Spaces:
Running
Running
File size: 11,254 Bytes
1367957 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 |
"""
LangChain-based conversation memory management (v0.2+ compatible)
"""
from langchain_core.chat_history import BaseChatMessageHistory
from langchain_community.chat_message_histories import ChatMessageHistory
from langchain_classic.memory import ConversationBufferWindowMemory # Keep classic for now
from langchain_core.messages import HumanMessage, AIMessage, BaseMessage
from typing import List, Dict, Any, Optional
import json
import os
import pickle
from datetime import datetime
class ConversationMemory:
"""
Manages conversation memory using LangChain with persistent storage
Fixed for LangChain v0.2+ Pydantic v2 validation
"""
def __init__(self, session_id: str = "default", memory_window: int = 10):
self.session_id = session_id
self.memory_window = memory_window
# ✅ FIX: Create ChatMessageHistory INSTANCE (required by Pydantic v2)
chat_history: BaseChatMessageHistory = ChatMessageHistory()
# Initialize LangChain memory with proper chat_history
self.memory = ConversationBufferWindowMemory(
chat_memory=chat_history, # Pass INSTANCE, not dict
k=memory_window,
return_messages=True,
memory_key="chat_history",
output_key="output"
)
# Additional metadata storage
self.conversation_metadata = {
'session_id': session_id,
'domains_discussed': set(),
'query_types_used': set(),
'previously_used_papers': set(),
'interaction_count': 0
}
# Load existing memory if available
self._load_memory()
def add_interaction(self, user_message: str, ai_response: str, metadata: Dict[str, Any] = None):
"""Add a new interaction to memory"""
# Add to LangChain memory
self.memory.save_context(
{"input": user_message},
{"output": ai_response}
)
# Update metadata
self.conversation_metadata['interaction_count'] += 1
if metadata:
if 'domain' in metadata:
self.conversation_metadata['domains_discussed'].add(metadata['domain'])
if 'query_type' in metadata:
self.conversation_metadata['query_types_used'].add(metadata['query_type'])
if 'papers_used' in metadata:
# Track recently used papers to avoid repetition
paper_ids = metadata.get('paper_ids', [])
self.conversation_metadata['previously_used_papers'].update(paper_ids)
# Keep only recent papers (last 20)
recent_papers = list(self.conversation_metadata['previously_used_papers'])[-20:]
self.conversation_metadata['previously_used_papers'] = set(recent_papers)
# Save memory to persistent storage
self._save_memory()
def get_conversation_history(self, limit: Optional[int] = None) -> List[Dict[str, str]]:
"""Get conversation history"""
chat_history = self.memory.chat_memory.messages
history = []
for i in range(0, len(chat_history), 2):
if i + 1 < len(chat_history):
history.append({
'user': chat_history[i].content,
'assistant': chat_history[i + 1].content,
'turn': i // 2 + 1
})
if limit:
history = history[-limit:]
return history
def get_conversation_context(self) -> Dict[str, Any]:
"""Get current conversation context for query enhancement"""
history = self.get_conversation_history(limit=3) # Last 3 exchanges
context = {
'session_id': self.session_id,
'interaction_count': self.conversation_metadata['interaction_count'],
'domains_discussed': list(self.conversation_metadata['domains_discussed']),
'query_types_used': list(self.conversation_metadata['query_types_used']),
'previously_used_papers': list(self.conversation_metadata['previously_used_papers']),
'recent_history': history
}
# Extract last topic for context
if history:
last_interaction = history[-1]
context['last_user_message'] = last_interaction['user']
context['last_assistant_response'] = last_interaction['assistant']
context['last_topic'] = self._extract_topic(last_interaction['user'])
# Get last query type from metadata
if self.conversation_metadata['query_types_used']:
context['last_query_type'] = list(self.conversation_metadata['query_types_used'])[-1]
# Add last_domain from domains_discussed
if self.conversation_metadata['domains_discussed']:
context['last_domain'] = list(self.conversation_metadata['domains_discussed'])[-1]
return context
def get_conversation_summary(self) -> Dict[str, Any]:
"""Get summary of the conversation"""
history = self.get_conversation_history()
return {
'session_id': self.session_id,
'total_interactions': len(history),
'domains_covered': list(self.conversation_metadata['domains_discussed']),
'query_types_used': list(self.conversation_metadata['query_types_used']),
'papers_referenced': len(self.conversation_metadata['previously_used_papers']),
'recent_activity': [msg['user'][:50] + '...' for msg in history[-3:]]
}
def clear_memory(self):
"""Clear all conversation memory"""
self.memory.clear()
self.conversation_metadata = {
'session_id': self.session_id,
'domains_discussed': set(),
'query_types_used': set(),
'previously_used_papers': set(),
'interaction_count': 0
}
self._save_memory()
def _extract_topic(self, message: str) -> str:
"""Extract main topic from a message"""
# Simple topic extraction - can be enhanced
words = message.lower().split()
# Filter out common words and keep meaningful ones
stop_words = {'what', 'how', 'why', 'when', 'where', 'which', 'can', 'you', 'me', 'the', 'a', 'an', 'and', 'or',
'but'}
meaningful_words = [word for word in words if word not in stop_words and len(word) > 3]
return ' '.join(meaningful_words[:3]) if meaningful_words else 'general discussion'
def _get_memory_file_path(self) -> str:
"""Get file path for persistent memory storage"""
memory_dir = "./memory_data"
os.makedirs(memory_dir, exist_ok=True)
return f"{memory_dir}/memory_{self.session_id}.pkl"
def _save_memory(self):
"""Save memory to persistent storage"""
try:
# ✅ FIX: Use .dict() for serialization compatibility
memory_data = {
'langchain_memory': self.memory.dict(), # Fixed serialization
'conversation_metadata': self.conversation_metadata
}
with open(self._get_memory_file_path(), 'wb') as f:
pickle.dump(memory_data, f)
print(f"💾 Memory saved for session: {self.session_id}")
except Exception as e:
print(f"❌ Error saving memory: {e}")
def _load_memory(self):
"""Load memory from persistent storage"""
try:
memory_file = self._get_memory_file_path()
if os.path.exists(memory_file):
with open(memory_file, 'rb') as f:
memory_data = pickle.load(f)
# ✅ FIX: Recreate chat_history before initializing memory
chat_history = ChatMessageHistory()
memory_config = memory_data['langchain_memory']
memory_config['chat_memory'] = chat_history # Ensure proper instance
self.memory = ConversationBufferWindowMemory(**memory_config)
self.conversation_metadata = memory_data['conversation_metadata']
print(f"📂 Memory loaded for session: {self.session_id}")
except Exception as e:
print(f"❌ Error loading memory: {e}")
# Continue with fresh memory
# For Vercel serverless compatibility
class VercelMemoryManager:
"""
Memory manager optimized for Vercel serverless environment
Uses JSON files instead of pickle for compatibility
"""
def __init__(self, session_id: str = "default"):
self.session_id = session_id
self.memory_file = f"/tmp/memory_{session_id}.json"
self.conversation_history = []
self.load_memory()
def add_interaction(self, user_message: str, ai_response: str, metadata: Dict[str, Any] = None):
"""Add interaction to memory"""
interaction = {
'user': user_message,
'assistant': ai_response,
'metadata': metadata or {},
'timestamp': self._get_timestamp()
}
self.conversation_history.append(interaction)
# Keep only last 20 interactions in serverless environment
if len(self.conversation_history) > 20:
self.conversation_history = self.conversation_history[-20:]
self.save_memory()
def get_conversation_context(self) -> Dict[str, Any]:
"""Get conversation context"""
recent_history = self.conversation_history[-3:] if self.conversation_history else []
domains = set()
query_types = set()
for interaction in self.conversation_history:
if 'metadata' in interaction:
meta = interaction['metadata']
if 'domain' in meta:
domains.add(meta['domain'])
if 'query_type' in meta:
query_types.add(meta['query_type'])
return {
'session_id': self.session_id,
'interaction_count': len(self.conversation_history),
'domains_discussed': list(domains),
'query_types_used': list(query_types),
'recent_history': recent_history
}
def save_memory(self):
"""Save memory to JSON file"""
try:
with open(self.memory_file, 'w') as f:
json.dump(self.conversation_history, f)
except Exception as e:
print(f"❌ Error saving memory: {e}")
def load_memory(self):
"""Load memory from JSON file"""
try:
if os.path.exists(self.memory_file):
with open(self.memory_file, 'r') as f:
self.conversation_history = json.load(f)
except Exception as e:
print(f"❌ Error loading memory: {e}")
self.conversation_history = []
def _get_timestamp(self) -> str:
"""Get current timestamp"""
from datetime import datetime
return datetime.now().isoformat() |