Deleting items from the repository

[marcolarosa]

When an item is deposited into the repository the metadata for that item is indexed (the search index) so that a user can do a search for the content in the repo. In addition, the individual entities (people, org's, etc) are also indexed individually (the lookup index) so that they can be used for lookups when describing the other items.

In the case that an item is deleted from the repo, the metadata can be removed from the search index but the individual entities can't be removed from the lookup index.

Consider the case where I create a marco entity with my orcid in item A and deposit the item, then, that marco entity is available to me for lookup. In item B, i lookup and inject that entity into B's metadata and then deposit that item. If I then remove B from the repo, if i were to delete all the lookup entities the marco person would disappear even though it was made available by depositing item A. Put another way: the entities available for lookup have no connection to the datasource they came from. Which is why we can't clean those up.

The upshot is that an admin who deposits an item with metadata that shouldn't be made available can't actually remove it - yet. We would need to build a control in the UI to delete specific entities to solve this issue. I'm going to describe this in a ticket and we can tackle it if becomes a problem.