be offered as a service, we need a bit of info regarding the specific dataverse in the form of metadata which is injected into an image (json object located in the whitelist folder residing in the image folder) which dataverse broker eventually calls upon in the event of a service binding. In the dataverse-broker/pkg/broker/utils.go file there are 2 functions of which get the metadata for a dataverse ( DataverseMetadataIds, and DataverseMeta ) which if you run the DataverseMetadataIds it will obtain the metadata for the dataverse. From there you use this output and create a json object similar to that of the current json objects in the whitelist folder, and you inject it with the output from the function. The "service_id" and "plan_id" fields are just UUIDs that can be generated online, unique for each service/plan, which is also injected into the json object for the dataverse.
- Make it easy for clients to interact with Dataverse
- Access datasets for use in containerized applications