feat(跨平台): 优化消息处理并添加纯文本提取功能

添加 extract_text_only 函数过滤非文本标记
修改翻译逻辑仅处理纯文本内容
完善附件处理和消息内容拼接
修复仅包含表情时的消息处理问题

adapters/router.py

@@ -227,6 +227,9 @@ class DiscordToOneBotConverter:
         Returns:
             伪装后的 OneBot 事件对象
         """
+        # 在方法内部导入 logger，避免作用域问题
+        from core.utils.logger import logger
+        
         # 1. 提取基础信息
         user_id = discord_message.author.id
         message_id = discord_message.id
@@ -279,38 +282,43 @@ class DiscordToOneBotConverter:
             else:
                 message_list.append(OneBotMessageSegment.text(content))
         
+        # 如果消息只包含表情（没有文本），更新 raw_message 以包含表情信息
+        if not raw_message.strip() or raw_message.strip().startswith('<'):
+            import re
+            raw_message = re.sub(r'<a?:([^:]+):(\d+)>', r'[\1]', raw_message)
+        
         # 添加附件信息
         if discord_message.attachments:
-            self.logger.debug(f"[DiscordToOneBotConverter] 检测到 {len(discord_message.attachments)} 个附件")
+            logger.debug(f"[DiscordToOneBotConverter] 检测到 {len(discord_message.attachments)} 个附件")
             for attachment in discord_message.attachments:
                 filename = attachment.filename.lower()
-                self.logger.debug(f"[DiscordToOneBotConverter] 处理附件: {attachment.filename}, MIME: {attachment.content_type}")
+                logger.debug(f"[DiscordToOneBotConverter] 处理附件: {attachment.filename}, MIME: {attachment.content_type}")
                 # 检查是否是语音文件
                 if filename.endswith(('.amr', '.silk', '.mp3', '.wav', '.ogg', '.m4a')):
                     seg = OneBotMessageSegment.record(attachment.url)
                     seg.data["filename"] = attachment.filename
                     message_list.append(seg)
                     raw_message += f"\n[语音: {attachment.filename}]"
-                    self.logger.debug(f"[DiscordToOneBotConverter] 识别为语音文件: {attachment.filename}")
+                    logger.debug(f"[DiscordToOneBotConverter] 识别为语音文件: {attachment.filename}")
                 elif filename.endswith(('.mp4', '.avi', '.mkv', '.mov', '.flv', '.wmv')):
                     seg = OneBotMessageSegment.video(attachment.url)
                     seg.data["filename"] = attachment.filename
                     message_list.append(seg)
                     raw_message += f"\n[视频: {attachment.filename}]"
-                    self.logger.debug(f"[DiscordToOneBotConverter] 识别为视频文件: {attachment.filename}")
+                    logger.debug(f"[DiscordToOneBotConverter] 识别为视频文件: {attachment.filename}")
                 elif filename.endswith(('.png', '.jpg', '.jpeg', '.gif', '.webp')):
                     image_type = "gif" if filename.endswith('.gif') else None
                     seg = OneBotMessageSegment.image(attachment.url, image_type=image_type)
                     seg.data["filename"] = attachment.filename
                     message_list.append(seg)
                     raw_message += f"\n[图片: {attachment.filename}]"
-                    self.logger.debug(f"[DiscordToOneBotConverter] 识别为图片文件: {attachment.filename}")
+                    logger.debug(f"[DiscordToOneBotConverter] 识别为图片文件: {attachment.filename}")
                 else:
                     seg = OneBotMessageSegment.file(attachment.url)
                     seg.data["filename"] = attachment.filename
                     message_list.append(seg)
                     raw_message += f"\n[文件: {attachment.filename}]"
-                    self.logger.success(f"[DiscordToOneBotConverter] 识别为普通文件: {attachment.filename}")
+                    logger.success(f"[DiscordToOneBotConverter] 识别为普通文件: {attachment.filename}")
                     
         # 添加贴纸 (Stickers) 信息
         if hasattr(discord_message, 'stickers') and discord_message.stickers:

plugins/discord-cross/handlers.py

@@ -171,10 +171,12 @@ async def handle_discord_message_event(event: Any):
     logger.debug(f"[CrossPlatform] 开始处理 Discord 事件消息: channel_id={discord_channel_id}")
     
     if hasattr(event, 'message') and isinstance(event.message, list):
+        has_text_content = False
         for segment in event.message:
             if isinstance(segment, MessageSegment):
                 if segment.type == "text":
                     content += segment.data.get("text", "")
+                    has_text_content = True
                 elif segment.type == "image":
                     file_url = segment.data.get("url") or segment.data.get("file")
                     file_name = segment.data.get("filename")
@@ -183,6 +185,7 @@ async def handle_discord_message_event(event: Any):
                         attachment_item = {"type": "image", "url": str(file_url), "filename": file_name}
                         if attachment_item not in attachments:
                             attachments.append(attachment_item)
+                            content += f"\n[图片: {file_name}]\n"
                 elif segment.type == "video":
                     file_url = segment.data.get("url") or segment.data.get("file")
                     file_name = segment.data.get("filename")
@@ -191,6 +194,7 @@ async def handle_discord_message_event(event: Any):
                         attachment_item = {"type": "video", "url": str(file_url), "filename": file_name}
                         if attachment_item not in attachments:
                             attachments.append(attachment_item)
+                            content += f"\n[视频: {file_name}]\n"
                 elif segment.type == "record":
                     file_url = segment.data.get("url") or segment.data.get("file")
                     file_name = segment.data.get("filename")
@@ -199,6 +203,7 @@ async def handle_discord_message_event(event: Any):
                         attachment_item = {"type": "record", "url": str(file_url), "filename": file_name}
                         if attachment_item not in attachments:
                             attachments.append(attachment_item)
+                            content += f"\n[语音: {file_name}]\n"
                 elif segment.type == "file":
                     file_url = segment.data.get("url") or segment.data.get("file")
                     file_name = segment.data.get("filename")
@@ -207,12 +212,17 @@ async def handle_discord_message_event(event: Any):
                         attachment_item = {"type": "file", "url": str(file_url), "filename": file_name}
                         if attachment_item not in attachments:
                             attachments.append(attachment_item)
+                            content += f"\n[文件: {file_name}]\n"
                             logger.debug(f"[CrossPlatform] Discord 消息识别到文件: {file_name}, URL: {file_url}")
     else:
         content = event.raw_message or ""
     
     content = content.strip()
     
+    # 如果 content 为空但有附件（如只有表情），使用 raw_message 作为 content
+    if not content and attachments:
+        content = event.raw_message or ""
+    
     logger.debug(f"[CrossPlatform] Discord 消息内容: '{content}', 附件数量: {len(attachments)}")
     
     discord_username = getattr(event, 'discord_username', 'Unknown')

plugins/discord-cross/parser.py

@@ -4,10 +4,26 @@
 """
 import os
 import json
+import re
 from typing import Dict, List, Any
 from models.message import MessageSegment
+from core.utils.logger import logger
 from .config import config
 
+
+def extract_text_only(content: str) -> str:
+    """从消息内容中提取纯文本，过滤掉非文本标记"""
+    if not content:
+        return ""
+    
+    # 移除所有 [图片: xxx]、[视频: xxx]、[语音: xxx]、[文件: xxx] 等标记
+    text_only = re.sub(r'\s*\[(图片|视频|语音|文件):[^\]]+\]\s*', ' ', content)
+    
+    # 移除连续空格
+    text_only = re.sub(r'\s+', ' ', text_only).strip()
+    
+    return text_only
+
 async def parse_forward_nodes(nodes: List[Dict[str, Any]]) -> tuple[str, List[dict]]:
     """解析 OneBot 合并转发消息节点"""
     content_parts = []

plugins/discord-cross/sender.py

@@ -8,7 +8,7 @@ from core.utils.logger import logger
 from core.managers.redis_manager import redis_manager
 from .config import config
 from .translator import translate_with_deepseek
-from .parser import format_discord_to_qq_content, format_qq_to_discord_content
+from .parser import format_discord_to_qq_content, format_qq_to_discord_content, extract_text_only
 
 async def send_to_discord(channel_id: int, content: str, attachments: List[dict] = None, embed: dict = None):
     """发送消息到 Discord 频道"""
@@ -117,9 +117,13 @@ async def forward_discord_to_qq(
     logger.debug(f"[CrossPlatform] 格式化后的内容: '{formatted_content}', 图片列表: {image_list}")
     
     if formatted_content:
-        translated_content = await translate_with_deepseek(formatted_content, "zh-CN", channel_id, "en2zh")
+        # 只提取文本进行翻译，过滤掉非文本内容
-        if translated_content != formatted_content:
+        text_only = extract_text_only(formatted_content)
-            formatted_content = f"{formatted_content}\n\n━━━━━ 翻译 ━━━━━\n{translated_content}"
+        if text_only:
+            translated_content = await translate_with_deepseek(text_only, "zh-CN", channel_id, "en2zh")
+            if translated_content != text_only:
+                # 将翻译后的文本替换回原文本位置
+                formatted_content = formatted_content.replace(text_only, translated_content)
     
     await send_to_qq(target_qq_group, formatted_content, image_list)
     logger.success(f"[CrossPlatform] Discord 频道 {channel_id} -> QQ 群 {target_qq_group}")
@@ -154,9 +158,12 @@ async def forward_qq_to_discord(
     
     if embed and embed.get("description"):
         original_text = embed["description"]
-        translated_text = await translate_with_deepseek(original_text, "en", group_id, "zh2en")
+        # 只提取文本进行翻译
-        if translated_text != original_text:
+        text_only = extract_text_only(original_text)
-            embed["description"] = f"{original_text}\n\n**Translation:**\n{translated_text}"
+        if text_only:
+            translated_text = await translate_with_deepseek(text_only, "en", group_id, "zh2en")
+            if translated_text != text_only:
+                embed["description"] = embed["description"].replace(text_only, translated_text)
     
     for channel_id in target_channels:
         await send_to_discord(channel_id, formatted_content, image_list, embed)