ML Engines

class mindsdb_sdk.ml_engines.MLEngine(name: str, handler: str, connection_data: dict)

Bases: object

Meta private

connection_data: dict = None

handler: str = None

name: str = None

class mindsdb_sdk.ml_engines.MLEngines(api)

Bases: mindsdb_sdk.utils.objects_collection.CollectionBase

ML engines collection

Examples of usage:

Get list

>>> ml_engines = con.ml_engines.list()

Get

>>> openai_engine = con.ml_engines.openai1

Create

>>> con.ml_engines.create(
...    'openai1',
...    'openai',
...    connection_data={'api_key': '111'}
...)

Drop

>>>  con.ml_engines.drop('openai1')

Upload BYOM model. After uploading a new ml engin will be availbe to create new model from it.

>>> model_code = open('/path/to/model/code').read()
>>> model_requirements = open('/path/to/model/requirements').read()
>>> ml_engine = con.ml_engines.create_byom(
...    'my_byom_engine',
...    code=model_code,
...    requirements=model_requirements
...)

create(name: str, handler: Union[str, mindsdb_sdk.handlers.Handler], connection_data: dict = None) → mindsdb_sdk.ml_engines.MLEngine

Create new ml engine and return it

Parameters

• name – ml engine name, string

• handler – handler name, string or Handler

• connection_data – parameters for ml engine, dict, optional

Returns

created ml engine object

create_byom(name: str, code: str, requirements: Union[str, List[str]] = None)

Create new BYOM ML engine and return it

Parameters

• code – model python code in string

• requirements – requirements for model. Optional if there is no special requirements. It can be content of ‘requirement.txt’ file or list of strings (item for every requirement).

Returns

created BYOM ml engine object

drop(name: str)

Drop ml engine by name

Parameters

name – name

get(name: str) → mindsdb_sdk.ml_engines.MLEngine

Get ml engine by name

:param name :return: ml engine object

list() → List[mindsdb_sdk.ml_engines.MLEngine]

Returns list of ml engines on server

Returns

list of ml engines