Feature request: Caching precomputations

[cascremers]

At the moment, a large model can take a long time to load due to complex precomputations. If a small change is made to the model (even if this does not impact precomputations, e.g. for a lemma) then recomputations are still redone.

It would be desirable to cache these and avoid needless precomputation.

One possible solution involves:

• Computing a fingerprint of the precomputation-relevant data (e.g. a hash value over partial input excluding lemmas etc)
• Storing precomputation data on disk (e.g. temp folder) by fingerprint.
• When precomputation is needed, check if present on disk, reload if present; otherwise precompute and save to disk.

Minor notes: might want to include Maude/Tamarin versions and command-line options in fingerprint.