analysis model

app.py

@@ -1,6 +1,7 @@
 # app.py
 # The core FastAPI application for our IGUDAR model
 
+import os
 import joblib
 import pandas as pd
 import numpy as np
@@ -9,6 +10,7 @@ from pydantic import BaseModel
 import warnings
 warnings.filterwarnings('ignore')
 
+MODEL_DIR = "./models"
 # --- 1. DEFINE APP AND LOAD MODELS ---
 
 # Initialize the FastAPI app
@@ -35,6 +37,26 @@ except FileNotFoundError:
     print("❌ ERROR: Model or preprocessing files not found. Ensure they are in the /models directory.")
     model = None # Set to None to handle errors gracefully
 
+# ... (existing code to load valuation model) ...
+
+# --- NEW: Load the Investment Analytics Model ---
+ANALYTICS_MODEL_PATH = os.path.join(MODEL_DIR, "analytics_model.joblib")
+try:
+    analytics_bundle = joblib.load(ANALYTICS_MODEL_PATH)
+    analytics_models = analytics_bundle['models']
+    analytics_scaler = analytics_bundle['scaler']
+    analytics_encoders = analytics_bundle['label_encoders']
+    analytics_feature_names = analytics_bundle['feature_names']
+    analytics_target_names = analytics_bundle['target_names']
+    
+    print("✅ Investment Analytics models loaded successfully.")
+
+except Exception as e:
+    print(f"❌ ERROR: Could not load analytics model. Error: {e}")
+    analytics_models = None
+
+# --- (The rest of your startup code remains the same) ---
+
 # --- 2. DEFINE THE INPUT DATA MODEL ---
 
 # Pydantic model for input data validation.
@@ -140,6 +162,80 @@ def predict_valuation(property_data: PropertyFeatures):
         "model_used": "igudar_valuation_v1_xgboost"
     }
 
+# --- 4. CREATE THE NEW INVESTMENT ANALYTICS ENDPOINT ---
+
+@app.post("/analytics")
+def predict_analytics(property_data: PropertyFeatures):
+    """
+    Predicts a full suite of 22+ financial and investment indicators for a property.
+    """
+    if analytics_models is None:
+        raise HTTPException(status_code=500, detail="Analytics model is not loaded. Check server logs.")
+
+    # Convert incoming data to a dictionary
+    data_dict = property_data.dict()
+
+    # --- Feature Preparation (must match analytics training script) ---
+    
+    # Start with a dictionary of all zeros for our feature vector
+    features = {name: 0 for name in analytics_feature_names}
+
+    # Map input data to the feature vector
+    features.update({
+        'price_mad': data_dict.get('price_mad', 1500000), # Assume a price if not provided for analytics
+        'size_m2': data_dict.get('size_m2', 100),
+        'bedrooms': data_dict.get('bedrooms', 2),
+        'bathrooms': data_dict.get('bathrooms', 1),
+        'age_years': data_dict.get('age_years', 5),
+        'infrastructure_score': data_dict.get('infrastructure_score', 50),
+        'economic_score': data_dict.get('economic_score', 50),
+        'lifestyle_score': data_dict.get('lifestyle_score', 50),
+        'investment_score': data_dict.get('investment_score', 50),
+        'neighborhood_tier': data_dict.get('neighborhood_tier', 3),
+        'total_amenities': data_dict.get('total_amenities', 20),
+        'data_quality': data_dict.get('data_quality', 0.8)
+    })
+    
+    # Calculate price_per_m2
+    if features['size_m2'] > 0:
+        features['price_per_m2'] = features['price_mad'] / features['size_m2']
+
+    # Encode categorical features
+    for col, le in analytics_encoders.items():
+        encoded_col_name = f"{col}_encoded"
+        if encoded_col_name in features:
+            try:
+                value = data_dict.get(col)
+                encoded_value = le.transform([value])[0]
+                features[encoded_col_name] = encoded_value
+            except:
+                features[encoded_col_name] = 0 # Default for unseen categories
+    
+    # Create DataFrame in the exact order of feature_names
+    df = pd.DataFrame([features])[analytics_feature_names]
+
+    # Scale the features using the analytics scaler
+    df_scaled = analytics_scaler.transform(df)
+
+    # --- Get predictions from all specialist models ---
+    all_predictions = {}
+    for group_name, model in analytics_models.items():
+        # Predict the targets for this group
+        group_predictions = model.predict(df_scaled)[0]
+        
+        # Get the names of the targets this model predicts
+        target_list = analytics_bundle['training_metrics'][group_name]['targets']
+        
+        # Map the predicted values to their names
+        for i, target_name in enumerate(target_list):
+            all_predictions[target_name] = float(group_predictions[i]) # Convert to standard Python float
+
+    # Here you can add the logic to format the predictions into a nice structure
+    # like in your training script's demo (e.g., grouping by 'returns_forecast', 'risk_assessment').
+    # For simplicity, we'll return the raw predictions for now.
+    
+    return {"predicted_analytics": all_predictions}
+
 @app.get("/")
 def read_root():
     return {"message": "Welcome to the IGUDAR AI Valuation API. Use the /docs endpoint to test."}

models/analytics_model.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d8a30c33bd6ff6508a7d7ee41728a12be9869c77f5dd6a71bebe1733cf17fc3
+size 62758342