[BUG] value error in matching

[tikhomirovd]

🐛 Bug Description

A ValueError occurs when executing the Matching model with a Dataset object in HypEx. The error message indicates a length mismatch in the expected vs. new index values.

Steps To Reproduce

1. Create a Dataset object with specific roles.
2. Pass the dataset to the Matching model.
3. Execute mtchng_mdl.execute(data=hypex_df).
4. Observe the ValueError.

Expected Behavior

The Matching model should execute without errors, correctly handling dataset indexing.

Environment

• HypEx Version: [e.g. 1.0.0]
• Python Version: [e.g. 3.8]
• Operating System: [e.g. iOS, Windows, Linux]

Additional Context

Full traceback:

ValueError: Length mismatch: Expected axis has 712395 elements, new values have 373945 elements

Possible Solution

Potential issues:

• Mismatch in indexing when concatenating treatment and control groups.
• Ensure index values align correctly between merged datasets.
• Investigate dataset manipulation within Bias.execute().

Code Sample

import hypex
from hypex.dataset import Dataset, InfoRole, TreatmentRole, TargetRole, FeatureRole

hypex_df = Dataset(
    roles={
        "epk_id":         InfoRole(str),
        "treatment_flag": TreatmentRole(int),
        "fake_target":    TargetRole(int)
    },
    data=hypex_mtchng_df,
    default_role=FeatureRole()
)

mtchng_mdl = hypex.Matching(
    group_match=False,
    distance="mahalanobis",
    bias_estimation=True,
    quality_tests="auto"
)
results = mtchng_mdl.execute(data=hypex_df)

Checklist

• I have described the bug in detail
• I have provided steps to reproduce
• I have provided the expected behavior
• I have provided screenshots (if applicable)
• I have provided my environment details
• I have suggested a possible solution (if applicable)