Open
Description
🐛 Describe the bug
Description
When running poetry install --with test on Windows, NumPy 2.0.2 installation fails due to missing C compiler. The current dependency requirements force NumPy to build from source instead of using pre-compiled wheels.
Environment
OS: Windows
Python version: 3.x (exact version to be specified)
Poetry version: 2.1.1
Project: mem0ai
Error Message
Installing dependencies from lock file
Package operations: 25 installs, 0 updates, 0 removals
- Installing numpy (2.0.2): Failed
PEP517 build of a dependency failed
Backend subprocess exited when trying to invoke build_wheel
ERROR: Unknown compiler(s): [['icl'], ['cl'], ['cc'], ['gcc'], ['clang'], ['clang-cl'], ['pgcc']]
The following exception(s) were encountered:
Running `icl ""` gave "[WinError 2] The system cannot find the file specified"
Running `cl /?` gave "[WinError 2] The system cannot find the file specified"
Running `cc --version` gave "[WinError 2] The system cannot find the file specified"
Running `gcc --version` gave "[WinError 2] The system cannot find the file specified"
Running `clang --version` gave "[WinError 2] The system cannot find the file specified"
Running `clang-cl /?` gave "[WinError 2] The system cannot find the file specified"
Running `pgcc --version` gave "[WinError 2] The system cannot find the file specified"
Steps to Reproduce
- Clone the repository
- Run poetry install --with test
Expected Behavior
Poetry should either:
- Use pre-compiled NumPy wheels instead of building from source
- Clearly document the need for Visual C++ build tools in the project's contributing guide
Suggested Solutions
- Pin NumPy to a version that has pre-compiled wheels for Windows
- Update contributing.md to mention the need for Visual C++ build tools
- Modify pyproject.toml to specify NumPy requirement in a way that prefers wheels over source builds
- Add a note about installing Visual Studio Build Tools in the setup instructions
Workarounds
For contributors who encounter this issue:
- Install Visual Studio Build Tools with "Desktop development with C++" workload
- Or install NumPy separately using pip install --only-binary=numpy numpy==2.0.2
Metadata
Metadata
Assignees
Labels
No labels