…dation

- BucketSelectorQueryBuilder: use public bucketsPathsMap getter from
  common-utils instead of reflection (opensearch-project/common-utils#949)
- BucketKeyFilter: use IncludeExclude.convertToStringFilter() instead
  of reflection on private fields
- TriggerService: replace blanket @Suppress(UNCHECKED_CAST) with
  runtime type checks using require()
- BucketLevelMonitorRunner: add 1-trigger validation at execution time
  to cover the _execute API path (in addition to create/update validation)

Signed-off-by: Surya Sashank Nistala <snistala@amazon.com>

* feat(bucket-trigger): Add bucket-level trigger evaluation using standard bucket_selector

Replace the custom BucketSelectorExt aggregation with standard
bucket_selector for bucket-level trigger evaluation when
multi_tenant_trigger_eval_enabled is true. This enables trigger
evaluation on user clusters that do not have the alerting plugin
installed.

**When the flag is on, bucket-level monitors are limited to 1 trigger.**
The standard bucket_selector is injected directly into the monitor's
search query so a single search call performs both data collection and
trigger evaluation. Include/exclude filtering from BucketSelectorExtFilter
is replicated post-response by BucketKeyFilter.

The existing BucketSelectorExt path is completely unchanged when the
flag is off (the default).

New files:
- BucketSelectorQueryBuilder: injects standard bucket_selector sub-agg
- BucketKeyFilter: post-response include/exclude bucket filtering
- TriggerService.runBucketLevelTriggerFromFilteredResponse(): reads
  remaining buckets from filtered response

Modified files:
- BucketLevelMonitorRunner: flag-gated single-query path
- InputService: useStandardBucketSelector parameter
- TransportIndexMonitorAction: 1-trigger validation for bucket-level
  monitors when flag is on
- AggregationQueryRewriter: skipBucketSelectorInjection parameter

Integration tests:
- RemoteBucketLevelTriggerIT: 13 tests covering composite/terms agg,
  alert lifecycle, dry run, nested parent path, include/exclude filter,
  validation of 1-trigger limit, Painless script in query
- RemoteBucketLevelTriggerRegressionIT: 5 tests verifying flag=false
  preserves existing BucketSelectorExt behavior

Signed-off-by: Surya Sashank Nistala <snistala@amazon.com>

* fix: Remove dead code and clean up PR

- Remove collectInputResultsForTrigger() from InputService (dead code
  from per-trigger query approach)
- Remove skipAllBucketSelectorInjection parameter from collectInputResults
  and getSearchRequest (no longer used)
- Collapse identical triggerCtx if/else branches in BucketLevelMonitorRunner
- Restore .gitignore to original state

Signed-off-by: Surya Sashank Nistala <snistala@amazon.com>

* fix: Address PR review comments — remove reflection, add runtime validation

- BucketSelectorQueryBuilder: use public bucketsPathsMap getter from
  common-utils instead of reflection (opensearch-project/common-utils#949)
- BucketKeyFilter: use IncludeExclude.convertToStringFilter() instead
  of reflection on private fields
- TriggerService: replace blanket @Suppress(UNCHECKED_CAST) with
  runtime type checks using require()
- BucketLevelMonitorRunner: add 1-trigger validation at execution time
  to cover the _execute API path (in addition to create/update validation)

Signed-off-by: Surya Sashank Nistala <snistala@amazon.com>

---------

Signed-off-by: Surya Sashank Nistala <snistala@amazon.com>