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
| import hashlib
from collections import OrderedDict
from typing import Optional
import tiktoken
class OptimizedContextManager:
"""优化的上下文管理器"""
def __init__(self, max_context_tokens: int = 8192, max_sessions: int = 1000):
self.max_context_tokens = max_context_tokens
self.max_sessions = max_sessions
self.conversations = OrderedDict()
self.tokenizer = tiktoken.get_encoding("cl100k_base")
self.memory_threshold = 0.8
self.cleanup_batch_size = 50
self.token_cache = {}
self.max_cache_size = 10000
self.stats = {
"total_messages": 0,
"context_compressions": 0,
"memory_cleanups": 0
}
def add_message(self, session_id: str, role: str, content: str) -> bool:
"""添加消息到上下文 - 优化版本"""
if self._is_memory_pressure():
self._emergency_cleanup()
content_tokens = self._count_tokens_cached(content)
if session_id not in self.conversations:
self.conversations[session_id] = {
"messages": [],
"total_tokens": 0,
"last_access": time.time()
}
conversation = self.conversations[session_id]
conversation["last_access"] = time.time()
self.conversations.move_to_end(session_id)
message = {
"role": role,
"content": content,
"timestamp": time.time(),
"tokens": content_tokens
}
conversation["messages"].append(message)
conversation["total_tokens"] += content_tokens
self.stats["total_messages"] += 1
if conversation["total_tokens"] > self.max_context_tokens:
self._compress_context(session_id)
if len(self.conversations) > self.max_sessions:
self._cleanup_lru_sessions()
return True
def get_context(self, session_id: str, max_tokens: Optional[int] = None) -> List[Dict]:
"""获取对话上下文 - 优化版本"""
if session_id not in self.conversations:
return []
conversation = self.conversations[session_id]
conversation["last_access"] = time.time()
self.conversations.move_to_end(session_id)
messages = conversation["messages"]
if max_tokens and conversation["total_tokens"] > max_tokens:
return self._truncate_to_token_limit(messages, max_tokens)
return messages.copy()
def _compress_context(self, session_id: str):
"""压缩上下文 - 梯度压缩策略"""
conversation = self.conversations[session_id]
messages = conversation["messages"]
if len(messages) <= 2:
return
recent_count = max(6, len(messages) // 4)
recent_messages = messages[-recent_count:]
older_messages = messages[:-recent_count]
if older_messages:
compressed_summary = self._create_context_summary(older_messages)
new_messages = [compressed_summary] + recent_messages
new_total_tokens = sum(msg["tokens"] for msg in new_messages)
conversation["messages"] = new_messages
conversation["total_tokens"] = new_total_tokens
self.stats["context_compressions"] += 1
print(f"🗜️ 上下文压缩: 会话 {session_id}, "
f"{len(older_messages)}条消息 -> 1条摘要")
def _create_context_summary(self, messages: List[Dict]) -> Dict:
"""创建上下文摘要"""
summary_content = f"[上下文摘要: {len(messages)}条消息]"
user_messages = [msg for msg in messages if msg["role"] == "user"]
if user_messages:
recent_content = user_messages[-1]["content"][:100] + "..."
summary_content += f" 最近话题: {recent_content}"
return {
"role": "system",
"content": summary_content,
"timestamp": time.time(),
"tokens": self._count_tokens_cached(summary_content),
"is_summary": True
}
def _truncate_to_token_limit(self, messages: List[Dict], max_tokens: int) -> List[Dict]:
"""动态截断到指定token限制"""
selected_messages = []
current_tokens = 0
for message in reversed(messages):
if current_tokens + message["tokens"] <= max_tokens:
selected_messages.insert(0, message)
current_tokens += message["tokens"]
else:
break
return selected_messages or messages[-1:] if messages else []
def _count_tokens_cached(self, text: str) -> int:
"""带缓存的token计算"""
text_hash = hashlib.md5(text.encode()).hexdigest()
if text_hash in self.token_cache:
return self.token_cache[text_hash]
tokens = len(self.tokenizer.encode(text))
if len(self.token_cache) < self.max_cache_size:
self.token_cache[text_hash] = tokens
return tokens
def _is_memory_pressure(self) -> bool:
"""检查内存压力"""
memory = psutil.virtual_memory()
return memory.percent > self.memory_threshold * 100
def _emergency_cleanup(self):
"""紧急内存清理"""
print("🚨 内存压力过高，执行紧急清理")
cleanup_count = min(self.cleanup_batch_size, len(self.conversations) // 4)
for _ in range(cleanup_count):
if self.conversations:
removed_session_id = next(iter(self.conversations))
self.conversations.pop(removed_session_id)
if len(self.token_cache) > self.max_cache_size // 2:
self.token_cache.clear()
self.stats["memory_cleanups"] += 1
def _cleanup_lru_sessions(self):
"""清理LRU会话"""
while len(self.conversations) > self.max_sessions:
oldest_session_id = next(iter(self.conversations))
self.conversations.pop(oldest_session_id)
def get_global_stats(self) -> Dict:
"""获取全局统计信息"""
total_tokens = sum(conv["total_tokens"] for conv in self.conversations.values())
return {
**self.stats,
"active_sessions": len(self.conversations),
"total_context_tokens": total_tokens,
"avg_tokens_per_session": total_tokens / max(len(self.conversations), 1),
"cache_size": len(self.token_cache),
"estimated_memory_mb": total_tokens * 4 / 1024 / 1024
}
|
