fix: skip non-numeric columns in calc_r2

[drussellmrichie]

Bug

calc_r2 in openavmkit/utilities/stats.py calls data[var].astype(float) on every variable in ind_vars without first checking whether the column is numeric. Any string/categorical column triggers:

ValueError: could not convert string to float: '100A'

This crash occurs in practice when ind_vars includes columns like luc (land use code) or bldg_com_struct (commercial structure type), which are legitimate categorical features for LightGBM models but cannot be cast to float for OLS R² computation.

Fix

Add a pd.api.types.is_numeric_dtype guard immediately after the existing ill-posed-model check (the len(data) < 3 or nunique() < 2 block). Non-numeric variables now receive NaN R² / adj-R² / coef_sign and are skipped cleanly via continue — consistent with the function's existing error-handling pattern for ill-posed models.

Before / After

Before:

        if len(data) < 3 or data[var].nunique() < 2:
            ...
            continue  # skip ill-posed models

        X = sm.add_constant(data[var].astype(float), has_constant='add')
        # ^ raises ValueError for string columns

After:

        if len(data) < 3 or data[var].nunique() < 2:
            ...
            continue  # skip ill-posed models

        # Skip non-numeric columns (e.g. string categoricals); .astype(float) would raise ValueError
        if not pd.api.types.is_numeric_dtype(data[var]):
            results["variable"].append(var)
            results["r2"].append(float("nan"))
            results["adj_r2"].append(float("nan"))
            results["coef_sign"].append(float("nan"))
            continue

        X = sm.add_constant(data[var].astype(float), has_constant='add')

Notes

• pd.api.types.is_numeric_dtype is already available (pandas is a core dependency).
• The NaN result for non-numeric vars is semantically correct — R² is undefined for categorical predictors in a simple OLS framework. Callers that rank variables by R² will naturally sort these to the bottom.
• No behavior change for numeric columns.

[github-actions]

Thank you for your contribution.
Please sign our CLA at the following link:
Click here to sign the CLA.
A maintainer will verify your signature and confirm it here by commenting with the following sentence:

---

I affirm that this contributor has signed the CLA

---

_{You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.}