I'm back already haha - sorry. The python API for creating a flow run seems like a hassle to me so I may be missing something. Using the client, I can script flow runs via:

client.create_flow_run(flow_id)

But this requires me to store or know the flow_id, which I'd rather not do between separate scripts. Is there a reason this type of submission isn't supported?:

client.create_flow_run(project_name, flow_name)

The CLI supports this so I expected the python API for client to as well. If it's alright, I'd like to open a feature-request on github. I think the fix will be easy. Something like:

def create_flow_run(flow_id=None, project_name=None, flow_name=None):
	
	if not flow_id and not (project_name and flow_name):
		raise ValueError("Either a flow_id or a project_name+flow_name must be provided")
	
	# then add if/else to handle proper graphql mutation

Maybe it'll be better for the client to have a

get_flow_id

method that returns a flow id or

None

given the project name and flow name. This way, you can get the flow id to pass to

create_flow_run

and should there be some other use cases requiring flow id, they'll benefit too.

I could definitely add a get_flow_id method as well -- for the Client class, no other methods need the flow_id so it may be worth just leaving that method to a graphql query until there's more repeated need for it. I still think a single method call is easier (as oppossed to two method calls each time with get_flow_id and then create_flow_run). It looks like the create_flow_run method already takes dynamic inputs with the version_group_id option.