Bug description

Passing stopping_criterion = None to convergence_plot() or process_benchmark_results() raises
UnboundLocalError: local variable 'is_converged' referenced before assignment, even though the API mentions that None can be passed.

To Reproduce

import plotly.io as pio

import optimagic as om

problems = om.get_benchmark_problems("example")
optimizers = [
    "nag_dfols",
    "scipy_neldermead",
    "scipy_truncated_newton",
]
results = om.run_benchmark(
    problems,
    optimizers,
)

fig = om.convergence_plot(
    problems=problems,
    results=results,
    n_cols=2,
    problem_subset=["rosenbrock_good_start", "box_3d"],
    stopping_criterion=None,
)

fig.show()

Expected behavior

The figure should be generated without any error. I am not sure whether the implementation is incorrect, or whether the docstring is outdated.

Additional Information

The error occurs in the _process_one_result() function in process_benchmark_results.py, due to the absence of logic to handle the case when stopping_criterion = None

Proposed Fix

This could be fixed by one of:

1. The case of stopping_criterion = None should be handled in the _process_one_result(),
2. The documentation should be updated to reflect that stopping_criterion can not be 'None'.