radioactiveshrimp 0.1.2__cp39-abi3-manylinux_2_28_x86_64.whl → 0.1.3__cp39-abi3-manylinux_2_28_x86_64.whl

radioactiveshrimp/__init__.py

@@ -1,5 +1,5 @@
 from radioactiveshrimp._core import hello_from_bin
-
+from radioactiveshrimp.model import LinearRegression
 
 def hello() -> str:
     return hello_from_bin()

radioactiveshrimp/model/__init__.py

@@ -0,0 +1,5 @@
+"""
+Model subpackage containing model functions.
+"""
+from .regression import LinearRegression
+__all__=['LinearRegression']

radioactiveshrimp/model/regression.py

@@ -0,0 +1,468 @@
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import numpy as np
+import matplotlib.pyplot as plt
+from scipy import stats
+from typing import Tuple, Optional
+import warnings
+
+class LinearRegression:
+    """
+    A PyTorch-based Linear Regression implementation for one variable.
+    
+    Model: y = w_1 * x + w_0
+    Loss: Mean Squared Error
+    
+    Features:
+    - Gradient-based optimization using PyTorch
+    - Confidence intervals for parameters w_1 and w_0
+    - Visualization with confidence bands
+    """
+    
+    def __init__(self, learning_rate: float = 0.01, max_epochs: int = 1000, 
+                 tolerance: float = 1e-6):
+        """
+        Initialize the Linear Regression model.
+        
+        Args:
+            learning_rate: Learning rate for gradient descent
+            max_epochs: Maximum number of training epochs
+            tolerance: Convergence tolerance for early stopping
+        """
+        self.learning_rate = learning_rate
+        self.max_epochs = max_epochs
+        self.tolerance = tolerance
+        
+        # Model parameters
+        self.w_1 = nn.Parameter(torch.randn(1, requires_grad=True))  # slope
+        self.w_0 = nn.Parameter(torch.randn(1, requires_grad=True))  # intercept
+        
+        # Training data storage
+        self.X_train = None
+        self.y_train = None
+        
+        # Model statistics for confidence intervals
+        self.n_samples = None
+        self.residual_sum_squares = None
+        self.X_mean = None
+        self.X_var = None
+        self.fitted = False
+        
+        # Loss function and optimizer
+        self.criterion = nn.MSELoss()
+        self.optimizer = optim.SGD([self.w_1, self.w_0], lr=self.learning_rate)
+        
+        # Training history
+        self.loss_history = []
+    
+    def forward(self, X: torch.Tensor) -> torch.Tensor:
+        """
+        Forward pass of the linear model.
+        
+        Args:
+            X: Input tensor of shape (n_samples,)
+            
+        Returns:
+            Predictions tensor of shape (n_samples,)
+        """
+        return self.w_1 * X + self.w_0
+    
+    def fit(self, X: np.ndarray, y: np.ndarray, test_x:np.ndarray=None, test_y:np.ndarray=None) -> 'LinearRegression':
+        """
+        Fit the linear regression model to the training data.
+        
+        Args:
+            X: Input features of shape (n_samples,)
+            y: Target values of shape (n_samples,)
+            
+        Returns:
+            self: Returns the fitted model instance
+        """
+        # Convert to PyTorch tensors
+        self.X_train = torch.tensor(X, dtype=torch.float32)
+        self.y_train = torch.tensor(y, dtype=torch.float32)
+        self.n_samples = len(X)
+        
+        # # Store statistics for confidence intervals
+        # self.X_mean = float(np.mean(X))
+        # self.X_var = float(np.var(X, ddof=1))  # Sample variance
+        
+        # Training loop
+        prev_loss = float('inf')
+        
+        for epoch in range(self.max_epochs):
+            # Zero gradients
+            self.optimizer.zero_grad()
+            
+            # Forward pass
+            y_pred = self.forward(self.X_train)
+            
+            # Compute loss
+            loss = self.criterion(y_pred, self.y_train)
+            
+            # Backward pass
+            loss.backward()
+            
+            # Update parameters
+            self.optimizer.step()
+            
+            # Store loss history
+            current_loss = loss.item()
+            self.loss_history.append(current_loss)
+            
+            # Check for convergence
+            if abs(prev_loss - current_loss) < self.tolerance:
+                print(f"Converged after {epoch + 1} epochs")
+                break
+            
+            prev_loss = current_loss
+        
+        # Compute residual sum of squares for confidence intervals
+        with torch.no_grad():
+            y_pred = self.forward(self.X_train)
+            residuals = self.y_train - y_pred
+            self.residual_sum_squares = float(torch.sum(residuals ** 2))
+        
+        self.fitted = True
+
+        # compute r^2 on optional test data, print result
+        if ((test_x is not None) and (test_y is not None)):
+            y_mean = float(torch.mean(self.y_train))
+            ss_tot = float(torch.sum((self.y_train - y_mean) ** 2))
+            r_squared = 1 - (self.residual_sum_squares / ss_tot)
+            print('R^2 for test data: ',r_squared)
+
+        return self
+    
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        """
+        Make predictions on new data.
+        
+        Args:
+            X: Input features of shape (n_samples,)
+            
+        Returns:
+            Predictions as numpy array
+        """
+        if not self.fitted:
+            raise ValueError("Model must be fitted before making predictions")
+        
+        X_tensor = torch.tensor(X, dtype=torch.float32)
+        
+        with torch.no_grad():
+            predictions = self.forward(X_tensor)
+        
+        return predictions.numpy()
+    
+    def get_parameters(self) -> Tuple[float, float]:
+        """
+        Get the fitted parameters.
+        
+        Returns:
+            Tuple of (w_1, w_0) - slope and intercept
+        """
+        if not self.fitted:
+            raise ValueError("Model must be fitted before accessing parameters")
+        
+        return float(self.w_1.item()), float(self.w_0.item())
+    
+    def parameter_confidence_intervals(self, confidence_level: float = 0.95) -> dict:
+        """
+        Compute confidence intervals for parameters w_1 and w_0.
+        
+        Args:
+            confidence_level: Confidence level (e.g., 0.95 for 95%)
+            
+        Returns:
+            Dictionary containing confidence intervals for both parameters
+        """
+        if not self.fitted:
+            raise ValueError("Model must be fitted before computing confidence intervals")
+        
+        # Degrees of freedom
+        df = self.n_samples - 2
+        
+        # Critical t-value
+        alpha = 1 - confidence_level
+        t_critical = stats.t.ppf(1 - alpha/2, df)
+        
+        # Standard error of regression
+        mse = self.residual_sum_squares / df
+        se_regression = np.sqrt(mse)
+        
+        # Standard error for w_1 (slope)
+        se_w1 = se_regression / np.sqrt(self.n_samples * self.X_var)
+        
+        # Standard error for w_0 (intercept)
+        se_w0 = se_regression * np.sqrt(1/self.n_samples + self.X_mean**2 / (self.n_samples * self.X_var))
+        
+        # Get current parameter values
+        w_1_val, w_0_val = self.get_parameters()
+        
+        # Compute confidence intervals
+        w_1_ci = (
+            w_1_val - t_critical * se_w1,
+            w_1_val + t_critical * se_w1
+        )
+        
+        w_0_ci = (
+            w_0_val - t_critical * se_w0,
+            w_0_val + t_critical * se_w0
+        )
+        
+        return {
+            'w_1_confidence_interval': w_1_ci,
+            'w_0_confidence_interval': w_0_ci,
+            'confidence_level': confidence_level,
+            'standard_errors': {
+                'se_w1': se_w1,
+                'se_w0': se_w0,
+                'se_regression': se_regression
+            }
+        }
+    
+    def plot_regression_with_confidence_band(self, confidence_level: float = 0.95, 
+                                           figsize: Tuple[int, int] = (10, 6),
+                                           title: Optional[str] = None) -> plt.Figure:
+        """
+        Plot the fitted regression line with confidence band.
+        
+        Args:
+            confidence_level: Confidence level for the band
+            figsize: Figure size tuple
+            title: Optional plot title
+            
+        Returns:
+            matplotlib Figure object
+        """
+        if not self.fitted:
+            raise ValueError("Model must be fitted before plotting")
+        
+        # Create figure
+        fig, ax = plt.subplots(figsize=figsize)
+        
+        # Convert training data to numpy for plotting
+        X_np = self.X_train.numpy()
+        y_np = self.y_train.numpy()
+        
+        # Create prediction range
+        X_range = np.linspace(X_np.min(), X_np.max(), 100)
+        y_pred_range = self.predict(X_range)
+        
+        # Compute confidence band
+        df = self.n_samples - 2
+        alpha = 1 - confidence_level
+        t_critical = stats.t.ppf(1 - alpha/2, df)
+        
+        mse = self.residual_sum_squares / df
+        se_regression = np.sqrt(mse)
+        
+        # Standard error for predictions (confidence band)
+        X_centered = X_range - self.X_mean
+        se_pred = se_regression * np.sqrt(1/self.n_samples + X_centered**2 / (self.n_samples * self.X_var))
+        
+        # Confidence band bounds
+        margin_of_error = t_critical * se_pred
+        y_upper = y_pred_range + margin_of_error
+        y_lower = y_pred_range - margin_of_error
+        
+        # Plot data points
+        ax.scatter(X_np, y_np, alpha=0.6, color='blue', label='Data points')
+        
+        # Plot regression line
+        ax.plot(X_range, y_pred_range, 'r-', linewidth=2, label='Fitted line')
+        
+        # Plot confidence band
+        ax.fill_between(X_range, y_lower, y_upper, alpha=0.3, color='red', 
+                       label=f'{int(confidence_level*100)}% Confidence band')
+        
+        # Get parameter values for display
+        w_1_val, w_0_val = self.get_parameters()
+        
+        # Labels and title
+        ax.set_xlabel('X')
+        ax.set_ylabel('y')
+        if title is None:
+            title = f'Linear Regression: y = {w_1_val:.3f}x + {w_0_val:.3f}'
+        ax.set_title(title)
+        ax.legend()
+        ax.grid(True, alpha=0.3)
+        
+        plt.tight_layout()
+        return fig
+    
+    def summary(self) -> dict:
+        """
+        Provide a summary of the fitted model.
+        
+        Returns:
+            Dictionary containing model summary statistics
+        """
+        if not self.fitted:
+            raise ValueError("Model must be fitted before generating summary")
+        
+        w_1_val, w_0_val = self.get_parameters()
+        
+        # R-squared calculation
+        y_mean = float(torch.mean(self.y_train))
+        ss_tot = float(torch.sum((self.y_train - y_mean) ** 2))
+        r_squared = 1 - (self.residual_sum_squares / ss_tot)
+        
+        # Adjusted R-squared
+        adj_r_squared = 1 - ((1 - r_squared) * (self.n_samples - 1) / (self.n_samples - 2))
+        
+        # RMSE
+        rmse = np.sqrt(self.residual_sum_squares / self.n_samples)
+        
+        return {
+            'parameters': {
+                'w_1 (slope)': w_1_val,
+                'w_0 (intercept)': w_0_val
+            },
+            'model_fit': {
+                'r_squared': r_squared,
+                'adjusted_r_squared': adj_r_squared,
+                'rmse': rmse,
+                'residual_sum_squares': self.residual_sum_squares
+            },
+            'training_info': {
+                'n_samples': self.n_samples,
+                'epochs_trained': len(self.loss_history),
+                'final_loss': self.loss_history[-1] if self.loss_history else None
+            }
+        }
+
+    def analysis_plot(self, figsize: Tuple[int, int] = (10, 6),
+                            title: Optional[str] = None) -> plt.Figure:
+        """
+        Plot the fitted regression line.
+        
+        Args:
+            figsize: Figure size tuple
+            title: Optional plot title
+            
+        Returns:
+            matplotlib Figure object
+        """
+        if not self.fitted:
+            raise ValueError("Model must be fitted before plotting")
+        
+        # Create figure
+        fig, ax = plt.subplots(3,1,figsize=figsize)
+        
+        # Convert training data to numpy for plotting
+        X_np = self.X_train.numpy()
+        y_np = self.y_train.numpy()
+        
+        # Create prediction range
+        X_range = np.linspace(X_np.min(), X_np.max(), 100)
+        y_pred_range = self.predict(X_range)
+        
+        # Plot data points
+        ax[0].scatter(X_np, y_np, alpha=0.6, color='blue', label='Data points')
+        
+        # Plot regression line with original data
+        ax[1].plot(X_range, y_pred_range, 'r-', linewidth=2, label='Fitted line')
+        ax[1].scatter(X_np, y_np, alpha=0.6, color='blue', label='Data points')
+
+        
+        # Get parameter values for display
+        w_1_val, w_0_val = self.get_parameters()
+
+        # Plot loss as training occured
+        ax[2].plot(self.loss_history, label='Training Loss', color='green')
+        
+        # Labels and title
+        ax[0].set_xlabel('X')
+        ax[0].set_ylabel('y')
+        ax[0].set_title("Original Data")
+        ax[0].legend()
+        ax[0].grid(True, alpha=0.3)
+
+        ax[1].set_xlabel('X')
+        ax[1].set_ylabel('y')
+        if title is None:
+            title = f'Linear Regression: y = {w_1_val:.3f}x + {w_0_val:.3f}'
+        ax[1].set_title(title)
+        ax[1].legend()
+        ax[1].grid(True, alpha=0.3)
+
+        ax[2].set_xlabel('Epoch')
+        ax[2].set_ylabel('Loss')
+        ax[2].set_title('Loss Though Training')
+        ax[2].legend()
+        ax[2].grid(True, alpha=0.3)
+        
+        plt.tight_layout()
+        plt.show()
+        return fig
+
+
+#-------------------------------------------------------------------------------------
+
+# import polars as p
+# from sklearn.model_selection import train_test_split
+
+# L = LinearRegression()
+# data = p.read_csv('/home/radioactiveshrimp/radioactiveshrimp/Hydropower.csv')
+# x = data.get_column('BCR')
+# y = data.get_column('AnnualProduction')
+
+# X_train, X_test, y_train, y_test = train_test_split(x, y, test_size = 0.25)
+
+# y_pred = L.forward(.9)
+
+
+# fittedMod = L.fit(X_train, y_train, X_test, y_test)
+# # fittedMod = L.fit(X_train, y_train)
+# print(L.predict(.9))
+# # print(L.loss_history)
+
+# fig = L.analysis_plot()
+# plt.show()
+
+#----------------------------------------------------------------
+# np.random.seed(42)
+# torch.manual_seed(42)
+
+# n_samples = 100
+# X = np.random.uniform(-5, 5, n_samples)
+# true_slope = 2.5
+# true_intercept = -1.2
+# noise = np.random.normal(0, 1, n_samples)
+# y = true_slope * X + true_intercept + noise
+
+# print("=== Linear Regression with PyTorch ===")
+# print(f"True parameters: slope={true_slope}, intercept={true_intercept}")
+# print()
+
+# # Create and fit model
+# model = LinearRegression(learning_rate=0.01, max_epochs=10)
+# model.fit(X, y)
+
+# # Get fitted parameters
+# fitted_slope, fitted_intercept = model.get_parameters()
+# print(f"Fitted parameters: slope={fitted_slope:.4f}, intercept={fitted_intercept:.4f}")
+
+# # Model summary
+# summary = model.summary()
+# print(f"\nModel Summary:")
+# print(f"R-squared: {summary['model_fit']['r_squared']:.4f}")
+# print(f"RMSE: {summary['model_fit']['rmse']:.4f}")
+# print(f"Training epochs: {summary['training_info']['epochs_trained']}")
+
+# # Confidence intervals
+# ci_results = model.parameter_confidence_intervals(confidence_level=0.95)
+# print(f"\n95% Confidence Intervals:")
+# print(f"w_1 (slope): [{ci_results['w_1_confidence_interval'][0]:.4f}, {ci_results['w_1_confidence_interval'][1]:.4f}]")
+# print(f"w_0 (intercept): [{ci_results['w_0_confidence_interval'][0]:.4f}, {ci_results['w_0_confidence_interval'][1]:.4f}]")
+
+# # Plot results
+# fig = model.plot_regression_with_confidence_band(confidence_level=0.95)
+# plt.show()
+
+# # Test different confidence levels
+# for conf_level in [0.90, 0.95, 0.99]:
+#     ci = model.parameter_confidence_intervals(confidence_level=conf_level)
+#     print(f"\n{int(conf_level*100)}% CI for slope: [{ci['w_1_confidence_interval'][0]:.4f}, {ci['w_1_confidence_interval'][1]:.4f}]")

{radioactiveshrimp-0.1.2.dist-info → radioactiveshrimp-0.1.3.dist-info}/METADATA

@@ -1,8 +1,11 @@
 Metadata-Version: 2.4
 Name: radioactiveshrimp
-Version: 0.1.2
+Version: 0.1.3
 Requires-Dist: matplotlib>=3.10.6
 Requires-Dist: numpy>=2.3.3
+Requires-Dist: polars>=1.33.1
+Requires-Dist: scikit-learn>=1.7.2
+Requires-Dist: scipy>=1.16.2
 Requires-Dist: torch>=2.8.0
 Requires-Dist: twine>=6.1.0
 Summary: Add your description here

radioactiveshrimp-0.1.3.dist-info/RECORD

@@ -0,0 +1,15 @@
+radioactiveshrimp-0.1.3.dist-info/METADATA,sha256=cl4YyqMwZB_g-18Sqq8wDLp6YKuNdGhHgjjq_lUTnYc,506
+radioactiveshrimp-0.1.3.dist-info/WHEEL,sha256=whVUlzvauXZI_AddfJsPvOGSCqk5SsvrB6oOXoKFJ_s,97
+radioactiveshrimp/__init__.py,sha256=Np_SVIHI6cGztPNmaAjmTydRaDgmbgW5SV8qmGEFMBY,153
+radioactiveshrimp/_core.abi3.so,sha256=3IMwsP7UrgNWdKdw2Pl11_VUQF8NqPy1xyyPuzkREdA,489672
+radioactiveshrimp/_core.pyi,sha256=b6oJaUXUzEzqUE5rpqefV06hl8o_JCU8pgKgIIzQgmc,33
+radioactiveshrimp/differential/__init__.py,sha256=jOhQKr6xM_7TVUt0GWRddtkgBvNn5I873WZcArT-GtE,101
+radioactiveshrimp/differential/discrete.py,sha256=9-w6Avz4MZfu8gFhJU8kTvryn9oT2xS1TvvYQ_8pmPU,1371
+radioactiveshrimp/distributions/__init__.py,sha256=hZX2UFt54wJG4zDLuTTTonDSA-anWvfjhDmR-akYcn4,182
+radioactiveshrimp/distributions/cvdistributions.py,sha256=UEiO_RYFQYqHybI_wcOdpOrxHToXRpLFsB58dwvI4vU,2599
+radioactiveshrimp/matrix/__init__.py,sha256=EoerrXlfmd2lI9grb3jxqIhHn-bzLtYhRKyxI0RVgaE,166
+radioactiveshrimp/matrix/elementary.py,sha256=zPiBG24F5zIncrrVWbiXEp1oFnHw2xNKy5quF0GBOuc,6333
+radioactiveshrimp/model/__init__.py,sha256=OLj9ObpV0i7DFJOb8U12qnJKVAgFb8_hB9yBFV57IZQ,122
+radioactiveshrimp/model/regression.py,sha256=i2CF6OVOb7eufzckAltRHJaoJlpproxLHqOPYSjpSJI,15685
+radioactiveshrimp/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+radioactiveshrimp-0.1.3.dist-info/RECORD,,

{radioactiveshrimp-0.1.2.dist-info → radioactiveshrimp-0.1.3.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: maturin (1.9.4)
+Generator: maturin (1.9.6)
 Root-Is-Purelib: false
 Tag: cp39-abi3-linux_x86_64

radioactiveshrimp-0.1.2.dist-info/RECORD

@@ -1,13 +0,0 @@
-radioactiveshrimp-0.1.2.dist-info/METADATA,sha256=eIFvbRoU_QV2ADRG--hldQQI256vfOcPQ6leK0jYZXI,412
-radioactiveshrimp-0.1.2.dist-info/WHEEL,sha256=VCoUqQjnl5wExkC-2PG3I2s7-FBXYum-vf1jh8qiYg8,97
-radioactiveshrimp/__init__.py,sha256=QY4ait-Lo2WQdyx7MU8N6a1v2ZYyAmYRqzYqdNbiavI,101
-radioactiveshrimp/_core.abi3.so,sha256=3IMwsP7UrgNWdKdw2Pl11_VUQF8NqPy1xyyPuzkREdA,489672
-radioactiveshrimp/_core.pyi,sha256=b6oJaUXUzEzqUE5rpqefV06hl8o_JCU8pgKgIIzQgmc,33
-radioactiveshrimp/differential/__init.py__,sha256=jOhQKr6xM_7TVUt0GWRddtkgBvNn5I873WZcArT-GtE,101
-radioactiveshrimp/differential/discrete.py,sha256=9-w6Avz4MZfu8gFhJU8kTvryn9oT2xS1TvvYQ_8pmPU,1371
-radioactiveshrimp/distributions/__init.py__,sha256=hZX2UFt54wJG4zDLuTTTonDSA-anWvfjhDmR-akYcn4,182
-radioactiveshrimp/distributions/cvdistributions.py,sha256=UEiO_RYFQYqHybI_wcOdpOrxHToXRpLFsB58dwvI4vU,2599
-radioactiveshrimp/matrix/__init.py__,sha256=EoerrXlfmd2lI9grb3jxqIhHn-bzLtYhRKyxI0RVgaE,166
-radioactiveshrimp/matrix/elementary.py,sha256=zPiBG24F5zIncrrVWbiXEp1oFnHw2xNKy5quF0GBOuc,6333
-radioactiveshrimp/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-radioactiveshrimp-0.1.2.dist-info/RECORD,,