Consider falling back to Spark on unsupported data sources

[nevi-me]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.

It appears that when Blaze encounters a data source that's not yet supported, it panics. An example is the Databricks file system (dbfs), which is interpreted as an object store. I haven't yet tested JDBC sources.

Describe the solution you'd like

If Blaze encounters an unsupported object store or other data source, it should defer back to Spark reading the data.
I appreaciate that this might not result in good performance gains on some workloads, but it at least prevents the entire query from failing.

Describe alternatives you've considered

Supporting custom data sources directly, which will take time. I have a DataFusion RDBMS WIP that I'd be able to extend Blaze with, but it's still far away from being generally usable.

Additional context

I was trying out some TPC-H queries with Blaze, so i'm opening issues for things that I observed.

[richox]

maybe we can add some additional check of the file uri in convertFileSourceScanExec?

[richox]

it is supported now, all unsupported operators (including scan) we fallback to spark execution.