Passing Static Parameters¶

Aim: pass parameters to the cycle components, when they are needed.

In [ ]:

# Uncomment the following line when running on Google Colab
# !pip install autora

# Uncomment the following line when running on Google Colab # !pip install autora

In [ ]:

import numpy as np
import pandas as pd
from autora.experimentalist.pipeline import make_pipeline
from autora.variable import VariableCollection, Variable
from sklearn.linear_model import LinearRegression

from autora.workflow import Cycle

import numpy as np import pandas as pd from autora.experimentalist.pipeline import make_pipeline from autora.variable import VariableCollection, Variable from sklearn.linear_model import LinearRegression from autora.workflow import Cycle

Here we have an experimentalist which takes a parameter:

In [ ]:

uniform_random_rng = np.random.default_rng(180)
def uniform_random_sample(n):
    new_conditions = uniform_random_rng.uniform(low=0, high=11, size=n)
    conditions_df = pd.DataFrame({"x": new_conditions})
    return conditions_df
example_experimentalist_with_parameters = make_pipeline([uniform_random_sample])

uniform_random_rng = np.random.default_rng(180) def uniform_random_sample(n): new_conditions = uniform_random_rng.uniform(low=0, high=11, size=n) conditions_df = pd.DataFrame({"x": new_conditions}) return conditions_df example_experimentalist_with_parameters = make_pipeline([uniform_random_sample])

In [ ]:

example_experimentalist_with_parameters(**{"uniform_random_sample": {"n": 1}})

example_experimentalist_with_parameters(**{"uniform_random_sample": {"n": 1}})

Out[ ]:

	x
0	6.33662

In [ ]:

def ground_truth(x):
    return x + 1
metadata = VariableCollection(
   independent_variables=[Variable(name="x", allowed_values=range(11))],
   dependent_variables=[Variable(name="y", value_range=(-20, 20))],
   )
example_experimentalist = make_pipeline([metadata.independent_variables[0].allowed_values])

def get_example_synthetic_experiment_runner():
    rng = np.random.default_rng(seed=180)
    def runner(x):
        return ground_truth(x) + rng.normal(0, 0.1, x.shape)
    def as_dataframe(conditions_df: pd.DataFrame):
        observations_df = conditions_df.copy()
        observations_df["y"] = runner(conditions_df["x"])
        return observations_df
    return as_dataframe

example_synthetic_experiment_runner = get_example_synthetic_experiment_runner()
example_theorist = LinearRegression()

def ground_truth(x): return x + 1 metadata = VariableCollection( independent_variables=[Variable(name="x", allowed_values=range(11))], dependent_variables=[Variable(name="y", value_range=(-20, 20))], ) example_experimentalist = make_pipeline([metadata.independent_variables[0].allowed_values]) def get_example_synthetic_experiment_runner(): rng = np.random.default_rng(seed=180) def runner(x): return ground_truth(x) + rng.normal(0, 0.1, x.shape) def as_dataframe(conditions_df: pd.DataFrame): observations_df = conditions_df.copy() observations_df["y"] = runner(conditions_df["x"]) return observations_df return as_dataframe example_synthetic_experiment_runner = get_example_synthetic_experiment_runner() example_theorist = LinearRegression()

The cycle can handle that using the params keyword:

In [ ]:

cycle_with_parameters = Cycle(
    variables=metadata,
    theorist=example_theorist,
    experimentalist=example_experimentalist_with_parameters,
    experiment_runner=example_synthetic_experiment_runner,
    params={"experimentalist": {"uniform_random_sample": {"n": 3}}}
)
cycle_with_parameters.run()
cycle_with_parameters.data.conditions[-1]

cycle_with_parameters = Cycle( variables=metadata, theorist=example_theorist, experimentalist=example_experimentalist_with_parameters, experiment_runner=example_synthetic_experiment_runner, params={"experimentalist": {"uniform_random_sample": {"n": 3}}} ) cycle_with_parameters.run() cycle_with_parameters.data.conditions[-1]

Out[ ]:

	x
0	7.349166
1	6.085965
2	2.285666

For the next cycle, if we wish, we can change the parameter value:

In [ ]:

cycle_with_parameters.params["experimentalist"]["uniform_random_sample"]["n"] = 2
cycle_with_parameters.run()
cycle_with_parameters.data.conditions[-1]

cycle_with_parameters.params["experimentalist"]["uniform_random_sample"]["n"] = 2 cycle_with_parameters.run() cycle_with_parameters.data.conditions[-1]

Out[ ]:

	x
0	1.955397
1	5.800231