added plot_gp_slice.py

[Jimbo994]

Added a plot function to plot the "slice" of a GP prediction given some fixed input features, two input features that are varied between the bounds of the domain and the output feature of interest.

To run a quick test:

import pandas as pd
import numpy as np
from bofire.data_models.features.api import ContinuousInput, ContinuousOutput
from bofire.data_models.constraints.api import LinearInequalityConstraint, NChooseKConstraint
from bofire.data_models.domain.api import Domain, Inputs, Outputs, Constraints
from bofire.data_models.surrogates.api import SingleTaskGPSurrogate
from bofire.runners.api import hyperoptimize

from bofire.plot.plot_gp_slice import plot_gp_slice

import plotly.graph_objects as go

import bofire.surrogates.api as surrogates

# Define the domain
inputs = Inputs(
    features=[
        ContinuousInput(key="x1", bounds=(0, 1)),
        ContinuousInput(key="x2", bounds=(0, 1)),
        ContinuousInput(key="x3", bounds=(0, 1)),
    ]
)

constraints = Constraints(
    constraints=[
        LinearInequalityConstraint(features=["x1", "x2", "x3"], coefficients=[1, 1, 1], rhs=2),
        NChooseKConstraint(features=["x1", "x2"], min_count=1, max_count=2, none_also_valid=False),
    ]
)

outputs = Outputs(
    features=[
        ContinuousOutput(key="y")
    ]
)

domain = Domain(inputs=inputs, constraints=constraints, outputs=outputs)

# Generate some synthetic data
data = pd.DataFrame({
    "x1": np.random.rand(100),
    "x2": np.random.rand(100),
    "x3": np.random.rand(100),
})
data["y"] = data["x1"] + data["x2"] + np.random.normal(0, 0.1, 100)

# add a valid_y column
data["valid_y"] = 1

# Define the surrogate model
surrogate_data = SingleTaskGPSurrogate(inputs=domain.inputs.get_by_keys(domain.inputs.get_keys()), outputs=domain.outputs)

# Optimize the surrogate model
opt_surrogate_data, perf = hyperoptimize(surrogate_data=surrogate_data, training_data=data, folds=2)

# Fit the surrogate model
surrogate_gp = surrogates.map(opt_surrogate_data)

surrogate_gp.fit(data)

# Define fixed input features
fixed_input_features = pd.Series({"x3": 0.5})

# Plot the GP slice
fig, fig_sd = plot_gp_slice(
    domain=domain,
    model=surrogate_gp,
    fixed_input_features=fixed_input_features,
    input_features=["x1", "x2"],
    output_feature="y",
    resolution=100,
    observed_data=data
)

fig.show()
fig_sd.show()