svemnet 0.1.0__py3-none-any.whl

svemnet/__init__.py

@@ -0,0 +1,89 @@
+"""svemnet: Self-Validated Ensemble Models (SVEM) for Python.
+
+Gaussian SVEM regression for small-sample design-of-experiments workflows:
+SVEM lasso / elastic net (:func:`fit_svem`) and SVEM forward selection
+(:func:`fit_svem_forward`), tuned per bootstrap by validation-weighted
+information criteria, plus a deterministic forward-AICc benchmark
+(:func:`forward_select`).
+
+Interfaces
+----------
+* Matrix API: :func:`fit_svem`, :func:`fit_svem_forward`,
+  :func:`forward_select` on numeric arrays.
+* scikit-learn API: :class:`SVEMRegressor` for pipelines and
+  cross-validation.
+* Formula API (optional, ``pip install svemnet[formula]``): :func:`svem` and
+  :func:`forward_aicc` on DataFrames with R-style formulas, mirroring the R
+  package SVEMnet.
+
+References
+----------
+Lemkus, Gotwalt, Ramsey, and Weese (2021), "Self-Validated Ensemble Models
+for Design of Experiments", Chemometrics and Intelligent Laboratory Systems,
+doi:10.1016/j.chemolab.2021.104439. Karl (2024), doi:10.1016/j.chemolab.2024.105122.
+Karl (2026), "SVEMnet: ...", Chemometrics and Intelligent Laboratory Systems,
+doi:10.1016/j.chemolab.2026.105660. R package: https://CRAN.R-project.org/package=SVEMnet
+"""
+
+from __future__ import annotations
+
+from .core import (
+    SVEMGaussianResult,
+    fit_svem,
+    kish_effective_n,
+    make_svem_weights,
+    predict_svem,
+    support_size,
+    weighted_ic_scores,
+)
+from .estimator import SVEMRegressor
+from .expansion import response_surface_formula
+from .forward import (
+    ForwardICResult,
+    SVEMForwardResult,
+    fit_svem_forward,
+    forward_select,
+)
+
+__version__ = "0.1.0"
+
+# The formula-interface names (svem, forward_aicc, SVEMFormulaModel,
+# ForwardAICcModel) are provided lazily via __getattr__ and deliberately
+# excluded from __all__ so `from svemnet import *` works on installs
+# without the [formula] extra.
+__all__ = [
+    "SVEMGaussianResult",
+    "SVEMForwardResult",
+    "ForwardICResult",
+    "SVEMRegressor",
+    "fit_svem",
+    "fit_svem_forward",
+    "forward_select",
+    "predict_svem",
+    "make_svem_weights",
+    "kish_effective_n",
+    "support_size",
+    "weighted_ic_scores",
+    "response_surface_formula",
+    "__version__",
+]
+
+_FORMULA_EXPORTS = {"svem", "forward_aicc", "SVEMFormulaModel", "ForwardAICcModel"}
+
+
+def __getattr__(name: str):
+    if name in _FORMULA_EXPORTS:
+        try:
+            from . import formula as _formula
+        except ImportError as exc:
+            raise AttributeError(
+                f"svemnet.{name} requires the optional formula dependencies "
+                "(formulaic, pandas). Install them with: "
+                "pip install svemnet[formula]"
+            ) from exc
+        return getattr(_formula, name)
+    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
+
+
+def __dir__():
+    return sorted(set(globals()) | _FORMULA_EXPORTS)