Update app.py

app.py

@@ -1,3 +1,5 @@
+import os
+import gc
 from flask import Flask, request, jsonify, render_template
 from flask_cors import CORS
 from brain import MairaBrain
@@ -5,27 +7,57 @@ from brain import MairaBrain
 app = Flask(__name__)
 CORS(app)
 
-# Initialize models here... (keep your current initialization code)
+# --- Configuration for the 5 Neural Cores ---
+# These match the filenames downloaded in your Dockerfile
+MODELS = {
+    "small": {"repo": "CyberCoder225/maira-model", "file": "SmolLM2-360M-Instruct.Q4_K_M.gguf"},
+    "medium": {"repo": "bartowski/Llama-3.2-1B-Instruct-GGUF", "file": "Llama-3.2-1B-Instruct-Q4_K_M.gguf"},
+    "qwen": {"repo": "Qwen/Qwen2.5-1.5B-Instruct-GGUF", "file": "qwen2.5-1.5b-instruct-q4_k_m.gguf"},
+    "danube": {"repo": "h2oai/h2o-danube3-500m-chat-GGUF", "file": "h2o-danube3-500m-chat-Q4_K_M.gguf"},
+    "granite": {"repo": "bartowski/granite-3.0-2b-instruct-GGUF", "file": "granite-3.0-2b-instruct-Q4_K_M.gguf"}
+}
+
+# --- Initialize Neural Cores (Lazy State) ---
+print("🌌 Initializing Neural Core Registry...")
+cores = {name: MairaBrain(cfg["repo"], cfg["file"]) for name, cfg in MODELS.items()}
+print("✅ Registry Ready. Standby for user input.")
 
 @app.route('/')
-def index():
-    # Looks for /templates/index.html
+def home():
+    """Serves the Chat UI from /templates/index.html"""
     return render_template('index.html')
 
 @app.route('/chat', methods=['POST'])
 def chat():
     try:
         data = request.json
-        msg = data.get("message")
-        m_type = data.get("model_type", "small")
-        
-        # Select brain based on user's manual choice
-        if m_type == "qwen": answer = maira_logic.get_response("boss", msg)
-        elif m_type == "granite": answer = maira_art.get_response("boss", msg)
-        elif m_type == "medium": answer = maira_prime.get_response("boss", msg)
-        elif m_type == "danube": answer = maira_chat.get_response("boss", msg)
-        else: answer = maira_lite.get_response("boss", msg)
-
-        return jsonify({"maira": answer})
+        user_input = data.get("message", "")
+        model_type = data.get("model_type", "small")
+        user_id = data.get("user_id", "CyberCoder225")
+
+        # --- RAM PROTECTION: UNLOAD OTHER CORES ---
+        # Hugging Face Free Tier has 16GB. We only keep ONE core in RAM at a time.
+        for name, core in cores.items():
+            if name != model_type:
+                core.unload() # Ensure brain.py has this method!
+
+        # --- GET RESPONSE ---
+        if model_type in cores:
+            active_core = cores[model_type]
+            answer = active_core.get_response(user_id, user_input)
+            return jsonify({
+                "maira": answer,
+                "metadata": {"core": model_type, "status": "active"}
+            })
+        else:
+            return jsonify({"error": "Invalid neural core selection"}), 400
+
     except Exception as e:
-        return jsonify({"error": str(e)}), 500
+        print(f"❌ CRITICAL ERROR: {str(e)}")
+        return jsonify({"error": "Neural Core Failure", "details": str(e)}), 500
+
+# --- START SERVER ---
+if __name__ == "__main__":
+    # PYTHONUNBUFFERED=1 in Dockerfile ensures these logs show up instantly
+    print("🚀 Maira Quintessence starting on port 7860...")
+    app.run(host="0.0.0.0", port=7860, debug=False)