{
  "parameters": {
    "url": "http://training-service:8083/api/v1/train",
    "sendHeaders": true,
    "headerParameters": {
      "parameters": [
        {
          "name": "Authorization",
          "value": "Bearer {{ $env.TRAINING_SERVICE_TOKEN }}"
        }
      ]
    },
    "sendBody": true,
    "bodyParameters": {
      "parameters": [
        {
          "name": "feature_job_id",
          "value": "={{ $node['Trigger Feature Engineering'].json.job_id }}"
        },
        {
          "name": "mlflow_run_id",
          "value": "={{ $node['Create MLflow Run'].json.run.info.run_id }}"
        }
      ]
    }
  },
  "name": "🧠 Trigger Model Training",
  "type": "n8n-nodes-base.httpRequest"
}

// Complete ML Model Training Pipeline in n8n
const items = $input.all();
const trainingData = items.slice(1); // Skip summary

// Simple Linear Regression Implementation
class SimpleLinearRegression {
  constructor() {
    this.weights = null;
    this.bias = 0;
    this.learningRate = 0.01;
    this.epochs = 1000;
  }
  
  fit(X, y) {
    const numFeatures = X[0].length;
    this.weights = Array(numFeatures).fill(0).map(() => Math.random() * 0.01);
    
    for (let epoch = 0; epoch < this.epochs; epoch++) {
      for (let i = 0; i < X.length; i++) {
        const prediction = this.predict_single(X[i]);
        const error = prediction - y[i];
        
        // Update weights using gradient descent
        for (let j = 0; j < numFeatures; j++) {
          this.weights[j] -= this.learningRate * error * X[i][j];
        }
        this.bias -= this.learningRate * error;
      }
    }
  }
  
  predict_single(x) {
    let prediction = this.bias;
    for (let i = 0; i < x.length; i++) {
      prediction += this.weights[i] * x[i];
    }
    return prediction;
  }
}

// Train and validate model
const model = new SimpleLinearRegression();
model.fit(trainX, trainY);

// Return results for next node
return [{ 
  json: {
    model_id: model_1758755784697, // model_timestamp
    weights: model.weights,
    bias: model.bias,
    validation_r2: calculateR2(valY, valPredictions),
    trained_at: new Date().toISOString()
  }
}];