main.cpp

#include "engine.hpp"
#include "nn.hpp"
#include <iostream>
#include <vector>
#include <fstream>
#include <sstream>
#include <string>

using namespace micrograd;

// Helper to load CSV data
void load_data(const std::string& filename, std::vector<std::vector<double>>& inputs, std::vector<std::vector<double>>& targets) {
    std::ifstream file(filename);
    if (!file.is_open()) {
        std::cerr << "Error: Could not open file " << filename << std::endl;
        exit(1);
    }

    std::string line;
    while (std::getline(file, line)) {
        std::stringstream ss(line);
        std::string val;
        std::vector<double> row;
        while (std::getline(ss, val, ',')) {
            row.push_back(std::stod(val));
        }
        
        if (row.size() >= 2) {
            targets.push_back({row.back()});
            row.pop_back();
            inputs.push_back(row);
        }
    }
}

int main() {
    // 1. Load Data
    std::vector<std::vector<double>> X;
    std::vector<std::vector<double>> y;
    
    std::cout << "Loading data from moons.csv..." << std::endl;
    load_data("moons.csv", X, y);
    std::cout << "Loaded " << X.size() << " samples." << std::endl;

    // 2. Initialize Model
    MLP<double, ReLUActivation<double>, TanhActivation<double>, HeInit<double>> model(2, {16, 16, 1});
    
    // 3. Setup Optimizer
    SGD<decltype(model), double> optimizer(model, 0.001);

    // 4. Train
    std::cout << "Training..." << std::endl;
    train(model, optimizer, X, y, 100); 

    // 5. Evaluate and Export Grid for Visualization
    std::cout << "Generating decision boundary grid..." << std::endl;
    std::ofstream out_file("predictions.csv");
    
    // Grid ranges based on moons dataset
    double x_min = -1.5, x_max = 2.5;
    double y_min = -1.0, y_max = 1.5;
    double step = 0.1;

    for (double i = x_min; i <= x_max; i += step) {
        for (double j = y_min; j <= y_max; j += step) {
            std::vector<double> input = {i, j};
            auto output = model(input);
            out_file << i << "," << j << "," << output[0]->data << "\n";
        }
    }
    out_file.close();
    std::cout << "Predictions saved to predictions.csv" << std::endl;

    return 0;
}