Create market_analysis_service.py

backend/app/services/market_analysis_service.py

@@ -0,0 +1,99 @@
+import pandas as pd
+import numpy as np
+from sklearn.model_selection import train_test_split
+from sklearn.ensemble import RandomForestRegressor
+from sklearn.metrics import mean_squared_error
+import requests
+import joblib
+import logging
+
+# Configure logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+class MarketAnalysisService:
+    def __init__(self, data_source_url):
+        self.data_source_url = data_source_url
+        self.model = None
+        self.data = None
+
+    def fetch_data(self):
+        """Fetch market data from a specified URL."""
+        try:
+            logger.info("Fetching data from %s", self.data_source_url)
+            response = requests.get(self.data_source_url)
+            response.raise_for_status()
+            self.data = pd.DataFrame(response.json())
+            logger.info("Data fetched successfully.")
+        except Exception as e:
+            logger.error("Error fetching data: %s", e)
+            raise
+
+    def preprocess_data(self):
+        """Preprocess the data for training."""
+        logger.info("Preprocessing data...")
+        # Example preprocessing steps
+        self.data.dropna(inplace=True)  # Remove missing values
+        self.data['date'] = pd.to_datetime(self.data['date'])  # Convert date column
+        self.data.set_index('date', inplace=True)  # Set date as index
+        self.data['price_change'] = self.data['price'].pct_change()  # Calculate price change
+        self.data.dropna(inplace=True)  # Drop NaN values after calculation
+        logger.info("Data preprocessing completed.")
+
+    def train_model(self):
+        """Train the Random Forest model on the market data."""
+        logger.info("Training model...")
+        X = self.data.drop(columns=['price', 'price_change'])
+        y = self.data['price']
+
+        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
+
+        self.model = RandomForestRegressor(n_estimators=100, random_state=42)
+        self.model.fit(X_train, y_train)
+
+        # Evaluate the model
+        predictions = self.model.predict(X_test)
+        mse = mean_squared_error(y_test, predictions)
+        logger.info("Model training completed with MSE: %.2f", mse)
+
+    def predict(self, input_data):
+        """Make predictions using the trained model."""
+        if self.model is None:
+            logger.error("Model is not trained yet. Call train_model() first.")
+            raise Exception("Model not trained")
+
+        logger.info("Making predictions...")
+        input_df = pd.DataFrame(input_data)
+        predictions = self.model.predict(input_df)
+        return predictions
+
+    def save_model(self, file_path):
+        """Save the trained model to a file."""
+        joblib.dump(self.model, file_path)
+        logger.info("Model saved to %s", file_path)
+
+    def load_model(self, file_path):
+        """Load a trained model from a file."""
+        self.model = joblib.load(file_path)
+        logger.info("Model loaded from %s", file_path)
+
+# Example usage
+if __name__ == "__main__":
+    data_source = "https://api.example.com/market_data"  # Replace with actual data source
+    service = MarketAnalysisService(data_source)
+
+    service.fetch_data()
+    service.preprocess_data()
+    service.train_model()
+
+    # Example input for prediction
+    input_data = {
+        'feature1': [0.5],
+        'feature2': [1.2],
+        # Add other features as necessary
+    }
+    predictions = service.predict(input_data)
+    print("Predictions:", predictions)
+
+    # Save the model
+    service.save_model("market_analysis_model.pkl")