Logging with Constraints: `ScreenLogger` doesn't print information related to constraints

[till-m]

(Less of a bug and more of an oversight on my part)

Describe the bug
When using ScreenLogger, no information about the constrained optimization is printed.

To Reproduce

import numpy as np
from bayes_opt import BayesianOptimization
from scipy.optimize import NonlinearConstraint

def target_function(x, y):
    return np.cos(2*x)*np.cos(y) + np.sin(x)

def constraint_function(x, y):
    return np.cos(x) * np.cos(y) - np.sin(x) * np.sin(y)

constraint_limit = 0.5
constraint = NonlinearConstraint(constraint_function, -np.inf, constraint_limit)

# Bounded region of parameter space
pbounds = {'x': (0, 6), 'y': (0, 6)}

optimizer = BayesianOptimization(
    f=target_function,
    constraint=constraint,
    pbounds=pbounds,
    random_state=1,
)

optimizer.maximize(
    init_points=2,
    n_iter=10,
)

Expected behavior
I'm somewhat torn on what information should be output. I see a few options:

• Just the value(s) of the constraint function(s) at the specific point
• Just the "allowed"-ness of the point
• Both the specific values and the "allowed"-ness
• Colour disallowed points red and/or allowed points green
• Just the value(s) of the constraint function(s) at the specific point, colour coded green/red for allowed/disallowed

Personally I'm leaning towards the second option to reduce clutter, but I'm open to other opinions. It would look something like this:

|   iter    |  target   |  allowed  |     x     |     y     |
-------------------------------------------------------------
| 1         | 0.4872    | False     | 2.502     | 4.322     |
| 2         | -0.2401   | True      | 0.0006862 | 1.814     |
| 3         | 0.7821    | False     | 2.304     | 4.326     |
| 4         | 1.249     | False     | 1.899     | 4.321     |
| 5         | 0.7351    | False     | 1.351     | 4.982     |
| 6         | 1.814     | True      | 1.534     | 3.757     |
| 7         | 1.793     | False     | 1.854     | 2.982     |
| 8         | 0.02935   | False     | 2.834     | 1.913     |
| 9         | 1.496     | False     | 1.108     | 3.106     |
| 10        | 0.5423    | True      | 5.99      | 0.06449   |
| 11        | 0.4597    | True      | 5.985     | 5.858     |
| 12        | -1.095    | True      | 5.986     | 2.891     |
=============================================================

[bwheelz36]

I agree for the screen logger.
What about the json logger - there we should probably have the ability to read back in the constraint model the same way we can read back in the objective model?

[till-m]

What about the json logger - there we should probably have the ability to read back in the constraint model the same way we can read back in the objective model?

good point, didn't think of that at the time. Should this be in a separate issue?

@bwheelz36 do you have any preference as to how the ScreenLogger should behave?

[bwheelz36]

I think the example you posted above with just True/False is adequate