Merge remote-tracking branch 'upstream/debug' into debug

src/plugins/chat/config.py

@@ -104,6 +104,12 @@ class BotConfig:
     memory_ban_words: list = field(
         default_factory=lambda: ["表情包", "图片", "回复", "聊天记录"]
     )  # 添加新的配置项默认值
+    
+    # 是否优先使用当前群组的记忆
+    memory_group_priority: bool = True  # 默认开启群组记忆优先
+    
+    # 群组记忆私有化
+    memory_private_groups: dict = field(default_factory=dict)  # 群组私有记忆配置
 
     @staticmethod
     def get_config_dir() -> str:
@@ -304,6 +310,12 @@ class BotConfig:
                 config.memory_forget_time = memory_config.get("memory_forget_time", config.memory_forget_time)
                 config.memory_forget_percentage = memory_config.get("memory_forget_percentage", config.memory_forget_percentage)
                 config.memory_compress_rate = memory_config.get("memory_compress_rate", config.memory_compress_rate)
+                # 添加对memory_group_priority配置项的加载
+                config.memory_group_priority = memory_config.get("memory_group_priority", config.memory_group_priority)
+            
+            if config.INNER_VERSION in SpecifierSet(">=0.0.9"):
+                # 添加群组记忆私有化配置项的加载
+                config.memory_private_groups = memory_config.get("memory_private_groups", {})
 
         def mood(parent: dict):
             mood_config = parent["mood"]

src/plugins/chat/emoji_manager.py

@@ -6,6 +6,8 @@ import random
 import time
 import traceback
 from typing import Optional, Tuple
+from PIL import Image
+import io
 
 from loguru import logger
 from nonebot import get_driver
@@ -192,11 +194,11 @@ class EmojiManager:
             logger.error(f"获取标签失败: {str(e)}")
             return None
 
-    async def _check_emoji(self, image_base64: str) -> str:
+    async def _check_emoji(self, image_base64: str, image_format: str) -> str:
         try:
             prompt = f'这是一个表情包，请回答这个表情包是否满足\"{global_config.EMOJI_CHECK_PROMPT}\"的要求，是则回答是，否则回答否，不要出现任何其他内容'
 
-            content, _ = await self.vlm.generate_response_for_image(prompt, image_base64)
+            content, _ = await self.vlm.generate_response_for_image(prompt, image_base64, image_format)
             logger.debug(f"输出描述: {content}")
             return content
 
@@ -237,7 +239,7 @@ class EmojiManager:
                 
                 image_bytes = base64.b64decode(image_base64)
                 image_hash = hashlib.md5(image_bytes).hexdigest()
-                
+                image_format = Image.open(io.BytesIO(image_bytes)).format.lower()
                 # 检查是否已经注册过
                 existing_emoji = self.db['emoji'].find_one({'filename': filename})
                 description = None
@@ -278,7 +280,7 @@ class EmojiManager:
 
 
                 if global_config.EMOJI_CHECK:
-                    check = await self._check_emoji(image_base64)
+                    check = await self._check_emoji(image_base64, image_format)
                     if '是' not in check:
                         os.remove(image_path)
                         logger.info(f"描述: {description}")

src/plugins/chat/prompt_builder.py

@@ -91,12 +91,20 @@ class PromptBuilder:
         memory_prompt = ''
         start_time = time.time()
 
+        # 获取群组ID
+        group_id = None
+        if stream_id:
+            chat_stream = chat_manager.get_stream(stream_id)
+            if chat_stream and chat_stream.group_info:
+                group_id = chat_stream.group_info.group_id
+
         # 调用 hippocampus 的 get_relevant_memories 方法
         relevant_memories = await hippocampus.get_relevant_memories(
             text=message_txt,
             max_topics=5,
             similarity_threshold=0.4,
-            max_memory_num=5
+            max_memory_num=5,
+            group_id=group_id  # 传递群组ID
         )
 
         if relevant_memories:

src/plugins/chat/utils_image.py

@@ -4,6 +4,8 @@ import time
 import aiohttp
 import hashlib
 from typing import Optional, Union
+from PIL import Image
+import io
 
 from loguru import logger
 from nonebot import get_driver
@@ -119,6 +121,7 @@ class ImageManager:
                     
             # 计算哈希值
             image_hash = hashlib.md5(image_bytes).hexdigest()
+            image_format = Image.open(io.BytesIO(image_bytes)).format.lower()
             
             # 查重
             existing = self.db.images.find_one({'hash': image_hash})
@@ -127,7 +130,7 @@ class ImageManager:
                 
             # 生成文件名和路径
             timestamp = int(time.time())
-            filename = f"{timestamp}_{image_hash[:8]}.jpg"
+            filename = f"{timestamp}_{image_hash[:8]}.{image_format}"
             file_path = os.path.join(self.IMAGE_DIR, filename)
             
             # 保存文件
@@ -238,7 +241,8 @@ class ImageManager:
             # 计算图片哈希
             image_bytes = base64.b64decode(image_base64)
             image_hash = hashlib.md5(image_bytes).hexdigest()
-            
+            image_format = Image.open(io.BytesIO(image_bytes)).format.lower()
+
             # 查询缓存的描述
             cached_description = self._get_description_from_db(image_hash, 'emoji')
             if cached_description:
@@ -247,13 +251,13 @@ class ImageManager:
 
             # 调用AI获取描述
             prompt = "这是一个表情包，使用中文简洁的描述一下表情包的内容和表情包所表达的情感"
-            description, _ = await self._llm.generate_response_for_image(prompt, image_base64)
+            description, _ = await self._llm.generate_response_for_image(prompt, image_base64, image_format)
             
             # 根据配置决定是否保存图片
             if global_config.EMOJI_SAVE:
                 # 生成文件名和路径
                 timestamp = int(time.time())
-                filename = f"{timestamp}_{image_hash[:8]}.jpg"
+                filename = f"{timestamp}_{image_hash[:8]}.{image_format}"
                 file_path = os.path.join(self.IMAGE_DIR, 'emoji',filename)
                 
                 try:
@@ -293,7 +297,8 @@ class ImageManager:
             # 计算图片哈希
             image_bytes = base64.b64decode(image_base64)
             image_hash = hashlib.md5(image_bytes).hexdigest()
-            
+            image_format = Image.open(io.BytesIO(image_bytes)).format.lower()
+
             # 查询缓存的描述
             cached_description = self._get_description_from_db(image_hash, 'image')
             if cached_description:
@@ -302,7 +307,7 @@ class ImageManager:
 
             # 调用AI获取描述
             prompt = "请用中文描述这张图片的内容。如果有文字，请把文字都描述出来。并尝试猜测这个图片的含义。最多200个字。"
-            description, _ = await self._llm.generate_response_for_image(prompt, image_base64)
+            description, _ = await self._llm.generate_response_for_image(prompt, image_base64, image_format)
             
             print(f"描述是{description}")
             
@@ -314,7 +319,7 @@ class ImageManager:
             if global_config.EMOJI_SAVE:
                 # 生成文件名和路径
                 timestamp = int(time.time())
-                filename = f"{timestamp}_{image_hash[:8]}.jpg"
+                filename = f"{timestamp}_{image_hash[:8]}.{image_format}"
                 file_path = os.path.join(self.IMAGE_DIR,'image', filename)
                 
                 try:

src/plugins/memory_system/memory.py

@@ -4,6 +4,7 @@ import math
 import random
 import time
 import os
+from typing import Optional
 
 import jieba
 import networkx as nx
@@ -209,15 +210,31 @@ class Hippocampus:
 
         return chat_samples
 
-    async def memory_compress(self, messages: list, compress_rate=0.1):
+    async def memory_compress(self, messages: list, compress_rate=0.1, group_id=None):
         """压缩消息记录为记忆
         
+        Args:
+            messages: 消息记录列表
+            compress_rate: 压缩率
+            group_id: 群组ID，用于标记记忆来源
+            
         Returns:
             tuple: (压缩记忆集合, 相似主题字典)
         """
+        from ..chat.config import global_config  # 导入配置
+        
         if not messages:
             return set(), {}
 
+        # 确定记忆所属的群组
+        memory_group_tag = None
+        if group_id is not None:
+            # 查找群聊所属的群组
+            for group_name, group_ids in global_config.memory_private_groups.items():
+                if str(group_id) in group_ids:
+                    memory_group_tag = f"[群组:{group_name}]"
+                    break
+        
         # 合并消息文本，同时保留时间信息
         input_text = ""
         time_info = ""
@@ -267,7 +284,17 @@ class Hippocampus:
         for topic, task in tasks:
             response = await task
             if response:
-                compressed_memory.add((topic, response[0]))
+                memory_content = response[0]
+                
+                # 添加标记
+                # 优先使用群组标记
+                if memory_group_tag:
+                    memory_content = f"{memory_group_tag}{memory_content}"
+                # 如果没有群组标记但有群组ID，添加简单的群组ID标记
+                elif group_id is not None:
+                    memory_content = f"[群组:{group_id}]{memory_content}"
+                
+                compressed_memory.add((topic, memory_content))
                 # 为每个话题查找相似的已存在主题
                 existing_topics = list(self.memory_graph.G.nodes())
                 similar_topics = []
@@ -316,7 +343,17 @@ class Hippocampus:
             logger.debug(f"进度: [{bar}] {progress:.1f}% ({i}/{len(memory_samples)})")
 
             compress_rate = global_config.memory_compress_rate
-            compressed_memory, similar_topics_dict = await self.memory_compress(messages, compress_rate)
+            
+            # 尝试从消息中提取群组ID
+            group_id = None
+            if messages and len(messages) > 0:
+                first_msg = messages[0]
+                if 'group_id' in first_msg:
+                    group_id = first_msg['group_id']
+                    logger.info(f"检测到消息来自群组: {group_id}")
+            
+            # 传递群组ID到memory_compress
+            compressed_memory, similar_topics_dict = await self.memory_compress(messages, compress_rate, group_id)
             logger.info(f"压缩后记忆数量: {len(compressed_memory)}，似曾相识的话题: {len(similar_topics_dict)}")
             
             current_time = datetime.datetime.now().timestamp()
@@ -841,8 +878,21 @@ class Hippocampus:
         return activation
 
     async def get_relevant_memories(self, text: str, max_topics: int = 5, similarity_threshold: float = 0.4,
-                                    max_memory_num: int = 5) -> list:
-        """根据输入文本获取相关的记忆内容"""
+                                    max_memory_num: int = 5, group_id: Optional[int] = None) -> list:
+        """根据输入文本获取相关的记忆内容
+        
+        Args:
+            text: 输入文本
+            max_topics: 最大主题数
+            similarity_threshold: 相似度阈值
+            max_memory_num: 最大记忆数量
+            group_id: 群组ID，用于优先匹配当前群组的记忆
+            
+        Returns:
+            list: 相关记忆列表
+        """
+        from ..chat.config import global_config  # 导入配置
+        
         # 识别主题
         identified_topics = await self._identify_topics(text)
 
@@ -855,30 +905,134 @@ class Hippocampus:
 
         # 获取最相关的主题
         relevant_topics = self._get_top_topics(all_similar_topics, max_topics)
-
+        
+        # 确定记忆所属的群组
+        current_group_name = None
+        if group_id is not None:
+            # 查找群聊所属的群组
+            for group_name, group_ids in global_config.memory_private_groups.items():
+                if str(group_id) in group_ids:
+                    current_group_name = group_name
+                    break
+        
+        has_private_groups = len(global_config.memory_private_groups) > 0
+        
         # 获取相关记忆内容
         relevant_memories = []
+        group_related_memories = []  # 当前群聊的记忆
+        group_definition_memories = []  # 当前群组的记忆
+        public_memories = []  # 公共记忆
+        
         for topic, score in relevant_topics:
             # 获取该主题的记忆内容
             first_layer, _ = self.memory_graph.get_related_item(topic, depth=1)
             if first_layer:
                 # 如果记忆条数超过限制，随机选择指定数量的记忆
-                if len(first_layer) > max_memory_num / 2:
-                    first_layer = random.sample(first_layer, max_memory_num // 2)
+                if len(first_layer) > max_memory_num:
+                    first_layer = random.sample(first_layer, max_memory_num)
+                
                 # 为每条记忆添加来源主题和相似度信息
                 for memory in first_layer:
-                    relevant_memories.append({
+                    memory_info = {
                         'topic': topic,
                         'similarity': score,
                         'content': memory
-                    })
+                    }
+                    
+                    memory_text = str(memory)
+                    
+                    # 分类处理记忆
+                    if has_private_groups and group_id is not None:
+                        # 如果配置了私有群组且当前在群聊中
+                        if current_group_name:
+                            # 当前群聊属于某个群组
+                            if f"[群组:{current_group_name}]" in memory_text:
+                                # 当前群组的记忆
+                                group_definition_memories.append(memory_info)
+                            elif not any(f"[群组:" in memory_text for _ in range(1)):
+                                # 公共记忆
+                                public_memories.append(memory_info)
+                        else:
+                            # 当前群聊不属于任何群组
+                            if f"[群组:{group_id}]" in memory_text:
+                                # 当前群聊的特定记忆
+                                group_related_memories.append(memory_info)
+                            elif not any(f"[群组:" in memory_text for _ in range(1)):
+                                # 公共记忆
+                                public_memories.append(memory_info)
+                    elif global_config.memory_group_priority and group_id is not None:
+                        # 如果只启用了群组记忆优先
+                        if f"[群组:{group_id}]" in memory_text:
+                            # 当前群聊的记忆，放入群组相关记忆列表
+                            group_related_memories.append(memory_info)
+                        else:
+                            # 其他记忆，放入公共记忆列表
+                            public_memories.append(memory_info)
+                    else:
+                        # 如果没有特殊配置，所有记忆都放入相关记忆列表
+                        relevant_memories.append(memory_info)
 
-        # 如果记忆数量超过5个,随机选择5个
-        # 按相似度排序
-        relevant_memories.sort(key=lambda x: x['similarity'], reverse=True)
-
-        if len(relevant_memories) > max_memory_num:
-            relevant_memories = random.sample(relevant_memories, max_memory_num)
+        # 根据配置决定如何组合记忆
+        if has_private_groups and group_id is not None:
+            # 配置了私有群组且当前在群聊中
+            if current_group_name:
+                # 当前群聊属于某个群组
+                # 优先使用当前群组的记忆，如果不足再使用公共记忆
+                if len(group_definition_memories) >= max_memory_num:
+                    # 如果群组记忆足够，只使用群组记忆
+                    group_definition_memories.sort(key=lambda x: x['similarity'], reverse=True)
+                    relevant_memories = group_definition_memories[:max_memory_num]
+                else:
+                    # 如果群组记忆不足，添加公共记忆
+                    group_definition_memories.sort(key=lambda x: x['similarity'], reverse=True)
+                    public_memories.sort(key=lambda x: x['similarity'], reverse=True)
+                    
+                    relevant_memories = group_definition_memories.copy()
+                    remaining_count = max_memory_num - len(relevant_memories)
+                    if remaining_count > 0 and public_memories:
+                        selected_other = public_memories[:remaining_count]
+                        relevant_memories.extend(selected_other)
+            else:
+                # 当前群聊不属于任何群组
+                # 优先使用当前群聊的记忆，然后使用公共记忆
+                if len(group_related_memories) >= max_memory_num:
+                    # 如果当前群聊记忆足够，只使用当前群聊记忆
+                    group_related_memories.sort(key=lambda x: x['similarity'], reverse=True)
+                    relevant_memories = group_related_memories[:max_memory_num]
+                else:
+                    # 如果当前群聊记忆不足，添加公共记忆
+                    group_related_memories.sort(key=lambda x: x['similarity'], reverse=True)
+                    public_memories.sort(key=lambda x: x['similarity'], reverse=True)
+                    
+                    relevant_memories = group_related_memories.copy()
+                    remaining_count = max_memory_num - len(relevant_memories)
+                    if remaining_count > 0 and public_memories:
+                        selected_other = public_memories[:remaining_count]
+                        relevant_memories.extend(selected_other)
+        elif global_config.memory_group_priority and group_id is not None:
+            # 如果只启用了群组记忆优先
+            # 按相似度排序
+            group_related_memories.sort(key=lambda x: x['similarity'], reverse=True)
+            public_memories.sort(key=lambda x: x['similarity'], reverse=True)
+            
+            # 优先使用群组相关记忆，如果不足再使用其他记忆
+            if len(group_related_memories) >= max_memory_num:
+                # 如果群组相关记忆足够，只使用群组相关记忆
+                relevant_memories = group_related_memories[:max_memory_num]
+            else:
+                # 使用所有群组相关记忆
+                relevant_memories = group_related_memories.copy()
+                # 如果群组相关记忆不足，添加其他记忆
+                remaining_count = max_memory_num - len(relevant_memories)
+                if remaining_count > 0 and public_memories:
+                    # 从其他记忆中选择剩余需要的数量
+                    selected_other = public_memories[:remaining_count]
+                    relevant_memories.extend(selected_other)
+        else:
+            # 如果没有特殊配置，按相似度排序
+            relevant_memories.sort(key=lambda x: x['similarity'], reverse=True)
+            if len(relevant_memories) > max_memory_num:
+                relevant_memories = relevant_memories[:max_memory_num]
 
         return relevant_memories
 

src/plugins/models/utils_model.py

@@ -104,6 +104,7 @@ class LLM_request:
             endpoint: str,
             prompt: str = None,
             image_base64: str = None,
+            image_format: str = None,
             payload: dict = None,
             retry_policy: dict = None,
             response_handler: callable = None,
@@ -115,6 +116,7 @@ class LLM_request:
             endpoint: API端点路径 (如 "chat/completions")
             prompt: prompt文本
             image_base64: 图片的base64编码
+            image_format: 图片格式
             payload: 请求体数据
             retry_policy: 自定义重试策略
             response_handler: 自定义响应处理器
@@ -151,7 +153,7 @@ class LLM_request:
 
         # 构建请求体
         if image_base64:
-            payload = await self._build_payload(prompt, image_base64)
+            payload = await self._build_payload(prompt, image_base64, image_format)
         elif payload is None:
             payload = await self._build_payload(prompt)
 
@@ -172,7 +174,7 @@ class LLM_request:
                             if response.status == 413:
                                 logger.warning("请求体过大，尝试压缩...")
                                 image_base64 = compress_base64_image_by_scale(image_base64)
-                                payload = await self._build_payload(prompt, image_base64)
+                                payload = await self._build_payload(prompt, image_base64, image_format)
                             elif response.status in [500, 503]:
                                 logger.error(f"错误码: {response.status} - {error_code_mapping.get(response.status)}")
                                 raise RuntimeError("服务器负载过高，模型恢复失败QAQ")
@@ -294,7 +296,7 @@ class LLM_request:
                 new_params["max_completion_tokens"] = new_params.pop("max_tokens")
         return new_params
 
-    async def _build_payload(self, prompt: str, image_base64: str = None) -> dict:
+    async def _build_payload(self, prompt: str, image_base64: str = None, image_format: str = None) -> dict:
         """构建请求体"""
         # 复制一份参数，避免直接修改 self.params
         params_copy = await self._transform_parameters(self.params)
@@ -306,7 +308,7 @@ class LLM_request:
                         "role": "user",
                         "content": [
                             {"type": "text", "text": prompt},
-                            {"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{image_base64}"}}
+                            {"type": "image_url", "image_url": {"url": f"data:image/{image_format.lower()};base64,{image_base64}"}}
                         ]
                     }
                 ],
@@ -391,13 +393,14 @@ class LLM_request:
         )
         return content, reasoning_content
 
-    async def generate_response_for_image(self, prompt: str, image_base64: str) -> Tuple[str, str]:
+    async def generate_response_for_image(self, prompt: str, image_base64: str, image_format: str) -> Tuple[str, str]:
         """根据输入的提示和图片生成模型的异步响应"""
 
         content, reasoning_content = await self._execute_request(
             endpoint="/chat/completions",
             prompt=prompt,
-            image_base64=image_base64
+            image_base64=image_base64,
+            image_format=image_format
         )
         return content, reasoning_content