Feat/polynomial model

src/rock_physics_open/equinor_utilities/machine_learning_utilities/__init__.py

@@ -1,14 +1,27 @@
 from .dummy_vars import generate_dummy_vars
-from .exponential_model import CarbonateExponentialPressure
+from .exponential_model import ExponentialPressureModel
+from .friable_pressure_models import (
+    FriableDryBulkModulusPressureModel,
+    FriableDryShearModulusPressureModel,
+)
 from .import_ml_models import import_model
+from .patchy_cement_pressure_models import (
+    PatchyCementDryBulkModulusPressureModel,
+    PatchyCementDryShearModulusPressureModel,
+)
+from .polynomial_model import PolynomialPressureModel
 from .run_regression import run_regression
-from .sigmoidal_model import CarbonateSigmoidalPressure, Sigmoid
+from .sigmoidal_model import SigmoidalPressureModel
 
 __all__ = [
     "generate_dummy_vars",
-    "CarbonateExponentialPressure",
     "import_model",
     "run_regression",
-    "CarbonateSigmoidalPressure",
-    "Sigmoid",
+    "ExponentialPressureModel",
+    "PolynomialPressureModel",
+    "SigmoidalPressureModel",
+    "FriableDryBulkModulusPressureModel",
+    "FriableDryShearModulusPressureModel",
+    "PatchyCementDryShearModulusPressureModel",
+    "PatchyCementDryBulkModulusPressureModel",
 ]

src/rock_physics_open/equinor_utilities/machine_learning_utilities/base_pressure_model.py

@@ -0,0 +1,172 @@
+from __future__ import annotations
+
+import pickle
+from abc import ABC, abstractmethod
+from typing import Any, Self
+
+import numpy as np
+
+
+class BasePressureModel(ABC):
+    """
+    Abstract base class for pressure sensitivity models.
+
+    All pressure models follow the convention:
+    - predict(): returns differential change (depleted - in_situ)
+    - predict_abs(): returns absolute values for specified case
+    - predict_max(): uses model_max_pressure instead of depleted pressure
+
+    Input validation is delegated to concrete implementations since
+    each model has different column requirements.
+    """
+
+    def __init__(self, model_max_pressure: float | None = None, description: str = ""):
+        """
+        Initialize base pressure model.
+
+        Parameters
+        ----------
+        model_max_pressure : float | None
+            Maximum pressure for predict_max method. Required for predict_max to work.
+        description : str
+            Human-readable description of the model instance.
+        """
+        self._model_max_pressure = model_max_pressure
+        self._description = description
+
+    @property
+    def max_pressure(self) -> float | None:
+        """Maximum pressure setting for predict_max method."""
+        return self._model_max_pressure
+
+    @property
+    def description(self) -> str:
+        """Model description."""
+        return self._description
+
+    def predict(self, inp_arr: np.ndarray) -> np.ndarray:
+        """
+        Predict differential change: result(depleted) - result(in_situ).
+
+        Parameters
+        ----------
+        inp_arr : np.ndarray
+            Input array with pressure columns and other model-specific parameters.
+
+        Returns
+        -------
+        np.ndarray
+            Differential change values.
+        """
+        arr = self.validate_input(inp_arr)
+        return self.predict_abs(arr, case="depleted") - self.predict_abs(
+            arr, case="in_situ"
+        )
+
+    def predict_max(self, inp_arr: np.ndarray) -> np.ndarray:
+        """
+        Predict using model_max_pressure instead of depleted pressure.
+
+        Parameters
+        ----------
+        inp_arr : np.ndarray
+            Input array where last column (depleted pressure) will be replaced.
+
+        Returns
+        -------
+        np.ndarray
+            Values at model_max_pressure minus values at in_situ pressure.
+
+        Raises
+        ------
+        ValueError
+            If model_max_pressure is not set.
+        """
+        if self._model_max_pressure is None:
+            raise ValueError('Field "model_max_pressure" is not set')
+
+        arr = self.validate_input(inp_arr).copy()
+        # Replace last column (assumed to be depleted pressure) with max pressure
+        arr[:, -1] = self._model_max_pressure
+        return self.predict_abs(arr, case="depleted") - self.predict_abs(
+            arr, case="in_situ"
+        )
+
+    @abstractmethod
+    def validate_input(self, inp_arr: np.ndarray) -> np.ndarray:
+        """
+        Validate input array format for this specific model.
+
+        Parameters
+        ----------
+        inp_arr : np.ndarray
+            Input array to validate.
+
+        Returns
+        -------
+        np.ndarray
+            Validated input array.
+
+        Raises
+        ------
+        ValueError
+            If input format is invalid for this model.
+        """
+
+    @abstractmethod
+    def predict_abs(self, inp_arr: np.ndarray, case: str = "in_situ") -> np.ndarray:
+        """
+        Predict absolute values for specified pressure case.
+
+        Parameters
+        ----------
+        inp_arr : np.ndarray
+            Validated input array.
+        case : str
+            Either "in_situ" or "depleted" to specify which pressure to use.
+
+        Returns
+        -------
+        np.ndarray
+            Absolute predicted values.
+        """
+
+    @abstractmethod
+    def todict(self) -> dict[str, Any]:
+        """
+        Convert model to dictionary for serialization.
+
+        Returns
+        -------
+        dict[str, Any]
+            Dictionary containing all model parameters.
+        """
+
+    def save(self, file: str | bytes) -> None:
+        """
+        Save model to pickle file.
+
+        Parameters
+        ----------
+        file : str | bytes
+            File path for saving.
+        """
+        with open(file, "wb") as f_out:
+            pickle.dump(self.todict(), f_out)
+
+    @classmethod
+    @abstractmethod
+    def load(cls, file: str | bytes) -> Self:
+        """
+        Load model from pickle file.
+
+        Parameters
+        ----------
+        file : str | bytes
+            File path for loading.
+
+        Returns
+        -------
+        BasePressureModel
+            Loaded model instance.
+        """