Content Snapshots
Content snapshots assist with synchronising data with a remote source, typically another archival institution. While mechanisms for communicating removal of previously-existing material does exist in protocols such as OAI-PMH or ResourceSync, these are often unsupported in practice, and for non-harvested data these are obviously no help. Snapshots allow comparing the state of the data in the EHRI portal belonging to an institution before and after a set of import operations and can automatically suggest cleanup actions based on the results.
Snapshot cleanup is dangerous and should only be used with manual verification of the actions to be performed!
A content snapshot is essentially a list of portal identifiers and their corresponding local (indigenous) identifiers for all the material held by a given institution at a point in time:
Portal identifier |
Local identifier |
|---|---|
gb-003348-1 |
1 |
gb-003348-2 |
2 |
... |
... |
gb-003348-500 |
500 |
The idea is to take a snapshot, perform a sequence of import operations that span the full range of material in an institution, and then compare the before and after states to determine what needs to be deleted or where redirects need to be created for items that have moved.
Creating a snapshot
To create a new snapshot simply click the Create Snapshot button and provide a brief description. The description might state that the snapshot is prior to a particular set of imports about to be done.
Selecting a snapshot
Selecting a snapshot from the list allows you to perform cleanup actions.
Note: this is dangerous!
Cleanup consists of:
Creation of redirects for moved items
Deletion of defunct items
Detection of moved items will only be accurate of an institution has fully unique local identifiers, i.e. identifiers that only occur once within the full span of an institutions holdings, and are not re-used for items that exist within different collections. If this is not the case then detection of moved items will likely give inaccurate or misleading results.
Defunct items are assumed to be those which have not been imported since the creation of the snapshot. If any items are manually managed, or created by datasets not yet imported, these will be wrongly suggested for deletion.
Summary: don't use cleanup without manually sanity-checking the actions it suggests.