Refactor bigtable API to use v2 client

[RustedBones]

Update all bigtable APIs to use cloud v2 client.

Drop BigtableBulkWriter. This is TBD, but it looks to me can manually shard/groupBy and flatten to normal IO if they require this feature.

Change bigtable ScioContext api to use Admin.Table and Admin.Instance objects that cache clients instances.

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 18.94737% with 77 lines in your changes missing coverage. Please review.

Project coverage is 61.59%. Comparing base (25c6987) to head (fc6db20). Report is 11 commits behind head on main.

Files	Patch %	Lines
...c/main/scala/com/spotify/scio/bigtable/Admin.scala	0.00%	55 Missing :warning:
...n/scala/com/spotify/scio/bigtable/BigTableIO.scala	50.00%	14 Missing :warning:
...otify/scio/bigtable/syntax/ScioContextSyntax.scala	20.00%	8 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #5444 +/- ## ========================================== + Coverage 61.29% 61.59% +0.29% ========================================== Files 311 312 +1 Lines 11062 11031 -31 Branches 774 780 +6 ========================================== + Hits 6780 6794 +14 + Misses 4282 4237 -45 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.