[query] Remove BlockMatrix persist from Backend interface

[ehigham]

This change is split out from a larger refactoring effort on the various Backend
implementations. The goals of this effort are to provide query-level
configuration to the backend that's currently tied to the lifetime of a backend,
reduce code duplication and reduce state duplication.

In this change, I'm removing blockmatrix persist/unpersist from the Backend
interface by adding BlockMatrixCache: mutable.Map[String, BlockMatrix] to
ExecuteContext. The various reader/writer implementations simply fetch the
block matrix from this cache. For the spark backend, this is backed by a cache
whose lifetime is tied to the spark backend. Since block matrices are not
supported in the local and service backends, the cache is an empty map.

Note that block matrix persist is broken in python (#14689)

[ehigham]

• [query] Move is.hail.check to test code #14801 : 2 dependent PRs (#14693 , #14842 )
• [query] Move LoweredTableReaderCoercer into ExecuteContext #14696
• [query] Remove persistedIr from Backend interface #14692
• [query] Remove lookupOrCompileCachedFunction from Backend interface #14691
• [query] Remove BlockMatrix persist from Backend interface #14690 👈 (View in Graphite)
• [query] Expose references via ExecuteContext #14686
• [query] Extract Backend Methods called from Python into Py4JBackendExtensions #14767
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[chrisvittal]

Very straightforward. Looks great.

[patrick-schultz]

As we discussed, I think it's worth considering whether we should just get rid of the SparkBackend specific BlockMatrix persist behavior. But happy to merge this for now to keep our options open!