Update app.py

app.py

@@ -1,41 +1,43 @@
-# app.py
-import os
 import gradio as gr
-from agent_core import AgentCore
-
-# ensure model cache stays inside the space quota
-os.environ.setdefault("TRANSFORMERS_CACHE", "/home/user/.cache")
-
-agent = AgentCore()
-
-INTRO = (
-    "👋 **Welcome to Aventra OS (Phase 4)**\n\n"
-    "- I remember important facts you tell me (with weighting + decay).\n"
-    "- Ask questions and I’ll retrieve the most relevant memories.\n"
-    "- Try:\n"
-    "  • *“My name is Tirrek. I run MotionBoys.”*\n"
-    "  • *“My favorite sport is basketball.”*\n"
-    "  • *“What’s my brand?”* or *“What do you know about me?”*"
-)
-
-def chat_fn(message, history):
-    # history is a list of (user, assistant). We only need `message`.
-    reply = agent.handle(message)
-    return reply
-
-with gr.Blocks(title="Aventra OS — Memory Core") as demo:
-    gr.Markdown("# 🧠 Aventra OS — Memory Core (Phase 4)")
-    gr.Markdown(INTRO)
-
-    chat = gr.ChatInterface(
-        fn=chat_fn,
-        chatbot=gr.Chatbot(height=420),
-        textbox=gr.Textbox(placeholder="Tell me a fact or ask a question…", lines=2, autofocus=True),
-        submit_btn="Send",
-        retry_btn=None,
-        undo_btn=None,
-        clear_btn="Clear",
-    )
-
-if __name__ == "__main__":
-    demo.launch()
+from sentence_transformers import SentenceTransformer
+import numpy as np
+from memory_store import MemoryStore
+from graph_view import visualize_graph
+
+# Initialize
+model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
+memory = MemoryStore()
+
+def process_input(user_input, chat_history):
+    user_input = user_input.strip()
+
+    # visualization command
+    if user_input.lower() in ["visualize memory", "show graph", "see memory map"]:
+        memories = getattr(memory, "memories", [])
+        if not memories:
+            return chat_history + [[user_input, "No memories yet to visualize."]]
+        visualize_graph(memories)
+        return chat_history + [[user_input, "🧠 Memory graph generated. Check the output window."]]
+
+    # embed input
+    vector = model.encode(user_input)
+    save_status = memory.add_memory(user_input, vector)
+
+    # retrieve related
+    related = memory.retrieve_relevant(vector)
+    recall = "\n".join([f"• {txt} (score: {round(score,4)})" for txt, score in related])
+
+    response = f"{save_status}\nHere’s what I recall most related:\n{recall}"
+    chat_history.append([user_input, response])
+    return chat_history
+
+with gr.Blocks() as demo:
+    gr.Markdown("# 🧠 Aventra Memory Interface")
+    chatbot = gr.Chatbot(height=500)
+    msg = gr.Textbox(label="Type your message...")
+    clear = gr.Button("Clear Chat")
+
+    msg.submit(process_input, [msg, chatbot], [chatbot])
+    clear.click(lambda: [], None, chatbot)
+
+demo.launch()