opensource/dify
2
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-04-27 02:36:29 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat: Add Audio Content Support for MCP Tools (#27979)

Browse Source
This commit is contained in:
Will 2025-11-10 10:12:11 +08:00 committed by CodingOnStar
parent becabd84a9
commit ab39f39db0
1 changed files with 13 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
api/core/tools/mcp_tool/tool.py
View File

@ -1,16 +1,19 @@
import base64 import base64
import json import json
import logging
from collections.abc import Generator from collections.abc import Generator
from typing import Any from typing import Any
from core.mcp.auth_client import MCPClientWithAuthRetry from core.mcp.auth_client import MCPClientWithAuthRetry
from core.mcp.error import MCPConnectionError from core.mcp.error import MCPConnectionError
from core.mcp.types import CallToolResult, ImageContent, TextContent from core.mcp.types import AudioContent, CallToolResult, ImageContent, TextContent
from core.tools.__base.tool import Tool from core.tools.__base.tool import Tool
from core.tools.__base.tool_runtime import ToolRuntime from core.tools.__base.tool_runtime import ToolRuntime
from core.tools.entities.tool_entities import ToolEntity, ToolInvokeMessage, ToolProviderType from core.tools.entities.tool_entities import ToolEntity, ToolInvokeMessage, ToolProviderType
from core.tools.errors import ToolInvokeError from core.tools.errors import ToolInvokeError
logger = logging.getLogger(__name__)
class MCPTool(Tool): class MCPTool(Tool):
def __init__( def __init__(
@ -52,6 +55,11 @@ class MCPTool(Tool):
yield from self._process_text_content(content) yield from self._process_text_content(content)
elif isinstance(content, ImageContent): elif isinstance(content, ImageContent):
yield self._process_image_content(content) yield self._process_image_content(content)
elif isinstance(content, AudioContent):
yield self._process_audio_content(content)
else:
logger.warning("Unsupported content type=%s", type(content))
# handle MCP structured output # handle MCP structured output
if self.entity.output_schema and result.structuredContent: if self.entity.output_schema and result.structuredContent:
for k, v in result.structuredContent.items(): for k, v in result.structuredContent.items():
@ -97,6 +105,10 @@ class MCPTool(Tool):
"""Process image content and return a blob message.""" """Process image content and return a blob message."""
return self.create_blob_message(blob=base64.b64decode(content.data), meta={"mime_type": content.mimeType}) return self.create_blob_message(blob=base64.b64decode(content.data), meta={"mime_type": content.mimeType})
def _process_audio_content(self, content: AudioContent) -> ToolInvokeMessage:
"""Process audio content and return a blob message."""
return self.create_blob_message(blob=base64.b64decode(content.data), meta={"mime_type": content.mimeType})
def fork_tool_runtime(self, runtime: ToolRuntime) -> "MCPTool": def fork_tool_runtime(self, runtime: ToolRuntime) -> "MCPTool":
return MCPTool( return MCPTool(
entity=self.entity, entity=self.entity,