feat: added logging to dataidea

app.py

@@ -1,167 +1,181 @@
-import gradio as gr
-import joblib
-import numpy as np
-import pandas as pd
-from sklearn.ensemble import RandomForestRegressor
-from sklearn.preprocessing import MinMaxScaler
-
-# Define the ScalableModel class
-class ScalableModel:
-    def __init__(self, model, feature_scaler=None, target_scaler=None):
-        self.model = model
-        self.feature_scaler = feature_scaler if feature_scaler else MinMaxScaler()
-        self.target_scaler = target_scaler if target_scaler else MinMaxScaler()
-    
-    def predict(self, X_test):
-        X_test_scaled = self.feature_scaler.transform(X_test)
-        predictions_scaled = self.model.predict(X_test_scaled).reshape(-1, 1)
-        return self.target_scaler.inverse_transform(predictions_scaled).flatten()
-
-# Load the trained model and scalers
-xgboost_model = joblib.load("XGBoost_Model_v1.pkl")  # Ensure the model is saved
-
-
-# Feature names (MUST match your training data exactly)
-feature_names = [
-    'garlic_kgs_organic_pesticides', #
-    'ginger_kgs_organic_pesticides', #
-    'plastic_tanks_120_ltrs_liquid_manure', #
-    'sacks_liquid_manure', #
-    'hoes_tools', #
-    'spades_tools', #
-    'pick_axes_tools', #
-    'forked_hoes_tools', #
-    'pangas_tools', #
-    'wheelbarrows_tools', #
-    'trowels_tools', #
-    'watering_cans_tools', #
-    'spray_pumps_tools', #
-    'Beans_seeds', #
-    'Maize_seeds', #
-    'Soyabean_seeds', #
-    'Gnuts_seeds', #
-    'Irish Potatoes_seeds', #
-    # 'Cassava (bags)',
-    # 'Mugavu tree seedlings',
-    # 'Onions (Kg)_seeds',
-    # 'Millet.1_seeds',
-    'GPS-Altitude',
-    'Land_size_agriculture',
-    'Time_to_collect_Water_for_Household_use_Minutes',
-    'Beans_total_yield', #
-    'Cassava_total_yield',
-    'Maize_total_yield',
-    'Sweet_potatoes_total_yield',
-    'Food_banana_total_yield',
-    'Coffee_total_yield'
-]
-
-
-# Define prediction function
-def predict_agriculture_value(*features):
-    input_data = np.array(features).reshape(1, -1)
-    # Check if all input values are zero
-    if np.all(input_data == 0):
-        return 0.0  # Return zero directly
-    
-    df_input = pd.DataFrame(input_data, columns=feature_names)
-    
-    # Use the model's built-in predict function
-    prediction = xgboost_model.predict(df_input)
-    
-    return float(prediction[0])
-
-# Create Gradio interface
-input_components = [gr.Number(label=feature) for feature in feature_names]
-
-iface = gr.Interface(
-    fn=predict_agriculture_value,
-    inputs=input_components,
-    outputs=gr.Number(label="Predicted Agriculture Value (USD)"),
-    title="Agriculture Value Prediction",
-    description="Enter input values to predict Agriculture Value (USD) using the trained XGBoost model."
-)
-
-# Run the Gradio app
-if __name__ == "__main__":
-    iface.launch()
-
-
-# # XGBoost Model
-
-# import gradio as gr
-# import joblib
-# import numpy as np
-# import pandas as pd
-# from xgboost import XGBRegressor
-# from sklearn.preprocessing import MinMaxScaler
-
-# # Define ScalableModel class
-# class ScalableModel:
-#     def __init__(self, model, feature_scaler=None, target_scaler=None):
-#         self.model = model
-#         self.feature_scaler = feature_scaler if feature_scaler else MinMaxScaler()
-#         self.target_scaler = target_scaler if target_scaler else MinMaxScaler()
-
-#     def fit(self, X_train, y_train):
-#         self.feature_scaler.fit(X_train)
-#         self.target_scaler.fit(y_train.values.reshape(-1, 1))
-#         self.model.fit(self.feature_scaler.transform(X_train), self.target_scaler.transform(y_train.values.reshape(-1, 1)).flatten())
-
-#     def predict(self, X_test):
-#         X_test_scaled = self.feature_scaler.transform(X_test)
-#         predictions_scaled = self.model.predict(X_test_scaled)
-#         return self.target_scaler.inverse_transform(predictions_scaled.reshape(-1, 1)).flatten()
-
-# # Load the trained XGBoost model
-# xgboost_model = joblib.load("XGBoost_Model_v1.pkl")
-
-# # Load the scalers
-# scalers = joblib.load("Scalers_XGBoost.pkl")
-# feature_scaler = scalers["feature_scaler"]
-# target_scaler = scalers["target_scaler"]
-
-# # Define feature names (must match training data)
-# feature_names = [
-#     'garlic_kgs_organic_pesticides', 'ginger_kgs_organic_pesticides',
-#     'plastic_tanks_120_ltrs_liquid_manure', 'sacks_liquid_manure',
-#     'hoes_tools', 'spades_tools', 'pick_axes_tools', 'forked_hoes_tools',
-#     'pangas_tools', 'wheelbarrows_tools', 'trowels_tools', 'watering_cans_tools',
-#     'spray_pumps_tools', 'Beans_seeds', 'Maize_seeds', 'Soyabean_seeds',
-#     'Gnuts_seeds', 'Irish Potatoes_seeds', 'GPS-Altitude', 'Land_size_agriculture',
-#     'Time_to_collect_Water_for_Household_use_Minutes', 'Beans_total_yield',
-#     'Cassava_total_yield', 'Maize_total_yield', 'Sweet_potatoes_total_yield',
-#     'Food_banana_total_yield', 'Coffee_total_yield'
-# ]
-
-# # Define prediction function
-# def predict_agriculture_value(*features):
-#     input_data = np.array(features).reshape(1, -1)
-
-#     # Scale the input features
-#     input_data_scaled = feature_scaler.transform(input_data)
-
-#     # Get scaled predictions
-#     prediction_scaled = xgboost_model.predict(input_data_scaled).reshape(-1, 1)
-
-#     # Inverse transform to get predictions in original scale
-#     prediction = target_scaler.inverse_transform(prediction_scaled).flatten()
-
-#     return float(prediction[0])
-
-# # Create Gradio interface
-# input_components = [gr.Number(label=feature) for feature in feature_names]
-
-# iface = gr.Interface(
-#     fn=predict_agriculture_value,
-#     inputs=input_components,
-#     outputs=gr.Number(label="Predicted Agriculture Value (USD)"),
-#     title="Agriculture Value Prediction",
-#     description="Enter input values to predict Agriculture Value (USD) using the trained XGBoost model."
-# )
-
-# # Run the Gradio app
-# if __name__ == "__main__":
-#     iface.launch()
-
+import gradio as gr
+import joblib
+import numpy as np
+import pandas as pd
+from sklearn.ensemble import RandomForestRegressor
+from sklearn.preprocessing import MinMaxScaler
+from dataidea.logger import event_log
+
+# Define the ScalableModel class
+class ScalableModel:
+    def __init__(self, model, feature_scaler=None, target_scaler=None):
+        self.model = model
+        self.feature_scaler = feature_scaler if feature_scaler else MinMaxScaler()
+        self.target_scaler = target_scaler if target_scaler else MinMaxScaler()
+    
+    def predict(self, X_test):
+        X_test_scaled = self.feature_scaler.transform(X_test)
+        predictions_scaled = self.model.predict(X_test_scaled).reshape(-1, 1)
+        return self.target_scaler.inverse_transform(predictions_scaled).flatten()
+
+# Load the trained model and scalers
+xgboost_model = joblib.load("XGBoost_Model_v1.pkl")  # Ensure the model is saved
+
+
+# Feature names (MUST match your training data exactly)
+feature_names = [
+    'garlic_kgs_organic_pesticides', #
+    'ginger_kgs_organic_pesticides', #
+    'plastic_tanks_120_ltrs_liquid_manure', #
+    'sacks_liquid_manure', #
+    'hoes_tools', #
+    'spades_tools', #
+    'pick_axes_tools', #
+    'forked_hoes_tools', #
+    'pangas_tools', #
+    'wheelbarrows_tools', #
+    'trowels_tools', #
+    'watering_cans_tools', #
+    'spray_pumps_tools', #
+    'Beans_seeds', #
+    'Maize_seeds', #
+    'Soyabean_seeds', #
+    'Gnuts_seeds', #
+    'Irish Potatoes_seeds', #
+    # 'Cassava (bags)',
+    # 'Mugavu tree seedlings',
+    # 'Onions (Kg)_seeds',
+    # 'Millet.1_seeds',
+    'GPS-Altitude',
+    'Land_size_agriculture',
+    'Time_to_collect_Water_for_Household_use_Minutes',
+    'Beans_total_yield', #
+    'Cassava_total_yield',
+    'Maize_total_yield',
+    'Sweet_potatoes_total_yield',
+    'Food_banana_total_yield',
+    'Coffee_total_yield'
+]
+
+
+# Define prediction function
+def predict_agriculture_value(*features):
+    input_data = np.array(features).reshape(1, -1)
+    # Check if all input values are zero
+    if np.all(input_data == 0):
+        return 0.0  # Return zero directly
+    
+    df_input = pd.DataFrame(input_data, columns=feature_names)
+    
+    # Use the model's built-in predict function
+    prediction = xgboost_model.predict(df_input)
+
+    event_log({
+        'api_key': '1968c15b-ed45-4a2d-a7dc-90ce623324b8',
+        'project_name': 'IDS',
+        'user_id': 'Emmanuel Nsubuga',
+        'message': 'Prediction Log',
+        'level': 'info',
+        'metadata': {
+            'input_data': input_data.tolist(),
+            'prediction': prediction.tolist(),
+            'source': 'gradio'
+        } 
+    })
+    
+    return float(prediction[0])
+
+# Create Gradio interface
+input_components = [gr.Number(label=feature) for feature in feature_names]
+
+iface = gr.Interface(
+    fn=predict_agriculture_value,
+    inputs=input_components,
+    outputs=gr.Number(label="Predicted Agriculture Value (USD)"),
+    title="Agriculture Value Prediction",
+    description="Enter input values to predict Agriculture Value (USD) using the trained XGBoost model."
+)
+
+# Run the Gradio app
+if __name__ == "__main__":
+    iface.launch()
+
+
+# # XGBoost Model
+
+# import gradio as gr
+# import joblib
+# import numpy as np
+# import pandas as pd
+# from xgboost import XGBRegressor
+# from sklearn.preprocessing import MinMaxScaler
+
+# # Define ScalableModel class
+# class ScalableModel:
+#     def __init__(self, model, feature_scaler=None, target_scaler=None):
+#         self.model = model
+#         self.feature_scaler = feature_scaler if feature_scaler else MinMaxScaler()
+#         self.target_scaler = target_scaler if target_scaler else MinMaxScaler()
+
+#     def fit(self, X_train, y_train):
+#         self.feature_scaler.fit(X_train)
+#         self.target_scaler.fit(y_train.values.reshape(-1, 1))
+#         self.model.fit(self.feature_scaler.transform(X_train), self.target_scaler.transform(y_train.values.reshape(-1, 1)).flatten())
+
+#     def predict(self, X_test):
+#         X_test_scaled = self.feature_scaler.transform(X_test)
+#         predictions_scaled = self.model.predict(X_test_scaled)
+#         return self.target_scaler.inverse_transform(predictions_scaled.reshape(-1, 1)).flatten()
+
+# # Load the trained XGBoost model
+# xgboost_model = joblib.load("XGBoost_Model_v1.pkl")
+
+# # Load the scalers
+# scalers = joblib.load("Scalers_XGBoost.pkl")
+# feature_scaler = scalers["feature_scaler"]
+# target_scaler = scalers["target_scaler"]
+
+# # Define feature names (must match training data)
+# feature_names = [
+#     'garlic_kgs_organic_pesticides', 'ginger_kgs_organic_pesticides',
+#     'plastic_tanks_120_ltrs_liquid_manure', 'sacks_liquid_manure',
+#     'hoes_tools', 'spades_tools', 'pick_axes_tools', 'forked_hoes_tools',
+#     'pangas_tools', 'wheelbarrows_tools', 'trowels_tools', 'watering_cans_tools',
+#     'spray_pumps_tools', 'Beans_seeds', 'Maize_seeds', 'Soyabean_seeds',
+#     'Gnuts_seeds', 'Irish Potatoes_seeds', 'GPS-Altitude', 'Land_size_agriculture',
+#     'Time_to_collect_Water_for_Household_use_Minutes', 'Beans_total_yield',
+#     'Cassava_total_yield', 'Maize_total_yield', 'Sweet_potatoes_total_yield',
+#     'Food_banana_total_yield', 'Coffee_total_yield'
+# ]
+
+# # Define prediction function
+# def predict_agriculture_value(*features):
+#     input_data = np.array(features).reshape(1, -1)
+
+#     # Scale the input features
+#     input_data_scaled = feature_scaler.transform(input_data)
+
+#     # Get scaled predictions
+#     prediction_scaled = xgboost_model.predict(input_data_scaled).reshape(-1, 1)
+
+#     # Inverse transform to get predictions in original scale
+#     prediction = target_scaler.inverse_transform(prediction_scaled).flatten()
+
+#     return float(prediction[0])
+
+# # Create Gradio interface
+# input_components = [gr.Number(label=feature) for feature in feature_names]
+
+# iface = gr.Interface(
+#     fn=predict_agriculture_value,
+#     inputs=input_components,
+#     outputs=gr.Number(label="Predicted Agriculture Value (USD)"),
+#     title="Agriculture Value Prediction",
+#     description="Enter input values to predict Agriculture Value (USD) using the trained XGBoost model."
+# )
+
+# # Run the Gradio app
+# if __name__ == "__main__":
+#     iface.launch()
+
 # print("Expected feature order:", feature_names)