Use a better method to incrementally save metadata for vitality

[bfabio]

We are using bare clones to calculate the metadata from, without getting the same information from network every time
(#269).

Despite that, the cloned repos for the Italian catalog contains 40GB of data.

We should use a smarter way to persist the data we need to disk incrementally instead of cloning the repo, while maintaining the ability of getting just the deltas from the network.