Modified BigQueryDynamicDestination destination to support write API with exactly once semantics

[dedocibula]

Spanner change stream to BQ template fails to launch with error "ParDo requires a deterministic key coder in order to use state and timers" when using useStorageWriteApi=true and useStorageWriteApiAtLeastOnce=false with:

com.google.cloud.teleport.v2.common.UncaughtExceptionLogger - The template launch failed.
java.lang.IllegalArgumentException: ParDo requires a deterministic key coder in order to use state and timers, the reason is:
 ShardedKey$Coder(KvCoder(SerializableCoder(com.google.cloud.bigquery.TableId),TableRowJsonCoder)) is not deterministic because:
	Key coder must be deterministic
	at org.apache.beam.sdk.transforms.ParDo.validateStateApplicableForInput(ParDo.java:448)
	at org.apache.beam.sdk.transforms.ParDo.access$300(ParDo.java:394)
	at org.apache.beam.sdk.transforms.ParDo$MultiOutput.expand(ParDo.java:953)
	at org.apache.beam.sdk.transforms.ParDo$MultiOutput.expand(ParDo.java:872)
	at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:559)
	at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:491)
	at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:352)
	at org.apache.beam.sdk.transforms.ParDo$SingleOutput.expand(ParDo.java:797)
	at org.apache.beam.sdk.transforms.ParDo$SingleOutput.expand(ParDo.java:712)
	at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:559)
	at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:491)
	at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:352)
	at org.apache.beam.sdk.transforms.GroupIntoBatches.expand(GroupIntoBatches.java:338)
	at org.apache.beam.sdk.transforms.GroupIntoBatches.expand(GroupIntoBatches.java:103)
	at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:559)
	at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:508)
	at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:365)
	at org.apache.beam.sdk.io.gcp.bigquery.StorageApiLoads.expandTriggered(StorageApiLoads.java:231)
	at org.apache.beam.sdk.io.gcp.bigquery.StorageApiLoads.expand(StorageApiLoads.java:125)
	at org.apache.beam.sdk.io.gcp.bigquery.StorageApiLoads.expand(StorageApiLoads.java:46)
	at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:559)
	at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:508)
	at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:365)
	at org.apache.beam.sdk.io.gcp.bigquery.BigQueryIO$Write.continueExpandTyped(BigQueryIO.java:3718)
	at org.apache.beam.sdk.io.gcp.bigquery.BigQueryIO$Write.expandTyped(BigQueryIO.java:3502)
	at org.apache.beam.sdk.io.gcp.bigquery.BigQueryIO$Write.expand(BigQueryIO.java:3374)
	at org.apache.beam.sdk.io.gcp.bigquery.BigQueryIO$Write.expand(BigQueryIO.java:2214)
	at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:559)
	at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:508)
	at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:365)
	at com.google.cloud.teleport.v2.templates.spannerchangestreamstobigquery.SpannerChangeStreamsToBigQuery.run(SpannerChangeStreamsToBigQuery.java:382)
	at com.google.cloud.teleport.v2.templates.spannerchangestreamstobigquery.SpannerChangeStreamsToBigQuery.main(SpannerChangeStreamsToBigQuery.java:198)

Given that TableRowJsonCoder's verifyDeterministic method always throws we need to implement a custom deterministic coder for the value

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 22 lines in your changes missing coverage. Please review.

Project coverage is 43.98%. Comparing base (4f45d13) to head (10d11ac).
Report is 6 commits behind head on main.

Files with missing lines	Patch %	Lines
...streamstobigquery/BigQueryDynamicDestinations.java	0.00%	22 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #1885      +/-   ##
============================================
+ Coverage     43.95%   43.98%   +0.02%     
- Complexity     3448     3457       +9     
============================================
  Files           827      827              
  Lines         49001    49053      +52     
  Branches       5136     5141       +5     
============================================
+ Hits          21540    21574      +34     
- Misses        25810    25831      +21     
+ Partials       1651     1648       -3     

Components	Coverage Δ
spanner-templates	63.58% <ø> (+0.11%)	⬆️
spanner-import-export	63.82% <ø> (+0.03%)	⬆️
spanner-live-forward-migration	75.24% <ø> (+0.19%)	⬆️
spanner-live-reverse-replication	68.40% <ø> (+0.18%)	⬆️
spanner-bulk-migration	84.29% <ø> (+0.17%)	⬆️

Files with missing lines	Coverage Δ
...streamstobigquery/BigQueryDynamicDestinations.java	0.00% <0.00%> (ø)

... and 12 files with indirect coverage changes

[damondouglas]

Approved and merging:

• Java PR / JUnit Test Report (pull_request) failed checks have nothing to do with this PR
• Dataflow Integration Test failures show the same failures in the JUnit Test Report