Implementation of metadata-based freshness

[mikealfare]

resolves #938

Problem

The current implementation of source freshness requires querying the data and requires the user provide a datetime field. This is slower and more expensive than it should be. It doesn't scale across multiple models. And some models do not have an appropriate datetime field.

Solution

Use source metadata where available.

Checklist

• I have read the contributing guide and understand what's expected of me
• I have run this code in development and it appears to resolve the stated issue
• This PR includes tests, or tests are not required/relevant for this PR
• This PR has no interface changes (e.g. macros, cli, logs, json artifacts, config files, adapter interface, etc) or this PR has already received feedback and approval from Product or DX

[mikealfare]

I added this test because we're seeing odd behavior with the get_table method (a BQ Client method). If we run get_table during source freshness (line 726 in impl.py) on a table that does not exist, it hangs without returning the expected NotFound error, even when providing retry. However when we do the same here, it seems to behave as expected.

[colin-rogers-dbt]

do we understand this well enough to raise a bug with bigquery? https://github.com/googleapis/python-bigquery/issues

[colin-rogers-dbt]

There's this issue: googleapis/python-bigquery#1674

[mikealfare]

Removing backport 1.7.latest as this has actually been merged into 1.7.latest already.

[mikealfare]

This PR reflects the changes that were merged directly into 1.7.latest via #1072.