trieste.experimental.plotting.plotting_plotly#

Module Contents#

format_point_markers(num_pts: int, num_init: int, idx_best: Optional[int] = None, mask_fail: Optional[trieste.types.TensorType] = None, m_init: str = 'x', m_add: str = 'circle', c_pass: str = 'green', c_fail: str = 'red', c_best: str = 'darkmagenta') → tuple[trieste.types.TensorType, trieste.types.TensorType][source]#

Prepares point marker styles according to some BO factors

Parameters

  • num_pts – total number of BO points

  • num_init – initial number of BO points

  • idx_best – index of the best BO point

  • mask_fail – Bool vector, True if the corresponding observation violates the constraint(s)

  • m_init – marker for the initial BO points

  • m_add – marker for the other BO points

  • c_pass – color for the regular BO points

  • c_fail – color for the failed BO points

  • c_best – color for the best BO points

Returns

2 string vectors col_pts, mark_pts containing marker styles and colors

add_surface_plotly(xx: trieste.types.TensorType, yy: trieste.types.TensorType, f: trieste.types.TensorType, fig: plotly.graph_objects.Figure, alpha: float = 1.0, figrow: int = 1, figcol: int = 1) → plotly.graph_objects.Figure[source]#

Adds a surface to an existing plotly subfigure

Parameters

  • xx – [n, n] array (input)

  • yy – [n, n] array (input)

  • f – [n, n] array (output)

  • fig – the current plotly figure

  • alpha – transparency

  • figrow – row index of the subfigure

  • figcol – column index of the subfigure

Returns

updated plotly figure

add_bo_points_plotly(x: trieste.types.TensorType, y: trieste.types.TensorType, z: trieste.types.TensorType, fig: plotly.graph_objects.Figure, num_init: int, idx_best: Optional[int] = None, mask_fail: Optional[trieste.types.TensorType] = None, figrow: int = 1, figcol: int = 1) → plotly.graph_objects.Figure[source]#

Adds scatter points to an existing subfigure. Markers and colors are chosen according to BO factors. :param x: [N] x inputs :param y: [N] y inputs :param z: [N] z outputs :param fig: the current plotly figure :param num_init: initial number of BO points :param idx_best: index of the best BO point :param mask_fail: Bool vector, True if the corresponding observation violates the constraint(s) :param figrow: row index of the subfigure :param figcol: column index of the subfigure :return: a plotly figure

plot_model_predictions_plotly(model: trieste.models.interfaces.ProbabilisticModel, mins: trieste.types.TensorType, maxs: trieste.types.TensorType, grid_density: int = 100, num_samples: Optional[int] = None, alpha: float = 0.85) → plotly.graph_objects.Figure[source]#

Plots 2-dimensional plot of model’s predictions. We first create a regular grid of points and evaluate the model on these points. We then plot the mean and 2 standard deviations to show epistemic uncertainty.

For DeepGaussianProcess models num_samples should be used and set to some positive number. This is needed as predictions from deep GP’s are stochastic and we need to take more than one sample to estimate the mean and variance.

Parameters

  • model – A probabilistic model

  • mins – List of 2 lower bounds for creating a grid of points for model predictions.

  • maxs – List of 2 upper bounds for creating a grid of points for model predictions.

  • grid_density – Number of points per dimension. This will result in a grid size of grid_density^2.

  • num_samples – Number of samples to use with deep GPs.

  • alpha – Transparency.

Returns

A plotly figure.

plot_function_plotly(obj_func: Callable[[trieste.types.TensorType], trieste.types.TensorType], mins: trieste.types.TensorType, maxs: trieste.types.TensorType, grid_density: int = 100, title: Optional[str] = None, xlabel: Optional[str] = None, ylabel: Optional[str] = None, alpha: float = 1.0) → plotly.graph_objects.Figure[source]#

Plots 2-dimensional plot of an objective function. To illustrate the function we create a regular grid of points and evaluate the function on these points.

Parameters

  • obj_func – The vectorized objective function.

  • mins – List of 2 lower bounds for creating a grid of points for model predictions.

  • maxs – List of 2 upper bounds for creating a grid of points for model predictions.

  • grid_density – Number of points per dimension. This will result in a grid size of grid_density^2.

  • title – optional titles

  • xlabel – optional xlabel

  • ylabel – optional ylabel

  • alpha – transparency

Returns

A plotly figure.