Add constraints handling for contour plot

[toshihikoyanase]

Contributor License Agreement

This repository (optuna-dashboard) and Goptuna share common code.
This pull request may therefore be ported to Goptuna.
Make sure that you understand the consequences concerning licenses and check the box below if you accept the term before creating this pull request.

• I agree this patch may be ported to Goptuna by other Goptuna contributors.

Reference Issues/PRs

Distinguish between feasible and infeasible trials in contour plots, similar to the history plot updated in #529.
The visualization function in Optuna has already supported this feature since optuna/optuna#4975 was merged.

What does this implement/fix? Explain your changes.

• Infeasible trials are plotted with #cccccc dots.
• Feasible trials are plotted with #000000 dots, consistent with previous versions.
• Each dot has a Gray edge, independent of the dot type.

[toshihikoyanase]

Let me share the code example to create the plot above:

from __future__ import annotations

import optuna
from optuna import Trial
from optuna.samplers import TPESampler
from optuna.trial import FrozenTrial


def objective(trial: Trial) -> float:
    x = trial.suggest_float("x", -1, 1)
    y = trial.suggest_float("y", -1, 1)

    cx = x - 0.5  # x should be less than 0.5.
    cy = y - 0.5  # y should be less than 0.5.

    trial.set_user_attr("constraints", [cx, cy])
    # x = 0.5, y = -0.5
    return (x - 0.5) ** 2 + (y + 0.5) ** 2


def constraints_func(trial: FrozenTrial) -> list[float]:
    return trial.user_attrs["constraints"]


sampler = TPESampler(constraints_func=constraints_func)
study = optuna.create_study(sampler=sampler, storage="sqlite:///db-contour.sqlite3")
study.optimize(objective, n_trials=20)

optuna.visualization.plot_contour(study).show()

[keisuke-umezawa]

LGTM!