server.py•4.46 kB
# Interactive Feedback MCP
# Developed by Fábio Ferreira (https://x.com/fabiomlferreira)
# Inspired by/related to dotcursorrules.com (https://dotcursorrules.com/)
# Enhanced by Pau Oliva (https://x.com/pof) with ideas from https://github.com/ttommyth/interactive-mcp
import os
import sys
import json
import tempfile
import subprocess
import base64
from typing import Annotated, Dict, List, Optional
from fastmcp import FastMCP
from pydantic import Field
# The log_level is necessary for Cline to work: https://github.com/jlowin/fastmcp/issues/81
mcp = FastMCP("交互式反馈 MCP", log_level="ERROR")
def launch_feedback_ui(summary: str, predefinedOptions: list[str] | None = None) -> dict[str, str | list[str]]:
# 为反馈结果创建一个临时文件
with tempfile.NamedTemporaryFile(suffix=".json", delete=False) as tmp:
output_file = tmp.name
try:
# 获取相对于此脚本的feedback_ui.py路径
script_dir = os.path.dirname(os.path.abspath(__file__))
feedback_ui_path = os.path.join(script_dir, "feedback_ui.py")
# 作为单独的进程运行feedback_ui.py
# 注意：uv似乎有一个bug，所以我们需要
# 传递一堆特殊标志来使其工作
args = [
sys.executable,
"-u",
feedback_ui_path,
"--prompt", summary,
"--output-file", output_file,
"--predefined-options", "|||".join(predefinedOptions) if predefinedOptions else ""
]
result = subprocess.run(
args,
check=False,
shell=False,
stdout=subprocess.DEVNULL,
stderr=subprocess.DEVNULL,
stdin=subprocess.DEVNULL,
close_fds=True
)
if result.returncode != 0:
raise Exception(f"启动反馈UI失败: {result.returncode}")
# 从临时文件读取结果
with open(output_file, 'r') as f:
result_data = json.load(f)
os.unlink(output_file)
# 处理图片路径，将图片转换为base64
if 'image_paths' in result_data and result_data['image_paths']:
image_data = []
for img_path in result_data['image_paths']:
if os.path.exists(img_path):
try:
with open(img_path, 'rb') as img_file:
img_content = img_file.read()
img_base64 = base64.b64encode(img_content).decode('utf-8')
img_filename = os.path.basename(img_path)
image_data.append({
'filename': img_filename,
'content': img_base64,
'path': img_path
})
except Exception as e:
print(f"处理图片时出错: {e}")
# 添加图片数据到结果中
result_data['images'] = image_data
return result_data
except Exception as e:
if os.path.exists(output_file):
os.unlink(output_file)
raise e
@mcp.tool()
def interactive_feedback(
message: str = Field(description="向用户提出的具体问题"),
predefined_options: list = Field(default=None, description="提供给用户选择的预定义选项（可选）"),
) -> Dict[str, str | List[Dict[str, str]]]:
"""向用户请求交互式反馈，支持文本和图片"""
# 如果没有提供预定义选项，使用默认选项
predefined_options_list = predefined_options if isinstance(predefined_options, list) else None
# 确保预定义选项列表不为空
if not predefined_options_list:
predefined_options_list = [
"已解决当前问题",
"进一步优化程序",
"进一步优化界面",
"还有一些问题需要修复",
"没有修复任何错误",
]
result = launch_feedback_ui(message, predefined_options_list)
# 构建返回结果
response = {
'interactive_feedback': result.get('interactive_feedback', '')
}
# 如果有图片，添加到返回结果中
if 'images' in result and result['images']:
response['images'] = result['images']
return response
if __name__ == "__main__":
mcp.run(transport="stdio")
