lab.find_issues(features=features) outputs error for underperforming issue

[sanjanag]

lab.find_issues(features=features) output

[/Users/sanjana/cleanlab_home/fork_cleanlab/cleanlab/datalab/internal/issue_finder.py:457](https://file+.vscode-resource.vscode-cdn.net/Users/sanjana/cleanlab_home/fork_cleanlab/cleanlab/datalab/internal/issue_finder.py:457): UserWarning: No labels were provided. The 'label' issue type will not be run.
  warnings.warn("No labels were provided. " "The 'label' issue type will not be run.")
Finding null issues ...
Finding outlier issues ...
Fitting OOD estimator based on provided features ...
Finding near_duplicate issues ...
Finding non_iid issues ...
Finding underperforming_group issues ...
Error in underperforming_group: UnderperformingGroupIssueManager.find_issues() missing 1 required positional argument: 'pred_probs'
Failed to check for these issue types: [UnderperformingGroupIssueManager]

Audit complete. 984 issues found in the dataset.

Dataset: https://www.kaggle.com/datasets/laotse/credit-risk-dataset/data

Code

import pandas as pd
from cleanlab import Datalab
from sklearn.preprocessing import StandardScaler
import numpy as np

df = pd.read_csv("./credit_risk_dataset.csv")
df = df[~df.isnull().any(axis=1)].copy()
feature_columns = df.columns.to_list()
feature_columns.remove("loan_status")

X_raw = df[feature_columns]
labels = df["loan_status"]

cat_features = [
    "person_home_ownership",
    "loan_intent",
    "loan_grade",
    "cb_person_default_on_file",
]
numeric_features = [
    "person_age",
    "person_income",
    "person_emp_length",
    "loan_amnt",
    "loan_int_rate",
    "loan_percent_income",
    "cb_person_cred_hist_length",
]

X_encoded = pd.get_dummies(X_raw, columns=cat_features, drop_first=True, dtype='float')

scaler = StandardScaler()
X_processed = X_encoded.copy()
X_processed[numeric_features] = scaler.fit_transform(X_encoded[numeric_features])

lab = Datalab({"X": X_processed.to_numpy(), "y": labels})

lab.find_issues(features=X_processed.to_numpy())

[jwmueller]

@elisno seems like the mapping that decides what issue-types to run based on the supplied args is off. The Underperforming group check should only run if pred_probs were included in the supplied args.