Add shared-stream producer and trigger to Apache Kafka provider

Event-driven triggers and asset watchers that watch the same Kafka
topics each open their own consumer in the triggerer, which does not
scale as event-driven schedules grow. Backing them with the
shared-stream producer-side ack channel (#67523) lets siblings on the
same topics and connection share one consumer, while offset commits
remain gated on trigger-event persistence so a triggerer crash cannot
drop a message the broker already considers delivered (at-least-once).

Signed-off-by: PoAn Yang <payang@apache.org>

Author: FrankYang0529

providers/apache/kafka/provider.yaml

@@ -113,6 +113,7 @@ triggers:
     python-modules:
       - airflow.providers.apache.kafka.triggers.await_message
       - airflow.providers.apache.kafka.triggers.msg_queue
+      - airflow.providers.apache.kafka.triggers.shared_stream
 
 connection-types:
   - hook-class-name: airflow.providers.apache.kafka.hooks.base.KafkaBaseHook

providers/apache/kafka/src/airflow/providers/apache/kafka/get_provider_info.py

@@ -83,6 +83,7 @@ def get_provider_info():
                 "python-modules": [
                     "airflow.providers.apache.kafka.triggers.await_message",
                     "airflow.providers.apache.kafka.triggers.msg_queue",
+                    "airflow.providers.apache.kafka.triggers.shared_stream",
                 ],
             }
         ],