Axes labels don't reflect the overrided values
The parameter `axes_labels` of the pareto_plot function in `fairical.plot` module has an unexpected behavior when considered in processing of `use_axes_labels`. For instance, the code snippet below generates the plot of [issue.pdf](/uploads/33ab896748daa8123ec8c53e21027df5/issue.pdf) ``` import numpy from fairical.scores import Scores thresholds = list(numpy.linspace(0, 1, 5, dtype=float)) data = Scores.load("tests/data/sample/system_1.json") solutions = data.solutions(["eod+gender", "acc"], thresholds=thresholds).deduplicate() from fairical import plot nds, ds = solutions.non_dominated_solutions() solutions_ = {"system_1": (nds, ds)} fig, ax = plot.pareto_plot(solutions_, axes_labels={"eod+gender": "$Custom_{Gender}$", "acc": "Acc"}) fig.show() ``` The solution is to guard to populate `use_axes_labels` when having passed input into `axes_labels` as below: ``` for k in next(iter(solutions.values()))[0].keys(): if k in axes_labels: use_axes_labels.append(axes_labels[k]) else: parts = k.split("+", 2) ```
issue