Add input validation for Friedman-Popescu H statistic calculation

[Copilot]

getFriedmanPopescusH() throws NPE when called with non-numeric or non-existent column names. The error occurs in FriedmanPopescusH.filterFrame() when frame.vecs(cols) returns null for invalid columns.

# This produces a cryptic NPE
h = titanic_xgb.h(train, ['fair', 'age'])  # 'fair' is categorical
# Error: Cannot read field "_key" because "vecs[i]" is null

Changes

• Added default validation method in FriedmanPopescusHCollector interface

  • Validates vars is non-null and non-empty
  • Validates all columns exist and are numeric
  • Validates columns contain non-constant values

• Updated implementations to call validation before processing

  • GBMModel.getFriedmanPopescusH()
  • XGBoostModel.getFriedmanPopescusH()

• Added comprehensive test coverage

  • Java: GBMTest.testFriedmanPopescusHInputValidation()
  • Python: pyunit_h_input_validation.py (tests both GBM and XGBoost)
  • R: runit_GBM_h_statistic_input_validation.R and runit_XGBoost_h_statistic_input_validation.R

Error messages now clearly indicate the issue:

Calculating H statistics error: column 'fair' is not numeric. 
H statistics can only be calculated for numeric variables.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>H statistic: Improve input validation</issue_title>
<issue_description>This code throws NPE:

from h2o.estimators import H2OXGBoostEstimator

# Import the titanic dataset into H2O:
titanic = h2o.import_file("https://s3.amazonaws.com/h2o-public-test-data/smalldata/gbm_test/titanic.csv")

# Set the predictors and response; set the response as a factor:
titanic["survived"] = titanic["survived"].asfactor()
predictors = titanic.columns
response = "survived"

# Split the dataset into a train and valid set:
train, valid = titanic.split_frame(ratios=[.8], seed=1234)

# Build and train the model:
titanic_xgb = H2OXGBoostEstimator(booster='dart',
                                  normalize_type="tree",
                                  seed=1234)
titanic_xgb.train(x=predictors,
                  y=response,
                  training_frame=train,
                  validation_frame=valid)

# Eval performance:
perf = titanic_xgb.model_performance()

# Generate predictions on a test set (if necessary):
pred = titanic_xgb.predict(valid)

# Extract feature interactions:
feature_interactions = titanic_xgb.feature_interaction()

# Get Friedman and Popescu's H statistics
h = titanic_xgb.h(train, ['fair','age'])
print(h)

Server error java.lang.NullPointerException:
  Error: Cannot read field "_key" because "vecs[i]" is null
  Request: None

The input parameter validation should be done properly. </issue_description>

<agent_instructions>Validate input parameters of the Friedman Popescus H statistics. On the Java Backend, find the getFriedmanPopescusH() method, where the implementation is for the XGBoost and GBM algorithms. </agent_instructions>

Comments on the Issue (you are @copilot in this section)

• Fixes H statistic: Improve input validation #16759

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[maurever]

R and Python tests are missing.

[Copilot]

R and Python tests are missing.

Added Python and R tests for input validation in commit 3365711.

Python test: h2o-py/tests/testdir_hstats/pyunit_h_input_validation.py

• Tests null/None, empty, non-existent, and non-numeric column parameters
• Validates both GBM and XGBoost models

R tests:

• h2o-r/tests/testdir_algos/gbm/runit_GBM_h_statistic_input_validation.R (GBM)
• h2o-r/tests/testdir_algos/xgboost/runit_XGBoost_h_statistic_input_validation.R (XGBoost)
• Both test the same validation scenarios

All tests verify proper error messages for invalid input and correct behavior for valid input.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[maurever]

LGTM. @tomasfryda or @valenad1, could you review this PR, please?

[valenad1]

It should go to rel-3.46.0 if you want it in next release and fix check pull request test stage please. :)