yoneda_cache file naming robustness to changing data structures

[kris-brown]

We cache the output of yoneda using yoneda_cache, but this creates a file based on the name of the ACSet. This could lead to name collisions (two different ACSet schemas which happen to share a name) and mystifying errors when one tries to read from the cache after the data of the schema has changed. One possible solution to these issues is to make the file name a hash of the content of the Presentation.