I have a dumb question - how do I access in the example here: https://docs.kedro.org/en/stable/configuration/parameters.html the step_size parameter in the code?

step_size: 1
model_params:
    learning_rate: 0.01
    test_data_ratio: 0.2
    number_of_train_iterations: 10000

Example shows only the nested one - I have some problems with getting the non-nested parameters :

def train_model(data, model):
    lr = model["learning_rate"]
    test_data_ratio = model["test_data_ratio"]
    iterations = model["number_of_train_iterations"]
# in pipeline definition
node(
    func=train_model,
    inputs=["input_data", "params:model_params"],
    outputs="output_data",
)

Something like

parameters["step_size"]

gives me KeyError, is there an option to have non-nested global parameters? I would like something like this:

def load_train_data(parameters: dict[str, str | int | bool], data_ingestion_parameters: dict[str, str | bool]):
    ...
    step_size = parameters["step_size"]

Currently I have to nest it to work with it like

parameters["global_parameters"]["step_size"]

, but it is a bit ugly

Someone will reply to you shortly. In the meantime, this might help:

Hi @Łukasz Janiec, not a dumb question at all 🙂 If you want to access top-level parameters, you need to make sure you pass the

parameters

argument inside your pipeline, rather than specifying the lower level key

params:model_parameters

so in your case you'll have to change the code to:

def train_model(data, parameters):
    model_params = parameters["model_params"]
    lr = model_params["learning_rate"]
    test_data_ratio = model_params["test_data_ratio"]
    iterations = model_params["number_of_train_iterations"]
   --> step_size = parameters["step_size"]

# in pipeline definition
node(
    func=train_model,
    inputs=["input_data", "parameters"],
    outputs="output_data",
)

Okay, I thought that these global parameters in

parameters.yml

were somehow special, thank you for the answer