Remove mirror features

Mirror features have been added 3-4 years ago and, to my knowledge, it hasn't really been used until now.

When we stop using the Fuseki WAC extension, every resource will have its own graph. So it won't be possible anymore to store mirror data in a special `http://semapps.org/mirror` graph.

SemApps offer a lot of tools to [sync data](https://semapps.org/docs/middleware/sync/). We use these extensively in ActivityPods, for example to aggregate public data in Pods. If at some point a server want to mirror another one, it won't be difficult to do so. But IMO having this as a core SemApps feature was a mistake, because it makes the code more complex (also on the data provider, which executes complex SPARQL queries to find mirror data).

So instead of refactoring the whole mirror feature, I suggest to remove it.

No files or tests are named. Start by searching the repository for mirror features, the `http://semapps.org/mirror` graph, and the data-provider SPARQL queries, then review the linked sync documentation. Done means the mirror-specific behavior and its related complexity are removed without adding a replacement core mirror feature.

Written by the indexing model from the issue text.