User guide » Dataset Publishing and Exposures

Dataset Publishing and Exposures

Publishing a dataset makes it discoverable outside DAISY — in an external data catalog, or through a data access request system such as REMS. Publishing is controlled and auditable: it requires a Data Steward, generates a permanent identifier for the dataset, and produces an Exposure record that can be deprecated but never silently removed.

Endpoints

An Endpoint represents an external system that datasets can be published to. Endpoints are configured by a system administrator and are not editable by regular users. Each endpoint has:

• a name identifying the target system (e.g. a data catalog or a REMS instance)
• a URL pattern that resolves to the dataset's public page, using ${entity_id} as a placeholder for the dataset's accession number
• an API key used to authenticate calls from the external system back to DAISY

Endpoints are set up once and then available to Data Stewards when creating exposures.

Exposures

An Exposure is the record that connects a specific dataset to a specific endpoint. Creating an exposure is what makes a dataset "published". Each exposure records:

• which endpoint the dataset is exposed to
• the REMS form assigned for access requests (form ID and name, populated from REMS)
• who created the exposure and when
• whether to attach access requests as PDF to REMS remarks

Warning

Only one active exposure can exist per dataset–endpoint combination. If a dataset needs to be re-exposed to the same endpoint (e.g. after a form change), the existing exposure must first be deprecated.

Who can manage exposures

Only Data Stewards can create, edit, or deprecate exposures.

Publication Status

A dataset's publication status is derived from its exposures:

Status	Condition
Unpublished	No exposures exist for this dataset
Published	At least one active (non-deprecated) exposure exists
Deprecated	Exposures exist but all are deprecated

The status is visible on the dataset detail page and is used by the API to control which datasets are returned to external consumers.

ELU Accession Number

When a dataset is published for the first time — i.e. when its first exposure is created — DAISY automatically generates a permanent ELU accession number in the format ELU_I_XXX. This identifier:

• is used as the ${entity_id} in the endpoint's URL pattern, forming the dataset's public URL in the external catalog
• is used by the DAISY API to identify datasets in responses to external systems
• cannot be changed after it is assigned

Warning

The ELU accession number is permanent. Once assigned it cannot be changed, even if the dataset is later deprecated or re-published.

The accession number is also what links an approved REMS application back to the correct dataset in DAISY.

Deprecating an Exposure

Removing an exposure from a dataset is done by deprecating it, not deleting it. Deprecation requires a reason, which is stored alongside the deprecation timestamp. The exposure record is preserved in full — this is intentional, as it maintains an audit trail of which datasets were exposed, to which systems, and for how long.

Once all exposures for a dataset are deprecated, the dataset's publication status changes to Deprecated. It remains in DAISY but is no longer served to external systems via the API.

Lifecycle Summary

Dataset created
      │
      ▼
[Unpublished] ──── Data Steward creates exposure ────▶ [Published]
                                                             │
                                              Data Steward deprecates exposure
                                                             │
                                                             ▼
                                                       [Deprecated]

Tip

A deprecated dataset can be re-published by creating a new exposure. The previous deprecated exposure remains in the record history.