fix: sort labels by str() key to handle mixed int/str label types

[YousefZahran1]

What

ClassificationQualityMetric and ClassificationConfusionMatrix crash with TypeError: '<' not supported between instances of 'str' and 'int' when classification labels include strings that look like integers (e.g. '101', '102').

Why

Evidently's internal label_key_valudator coerces any label value that can be cast to int — including string labels like '101' — into an actual int. When other labels remain as str (e.g. 'foo', 'bar'), the labels list contains mixed types. Python 3's sorted() cannot compare int and str, so calculate_matrix crashes on line 324.

Reproduce

from evidently.report import Report
from evidently.metrics import ClassificationQualityMetric, ClassificationConfusionMatrix
import pandas as pd

label_target  = ['foo', 'bar', 'fun', 'foo', 'fun', 'foo', '101', '102']
label_predict = ['foo', 'bar', 'fun', 'bar', 'fun', 'fun', '101', '101']
data_df = pd.DataFrame({'target': label_target, 'prediction': label_predict}, dtype='string')

report = Report(metrics=[ClassificationQualityMetric(), ClassificationConfusionMatrix()])
report.run(reference_data=None, current_data=data_df)
# TypeError: '<' not supported between instances of 'str' and 'int'

Fix

Use sorted(labels, key=str) instead of sorted(labels) in calculate_matrix. This sorts labels by their string representation, avoiding the type comparison entirely while preserving a consistent, deterministic order. No API or behaviour change for datasets with homogeneous label types.

Testing

# After fix — report runs without error:
report.run(reference_data=None, current_data=data_df)
print("SUCCESS")  # ✓

Fixes #1085