pin: cleanup pin indexes and storage

Currently, the `pinning/pinner` package is built in such a way that the internals would technically allow for multiple pins of the same CID, while in practice that is not possible. In addition, it makes maintenance tasks or new features, such as #537, more complicated.

The idea here would be to simplify the current indexes to only having one value per key. In addition, when updates are made, avoid deleting and writing a new pin again to a different key in the datastore. Instead, reuse the same key and simply override the the information. Perhaps the ID could even be based on the CID (currently it is random).