Update project structure and code cleanup

Author: sairin1202

.dockerignore

@@ -1,2 +1,2 @@
 debug/
-*embeddings.json
+.env

Dockerfile

@@ -0,0 +1,117 @@
+from memu.llm import OpenAIClient
+from memu.memory import MemoryAgent
+import json
+from datetime import datetime
+
+import re
+import os
+import shutil
+
+character_test = None
+
+def load_locomo_conversation(sample_id: int, session_id: int|str|list, add_image_caption: bool = False, perserve_role: bool = True):
+
+    try:
+        with open(f"locomo/locomo10.json", "r") as f:
+            data = json.load(f)
+
+        sample = data[sample_id]
+        conversation = sample["conversation"]
+
+        speaker_a = conversation.get("speaker_a", "Speaker A")
+        speaker_b = conversation.get("speaker_b", "Speaker B")
+        speakers = [speaker_a, speaker_b]
+
+        if not perserve_role:
+            remap_role = {
+                speaker_a: "user",
+                speaker_b: "assistant"
+            }
+        
+        if isinstance(session_id, int):
+            session_ids = [session_id]
+        elif isinstance(session_id, str):
+            session_ids = [int(session_id)]
+        elif isinstance(session_id, list):
+            session_ids = [int(sid) for sid in session_id]
+
+        conversation_messages = []
+        for session_id in session_ids:
+            conversation_single = []
+            if f"session_{session_id+1}" in conversation:
+                session = conversation[f"session_{session_id+1}"]
+                
+                for message in session:
+                    speaker = message["speaker"]
+                    content = message["text"]
+
+                    if not perserve_role:
+                        speaker = remap_role.get(speaker, "Unknown")
+
+                    if add_image_caption and "blip_caption" in message:
+                        content = f"({speaker} shares {message['blip_caption']}.) {content}"
+
+                    conversation_single.append({
+                        "role": speaker,
+                        "content": content
+                    })
+            conversation_messages.append(conversation_single)
+
+        return conversation_messages, speakers
+
+    except Exception as e:
+        raise
+
+def load_debug_conversation(file_name: str):
+    if '.' not in file_name:
+        file_name = f"{file_name}.txt"
+    with open(f"debug/{file_name}", "r") as f:
+        raw = f.readlines()
+
+    conversation = []
+    speakers = set()
+    for line in raw:
+        if ": " in line:
+            role, content = line.split(": ", 1)
+            speakers.add(role)
+            conversation.append({"role": role, "content": content.strip()})
+
+    return conversation, list(speakers)
+
+def process_conversation(conversation: list[dict] = []):
+    """Process conversation with memory agent"""
+    
+    # Initialize LLM client
+    llm_client = OpenAIClient(model="gpt-4o-mini")
+    memory_agent = MemoryAgent(llm_client=llm_client, memory_dir="memory")
+    
+    # Process conversation
+    result = memory_agent.run(
+        conversation=conversation,
+        character_name=character_test or conversation[0]["role"],
+        max_iterations=20
+    )
+    
+    if result.get("success"):
+        print(f"✅ Processing completed - Iterations: {result.get('iterations', 0)}")
+        print(f"📁 Files generated: {len(result.get('files_generated', []))}")
+    else:
+        print(f"❌ Processing failed: {result.get('error')}")
+    
+    return result
+
+if __name__ == "__main__":
+    print("🌟 MEMORY AGENT DEMONSTRATION")
+    print("Loading and processing conversations...")
+    
+    conversations, speakers = load_locomo_conversation(sample_id=2, session_id=[0,1], add_image_caption=True, perserve_role=True)
+    character_test = speakers[1]
+    
+    print(f"Loaded {len(conversations)} conversations for character: {character_test}")
+    
+    # Process each conversation
+    for i, conversation in enumerate(conversations):
+        print(f"\n🔄 Processing conversation {i+1} of {len(conversations)}")
+        process_conversation(conversation)
+
+    print("\n✅ All conversations processed.")