[FEATURE]: Lint Scala notebooks

[nfx]

Is there an existing issue for this?

• I have searched the existing issues

Problem statement

Some of the ETL code that has to be migrated to UC is written in Scala. We need to speed that migration up. There's https://scalameta.org/docs/trees/guide.html + https://scalacenter.github.io/scalafix/, that is relatively stable.

We can limit the scope to:

• spark.read.XXX("fs://prefixpath/")
• df.write.save("fs://prefixpath/", ...)
• spark.table("...")
• spark.read.table("...")
• df.write.saveAsTable("...")
• spark.sql("...") with a call back to parent Python process for linting/rewrites
• value inference of dbutils.widgets.get(X) from a supplied Map[String,String]
• dbutils.fs. calls

Proposed Solution

Implement linters/fixers for Scala

Additional Context

No response

[ericvergnaud]

Development wise, it might be more effective to leverage ANTLR with Scala grammar https://github.com/antlr/grammars-v4/blob/master/scala/Scala.g4. This would avoid running the scala linter as a (slow) child process and/or recreating all the plumbing.