Fix matching of entities between manifest and catalog

[sweco]

This fixes two issues:

1. Some systems (e.g. Snowflake) uppercase all table and column names. In accordance with the dbt-docs official source code, we lowercase all table and column names before matching them.
2. There can be multiple tables with the same name but in different schemas. We therefore use schema_name.table_name to distinguish tables.

[sweco]

@mrshu, I refactored the code, I believe it's much cleaner now. The following was done:

• Introduce ColumnRef class to hold information about covered / not covered columns. This gets rid of the string parsing that we had.

• Move all coverage computing to the CoverageReport class because it probably belongs there anyway. Thanks to this we only compute coverage once, previously it was accidentally computed twice. In the example below, it was computed on lines 190 and 197-198 because both catalog.coverage and CoverageReport.from_table call table.coverage inside.

  dbt-coverage/dbt_coverage/__init__.py

  Lines 188 to 199 in e0a183f

  	@classmethod
  	def from_catalog(cls, catalog: Catalog, cov_type: CoverageType):
  	cov = catalog.coverage(cov_type)
  	return CoverageReport(
  	cls.Type.CATALOG,
  	cov_type,
  	None,
  	cov[0],
  	cov[1],
  	{table.name: CoverageReport.from_table(table, cov_type)
  	for table in catalog.tables.values()}
  	)

I believe this is a final code. Could I please bother you for one last round of review?

[mrshu]

Nicely done @sweco, let's get it in!