已关闭
[Usage]: 使用msagent命令拉起msagent功能后，出现大量报错：AttributeError：_ARRAY_API not found #254
libotao316创建于 28 天前关闭于 21 天前
Mrtutu
28 天前 评论：
28 天前 评论：
👋 您好,欢迎向 MindStudio-Agent 提交 Issue！
我们已收到您的反馈，感谢你对开源社区的支持。🎉
📅处理时效： 维护团队将在24小时内 查看并回复您的问题（工作日）。
🔍自助查询： 在等待期间，建议您先查阅以下资料，可能已有解决方案：
📖 MindStudio-Agent 官方文档
📝 贡献者指南
请确保 Issue 描述清晰，包含复现步骤和日志，这将帮助我们更快定位问题。谢谢！
28 天前 添加了label：triaged
28 天前 添加了label：usage
Wangang Yu
27 天前 评论：
27 天前 评论：
/label add pending
27 天前 添加了label：pending
25 天前 关联了pull request：feat(install): 新增一键安装脚本（curl|bash 与 irm|iex），uv tool 隔离安装避免依赖冲突
fanglanyue
21 天前 评论：
21 天前 评论：
用户环境里已经存在一组"不配套"的底层数据处理组件:新的NumPy 2.x,配上了按旧 NumPy 1.x编译的pyarrow,安装并启动 mindstudio-agent后,默认拉起 msprof-mcp;它导入pandas时,pandas发现环境里有 pyarrow 并加载它,才把这个原本潜伏的冲突暴露出来。
关键澄清:
- mindstudio-agent/msprof-mcp 没有直接声明依赖 pyarrow。因此不是"安装mindstudio-agent后,pyarrow要求 NumPy<2"，实际是环境中已有的旧 pyarrow=15.0.2自身要求 numpy<2,但环境最终是numpy=2.4.6
推荐使用一键安装的方式curl -LsSf https://raw.gitcode.com/Ascend/msagent/raw/master/scripts/install.sh | bash，使用uv tool install安装工具，隔离虚拟环境，避免污染现有环境，解决版本配套问题。详见：https://gitcode.com/Ascend/msagent/blob/master/docs/zh/getting_started/install_guide.md#31-一键安装推荐
21 天前 issue状态由 TODO 改变为 DONE
21 天前 关闭了 issue
21 天前 添加了label：resolved
欢迎加入社区，感谢您对社区的贡献 🎉!
环境信息
欢迎加入社区，感谢您对社区的贡献 🎉!
(base) [root@orker-1
(base) [root@worker-1 -]#
(base) [root@worker-1 -]# pip show pandas ] grep Version
Version: 3.0.5
Licensed under the Apache License, Version 2.0 (the "License");
Version 2.0, January 2004
Python Software Foundation License Version 2.
the documentation are dual licensed under the PSF License Version 2
(base) [root@worker-1 ~]# pip show numpy I grep Version
Version: 2.4.6
(base) [root@worker-1 -]#
(base) [root@worker-1 -]#
(base) [root@worker-1 -]# msagent
None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
. Loading...
A module that was compiled using NumPy 1.x cannot be run in
NumPy 2.4.6 as it may crash. To support both 1.x and 2.x
versions of NumPy, modules must be compiled with NumPy 2.0.
Some module may need to rebuild instead e.g with 'pybind11>=2.12'
If you are a user of the module, the easiest solution will be to
downgrade to numpy<2' or try to upgrade the affected module
We expect that some modules will need time to support NumPy 2.
Traceback (most recent call last): File "/root/anaconda3/bin/msprof-mcp", line 5, in
from msprof mcp.server import main
File "/root/anaconda3/lib/python3.11/site-packages/msprof mcp/server.py", line 9, in module>
from .tools.trace view.trace view analvze import TraceViewAnalyzeTool
File "/root/anaconda3/Lib/python3.11/site-packages/msprof mcp/tools/trace view/trace view analyze.py", line 30, in
from .perfetto tool import FlowDataTool, sliceInfoTool, sliceFinderTool, sqlQueryTool
File "/root/anaconda3/Lib/python3.11/site packages/msprof mcp/tools/trace view/perfetto tool py", line 7, in
from .connection manager import BaseTool, ToolError
File "/root/anaconda3/Lib/python3.11/site-packages/msprof mcp/tools/trace view/connection manager.py", line 7, in
from perfetto.trace processor import TraceProcessor, TraceProcessorConfig
File "/root/anaconda3/Lib/python3.11/site-packages/perfetto/trace processor/ init py", line 16, in module
from perfetto.trace processor.api import TraceProcessor
File "/root/anaconda3/Lib/python3.11/site-packages/perfetto/trace processor/api.py", line 20, in
from perfetto.common query result iterator import QueryResultIterator
File "/root/anaconda3/Lib/python3.11/site packages/perfetto/common/query result iterator.py", line 21, in
import pandas as pd
File "/root/anaconda3/lib/python3.11/site packages/pandas/init py", line 22, in module>
from pandas.compat import (
File "/root/anaconda3/lib/python3.11/site packages/pandas/compat/_ init py", line 28, in module>
from pandas.compat.pyarrow import (
File "/root/anaconda3/Lib/python3.11/site-packages/pandas/compat/pyarrow.py", line 12, in module>
import pyarrow as pa
File "/root/anaconda3/lib/python3.11/site packages/pyarrow/_ int py", line 65, in module>
import pyarrow.lib as lib
AttributeError: ARRAY_API not found
. Loading...
A module that was compiled using NumPy 1.x cannot be run in
NumPy 2.4.6 as it may crash. To support both 1.x and 2.x
versions of NumPy, modules must be compiled with NumPy 2.0.
Some module may need to rebuild instead e.g. with 'pybind11>=2.12'.
If you are a user of the module, the easiest solution will be to
downgrade to 'numpy<2' or try to upgrade the affected module.
We expect that some modules will need time to support NumPy 2.
使用场景及问题
欢迎加入社区，感谢您对社区的贡献 🎉!
通过使用社区文档进行msagent工具安装之后，运行msagent出现大量报错：
(base) [root@orker-1
(base) [root@worker-1 -]#
(base) [root@worker-1 -]# pip show pandas ] grep Version
Version: 3.0.5
Licensed under the Apache License, Version 2.0 (the "License");
Version 2.0, January 2004
Python Software Foundation License Version 2.
the documentation are dual licensed under the PSF License Version 2
(base) [root@worker-1 ~]# pip show numpy I grep Version
Version: 2.4.6
(base) [root@worker-1 -]#
(base) [root@worker-1 -]#
(base) [root@worker-1 -]# msagent
None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
. Loading...
A module that was compiled using NumPy 1.x cannot be run in
NumPy 2.4.6 as it may crash. To support both 1.x and 2.x
versions of NumPy, modules must be compiled with NumPy 2.0.
Some module may need to rebuild instead e.g with 'pybind11>=2.12'
If you are a user of the module, the easiest solution will be to
downgrade to numpy<2' or try to upgrade the affected module
We expect that some modules will need time to support NumPy 2.
Traceback (most recent call last): File "/root/anaconda3/bin/msprof-mcp", line 5, in
from msprof mcp.server import main
File "/root/anaconda3/lib/python3.11/site-packages/msprof mcp/server.py", line 9, in module>
from .tools.trace view.trace view analvze import TraceViewAnalyzeTool
File "/root/anaconda3/Lib/python3.11/site-packages/msprof mcp/tools/trace view/trace view analyze.py", line 30, in
from .perfetto tool import FlowDataTool, sliceInfoTool, sliceFinderTool, sqlQueryTool
File "/root/anaconda3/Lib/python3.11/site packages/msprof mcp/tools/trace view/perfetto tool py", line 7, in
from .connection manager import BaseTool, ToolError
File "/root/anaconda3/Lib/python3.11/site-packages/msprof mcp/tools/trace view/connection manager.py", line 7, in
from perfetto.trace processor import TraceProcessor, TraceProcessorConfig
File "/root/anaconda3/Lib/python3.11/site-packages/perfetto/trace processor/ init py", line 16, in module
from perfetto.trace processor.api import TraceProcessor
File "/root/anaconda3/Lib/python3.11/site-packages/perfetto/trace processor/api.py", line 20, in
from perfetto.common query result iterator import QueryResultIterator
File "/root/anaconda3/Lib/python3.11/site packages/perfetto/common/query result iterator.py", line 21, in
import pandas as pd
File "/root/anaconda3/lib/python3.11/site packages/pandas/init py", line 22, in module>
from pandas.compat import (
File "/root/anaconda3/lib/python3.11/site packages/pandas/compat/_ init py", line 28, in module>
from pandas.compat.pyarrow import (
File "/root/anaconda3/Lib/python3.11/site-packages/pandas/compat/pyarrow.py", line 12, in module>
import pyarrow as pa
File "/root/anaconda3/lib/python3.11/site packages/pyarrow/_ int py", line 65, in module>
import pyarrow.lib as lib
AttributeError: ARRAY_API not found
. Loading...
A module that was compiled using NumPy 1.x cannot be run in
NumPy 2.4.6 as it may crash. To support both 1.x and 2.x
versions of NumPy, modules must be compiled with NumPy 2.0.
Some module may need to rebuild instead e.g. with 'pybind11>=2.12'.
If you are a user of the module, the easiest solution will be to
downgrade to 'numpy<2' or try to upgrade the affected module.
We expect that some modules will need time to support NumPy 2.
请排查解决，务必使用户能通过手册能正常使用工具。
欢迎加入社区，感谢您对社区的贡献 🎉!