Fix duplicate notification endpoint setup in pipeline generation

**Why?**

When generating a pipeline, the code optimization introduced by PR awslabs#285
broke the notification_endpoint configuration. When set, it would try to
configure the notification in the main generator as well as in the
default ADF pipeline generator.

**What?**

Removing the notification endpoint configuration from the main template,
as that is replaced by the default ADF pipeline generator
implementation.

Author: sbkok

src/lambda_codebase/initial_commit/bootstrap_repository/adf-build/shared/cdk/cdk_stacks/main.py

@@ -7,7 +7,6 @@
 from aws_cdk import (
     core
 )
-from cdk_constructs import adf_notifications
 from logger import configure_logger
 
 from cdk_stacks.adf_default_pipeline import generate_adf_default_pipeline as generate_default_pipeline, PIPELINE_TYPE as DEFAULT_PIPELINE
@@ -21,9 +20,6 @@ def __init__(self, scope: core.Construct, stack_input: dict, **kwargs) -> None:
         LOGGER.info('Pipeline creation/update of %s commenced', stack_input['input']['name'])
         _pipeline_type = stack_input['input'].get('params', {}).get('type', DEFAULT_PIPELINE).lower()
 
-        if stack_input['input'].get('params', {}).get('notification_endpoint'):
-            stack_input['input']["topic_arn"] = adf_notifications.Notifications(self, 'adf_notifications', stack_input['input']).topic_arn
-
         self.generate_pipeline(_pipeline_type, stack_input)
 
     def generate_pipeline(self, _pipeline_type, stack_input):