Support creating a GlueCatalog with own client.

[rchowell]

Feature Request / Improvement

Sometimes I have my own boto3 client which I would like to use as the GlueCatalog client. Perhaps an optional client constructor parameter or an alternative constructor can address this.

class GlueCatalog(MetastoreCatalog):
 def __init__(self, name: str, client: GlueClient | None = None, **properties: Any):
 super().__init__(name, **properties)

 retry_mode_prop_value = get_first_property_value(properties, GLUE_RETRY_MODE)

 if client:
 self.glue = client
 else:
 session = boto3.Session(
 profile_name=properties.get(GLUE_PROFILE_NAME),
 region_name=get_first_property_value(properties, GLUE_REGION, AWS_REGION),
 botocore_session=properties.get(BOTOCORE_SESSION),
 aws_access_key_id=get_first_property_value(properties, GLUE_ACCESS_KEY_ID, AWS_ACCESS_KEY_ID),
 aws_secret_access_key=get_first_property_value(properties, GLUE_SECRET_ACCESS_KEY, AWS_SECRET_ACCESS_KEY),
 aws_session_token=get_first_property_value(properties, GLUE_SESSION_TOKEN, AWS_SESSION_TOKEN),
 )
 self.glue: GlueClient = session.client(
 "glue",
 endpoint_url=properties.get(GLUE_CATALOG_ENDPOINT),
 config=Config(
 retries={
 "max_attempts": properties.get(GLUE_MAX_RETRIES, MAX_RETRIES),
 "mode": retry_mode_prop_value if retry_mode_prop_value in EXISTING_RETRY_MODES else STANDARD_RETRY_MODE,
 }
 ),
 )

Currently I am using __new__ to set this myself, but perhaps poor practice, and it's not clear at the moment which missing properties may impact me later.

gc = GlueCatalog.__new__(GlueCatalog)
gc.name = my_name
gc.properties = {}
gc.glue = my_client # <-----

[Fokko]

@rchowell Thanks for raising this, I think it is reasonable to pass in your own client. Are you interested in raising a PR? Happy to review!

[rchowell]

I'll give it a go 👍