Open DBusAI opened 3 years ago
Hi @DBusAI, thanks for bringing some great ideas. I have been thinking about some of the same things, such as the API simplification. That should be easily doable.
Regarding your point 2, there is a method ModelCreator. construct_from_service_key
. I could overload this to also take a file name, that should do the trick. I have to admit that the construction of the objects is not particular well documented. There is issue #45 to address that.
When I have some more time, I will create individual issues from your five proposal to track them better.
Thanks again!
What do you think about API simplification? 1) Too long import: Current:
from sap.aibus.dar.client.model_manager_client import ModelManagerClient
Proposal: 'from sap.dar import MMClient' 2) Code cutting for model creation: Current:with open('dar_test.txt', 'r') as sk_file: sk_data = sk_file.read() json_data = json.loads(sk_data) creator = ModelCreator.construct_from_credentials( dar_url=json_data['url'], clientid=json_data['uaa']['clientid'], clientsecret=json_data['uaa']['clientsecret'], uaa_url=json_data['uaa']['url'], )
Proposal: creator = ModelCreator(filekey='dar_test.txt') 3) Config proposal: Current: new_schema = { "features": [ {"label": "manufacturer", "type": "CATEGORY"}, {"label": "description", "type": "TEXT"}, {"label": "price", "type": "NUMBER"}, ], "labels": [ {"label": "level1_category", "type": "CATEGORY"}, {"label": "level2_category", "type": "CATEGORY"}, {"label": "level3_category", "type": "CATEGORY"},} Proposal: Features = fc.Category(["manufacturer"])+fc.Text(["description"])+ fc.Number(["price]) Labels = fc.Category(["level1_category,level2_category,level3_category"]) Schema = Features+Labels ^ less code, easy to read, easy to change 4) Enums for templates: Current: model_template_id="d7810207-ca31-4d4d-9b5a-841a644fd81f" Proposal: model_template = 0 5) One class for rule them all: Current: ModelManager, ModelCreator, InferenceClient, etc. Proposal: DARModel DARModel.create() DARModel.deploy() DARModel.inference() DARModel.templates DARModel.deployments etc...