TinyTorch/tinytorch/core/_validation.py

# ╔═══════════════════════════════════════════════════════════════════════════════╗
# ║                        🚨 CRITICAL WARNING 🚨                                ║
# ║                     AUTOGENERATED! DO NOT EDIT!                              ║
# ║                                                                               ║
# ║  This file is AUTOMATICALLY GENERATED from source modules.                   ║
# ║  ANY CHANGES MADE HERE WILL BE LOST when modules are re-exported!            ║
# ║                                                                               ║
# ║  ✅ TO EDIT: modules/source/_validation/_validation_dev.py                   ║
# ║  ✅ TO EXPORT: Run 'tito module complete <module_name>'                      ║
# ║                                                                               ║
# ║  🛡️ STUDENT PROTECTION: This file contains critical fixes for Variable/      ║
# ║     Tensor compatibility. Editing it directly WILL break CIFAR-10 training.  ║
# ║                                                                               ║
# ║  🎓 LEARNING TIP: Work in modules/source/ - that's where real development    ║
# ║     happens! The tinytorch/ directory is just the compiled output.           ║
# ╚═══════════════════════════════════════════════════════════════════════════════╝

"""
TinyTorch Runtime Validation System

🛡️ **STUDENT PROTECTION SYSTEM**
This module provides runtime validation to detect when students accidentally
break critical Variable/Tensor compatibility in core functions.

**Purpose**: Prevent CIFAR-10 training failures due to core file modifications.
"""

import numpy as np
import warnings
from typing import Any, Callable, Optional


class TinyTorchValidationError(Exception):
    """Raised when critical TinyTorch functionality is broken."""
    pass


def validate_variable_tensor_compatibility():
    """
    🛡️ **STUDENT PROTECTION**: Validate that core functions handle Variables correctly.
    
    This function tests the critical Variable/Tensor compatibility that enables
    CIFAR-10 training. If this fails, students have likely edited core files.
    """
    try:
        # Import core components
        from tinytorch.core.tensor import Tensor
        from tinytorch.core.autograd import Variable
        from tinytorch.core.layers import matmul
        from tinytorch.core.activations import ReLU, Softmax
        from tinytorch.core.training import MeanSquaredError as MSELoss
        
        # Test 1: Matrix multiplication with Variables
        a = Variable(Tensor([[1, 2], [3, 4]]), requires_grad=True)
        b = Variable(Tensor([[5, 6], [7, 8]]), requires_grad=True)
        
        try:
            result = matmul(a, b)
            if not hasattr(result, 'requires_grad'):
                raise TinyTorchValidationError("matmul doesn't return Variables properly")
        except Exception as e:
            raise TinyTorchValidationError(f"Matrix multiplication with Variables failed: {e}")
        
        # Test 2: ReLU with Variables
        relu = ReLU()
        x = Variable(Tensor([[-1, 0, 1]]), requires_grad=True)
        
        try:
            relu_result = relu(x)
            if not hasattr(relu_result, 'requires_grad'):
                raise TinyTorchValidationError("ReLU doesn't return Variables properly")
        except Exception as e:
            raise TinyTorchValidationError(f"ReLU with Variables failed: {e}")
        
        # Test 3: Softmax with Variables
        softmax = Softmax()
        x = Variable(Tensor([[1, 2, 3]]), requires_grad=True)
        
        try:
            softmax_result = softmax(x)
            if not hasattr(softmax_result, 'requires_grad'):
                raise TinyTorchValidationError("Softmax doesn't return Variables properly")
            # Check if it's a valid probability distribution
            prob_sum = np.sum(softmax_result.data.data)
            if not np.isclose(prob_sum, 1.0, atol=1e-6):
                raise TinyTorchValidationError("Softmax doesn't produce valid probabilities")
        except Exception as e:
            raise TinyTorchValidationError(f"Softmax with Variables failed: {e}")
        
        # Test 4: Loss function with Variables
        loss_fn = MSELoss()
        pred = Variable(Tensor([[0.1, 0.2, 0.7]]), requires_grad=True)
        true = Variable(Tensor([[0.0, 0.0, 1.0]]), requires_grad=False)
        
        try:
            loss = loss_fn(pred, true)
            # Handle Variable/Tensor data access properly
            if hasattr(loss.data, 'data'):
                loss_value = float(loss.data.data)
            elif hasattr(loss.data, '_data'):
                loss_value = float(loss.data._data)
            else:
                loss_value = float(loss.data)
            if not isinstance(loss_value, (int, float)) or np.isnan(loss_value):
                raise TinyTorchValidationError("Loss function doesn't return valid scalar")
        except Exception as e:
            raise TinyTorchValidationError(f"Loss function with Variables failed: {e}")
        
        return True
        
    except ImportError as e:
        raise TinyTorchValidationError(f"Core modules not available: {e}")


def validate_training_pipeline():
    """
    🛡️ **STUDENT PROTECTION**: Validate complete training pipeline works.
    
    Tests the full forward pass that CIFAR-10 training requires.
    """
    try:
        from tinytorch.core.tensor import Tensor
        from tinytorch.core.autograd import Variable
        from tinytorch.core.layers import Dense
        from tinytorch.core.activations import ReLU, Softmax
        from tinytorch.core.training import MeanSquaredError as MSELoss
        from tinytorch.core.optimizers import Adam
        
        # Create a mini neural network
        fc1 = Dense(10, 5)
        relu = ReLU()
        fc2 = Dense(5, 3)
        softmax = Softmax()
        
        # Make it trainable
        fc1.weights = Variable(fc1.weights.data, requires_grad=True)
        fc1.bias = Variable(fc1.bias.data, requires_grad=True)
        fc2.weights = Variable(fc2.weights.data, requires_grad=True)
        fc2.bias = Variable(fc2.bias.data, requires_grad=True)
        
        # Test forward pass
        x = Variable(Tensor(np.random.randn(2, 10)), requires_grad=False)
        h1 = fc1(x)
        h1_act = relu(h1)
        h2 = fc2(h1_act)
        output = softmax(h2)
        
        # Test loss computation
        target = Variable(Tensor(np.random.randn(2, 3)), requires_grad=False)
        loss_fn = MSELoss()
        loss = loss_fn(output, target)
        
        # Test optimizer
        optimizer = Adam([fc1.weights, fc1.bias, fc2.weights, fc2.bias], learning_rate=0.001)
        
        # Validate shapes are preserved
        original_bias_shape = fc1.bias.data.shape
        optimizer.step()  # This should not corrupt shapes
        
        if fc1.bias.data.shape != original_bias_shape:
            raise TinyTorchValidationError(f"Bias shape corrupted: {original_bias_shape} -> {fc1.bias.data.shape}")
        
        return True
        
    except Exception as e:
        raise TinyTorchValidationError(f"Training pipeline validation failed: {e}")


def run_student_protection_checks(verbose: bool = False):
    """
    🛡️ **MAIN PROTECTION FUNCTION**: Run all validation checks.
    
    This function should be called before CIFAR-10 training to ensure
    students haven't accidentally broken core functionality.
    
    Args:
        verbose: If True, print detailed validation results
        
    Returns:
        bool: True if all checks pass
        
    Raises:
        TinyTorchValidationError: If any critical functionality is broken
    """
    checks = [
        ("Variable/Tensor Compatibility", validate_variable_tensor_compatibility),
        ("Training Pipeline", validate_training_pipeline),
    ]
    
    if verbose:
        print("🛡️ Running TinyTorch Student Protection Checks...")
        print("=" * 60)
    
    for check_name, check_func in checks:
        try:
            check_func()
            if verbose:
                print(f"✅ {check_name}: PASSED")
        except TinyTorchValidationError as e:
            error_msg = f"""
🚨 CRITICAL ERROR: {check_name} validation failed!

{e}

🛡️ STUDENT PROTECTION TRIGGERED:
This error suggests that core TinyTorch files have been accidentally modified.

📋 TO FIX:
1. Check if you've edited any files in tinytorch/core/ directory
2. Those files are auto-generated and should NOT be edited directly
3. Make changes in modules/source/ instead
4. Run 'tito module complete <module>' to regenerate core files

⚠️  CIFAR-10 training will FAIL until this is fixed!
"""
            if verbose:
                print(f"❌ {check_name}: FAILED")
                print(error_msg)
            raise TinyTorchValidationError(error_msg)
    
    if verbose:
        print("=" * 60)
        print("🎉 All protection checks passed! CIFAR-10 training should work.")
    
    return True


def auto_validate_on_import():
    """
    🛡️ **AUTOMATIC PROTECTION**: Run validation when core modules are imported.
    
    This provides automatic protection without requiring students to
    remember to run validation checks.
    """
    try:
        run_student_protection_checks(verbose=False)
    except TinyTorchValidationError:
        # Only warn on import, don't crash
        warnings.warn(
            "🚨 TinyTorch core functionality may be broken. "
            "Run 'from tinytorch.core._validation import run_student_protection_checks; "
            "run_student_protection_checks(verbose=True)' for details.",
            UserWarning
        )


# Run automatic validation when this module is imported
# This provides silent protection for students
try:
    auto_validate_on_import()
except Exception:
    # Don't crash on import, just warn
    pass