Linked Datasets

From TechWiki

Linked datasets are special kind of datasets in conStruct. These conStruct linked datasets are references to datasets that haven't been created using that conStruct node. Normally, if a dataset is created or imported on a conStruct node, conStruct will create the Dataset and generate its own URI (the dataset id). The URI will be based on the conStruct node's URL, and Drupal's internal auto-incrementing node counter.

However, datasets can be created from outside of a conStruct node. Datasets can directly be created on these structWSF instance by querying the Dataset: Create web service endpoint. Since each conStruct node uses one, or multiple, structWSF instances, one can link to these "external" datasets.

Requirements

The only requirement for a conStruct node to be able to link to a dataset is that the server hosting the conStruct instance has the proper rights on the dataset.

Creating a Linked Dataset

Linking a conStruct node to a particular dataset URI is quite easy. One has to follow these steps:

1. The user has to go to the dataset (content type) creation page: http://localhost/conStruct/dataset/
   1. Additionally, there is a "create a new dataset" link from the conStruct dataset module: http://localhost/conStruct/dataset/
2. The user has to provide 4 different things:
   1. The title of the dataset (this is the title used to refer to this dataset, on that conStruct node)
   2. The description of the dataset (this is the description used to refer to this dataset, on that conStruct node)
   3. The IP address / domain name of the structWSF instance where the dataset is hosted (note: if you specify the domain name, remove any protocol identifier such as "http://" and trailing slashes
   4. The URI of the dataset on that structWSF instance.
3. Finally click the "save" button to create that linked dataset.

Once the linked dataset is created, it will appear as is (a linked dataset) in the list of datasets available in the conStruct dataset module. Once that dataset is linked to a conStruct node, it will be available like any other dataset for searching, browsing and such. Multiple datasets, from different sources, can be aggregated and used on the same conStruct node. This is where the power of distributed structWSF nodes starts to appear.

Permissions

There are a few permissions usecases depending on the permissions your conStruct node have on the remote structWSF instance.

One can link to a dataset which is World Readable, or that the linking conStruct instance has access to.

However, if the conStruct node doesn't have full access to the structWSF instance (so, if it doesn't have access to the /wsf/ and /wsf/datasets/ internal datasets of the structWSF instance, it won't be able to create new permissions, on that linked datasets, for its own users.

Depending what is needed, and what operation you want to perform on such a linked dataset, you will have to negotiate the rights to do something additional than accessing (reading) them with the system administrator and managers of that remote structWSF node.