Skip to content

Latest commit

 

History

History
90 lines (68 loc) · 4.46 KB

faq.md

File metadata and controls

90 lines (68 loc) · 4.46 KB

FAQ & Troubleshooting

Is PythonCall/JuliaCall thread safe?

No.

However it is safe to use PythonCall with Julia with multiple threads, provided you only call Python code from the first thread. (Before v0.9.22, tricks such as disabling the garbage collector were required.)

From Python, to use JuliaCall with multiple threads you probably need to set [PYTHON_JULIACALL_HANDLE_SIGNALS=yes](@ref julia-config) before importing JuliaCall. This is because Julia intentionally causes segmentation faults as part of the GC safepoint mechanism. If unhandled, these segfaults will result in termination of the process. This is equivalent to starting julia with julia --handle-signals=yes, the default behavior in Julia. See discussion here for more information.

Related issues: #201, #202, #529

Issues when Numpy arrays are expected

When a Julia array is passed to Python, it is wrapped as a ArrayValue. This type satisfies the Numpy array interface and the buffer protocol, so can be used in most places where a numpy array is valid.

However, some Python packages have functions which assume the input is an actual Numpy array. You may see errors such as:

AttributeError: Julia: type Array has no field dtype

To fix this you can convert the array x to a Numpy array as follows

Py(x).to_numpy()

If the array is being mutated, you will need to pass the argument copy=false.

Related issues: #280

Heap corruption when using PyTorch

On some systems, you may see an error like the following when using torch and juliacall:

Python(65251,0x104cf8580) malloc: Heap corruption detected, free list is damaged at 0x600001c17280
*** Incorrect guard value: 1903002876
Python(65251,0x104cf8580) malloc: *** set a breakpoint in malloc_error_break to debug
[1]    65251 abort      ipython

A solution is to ensure that juliacall is imported before torch.

Related issues: #215

ccall requires the compiler error when importing some Python libraries

On some systems, you may see an error like the following when import e.g. matplotlib before juliacall:

ERROR: `ccall` requires the compilerTraceback (most recent call last):
  File "/home/dingraha/projects/pythoncall_import_error/run.py", line 2, in <module>
    from foo import Foo
  File "/home/dingraha/projects/pythoncall_import_error/foo.py", line 4, in <module>
    import juliacall; jl = juliacall.newmodule("FooModule")
  File "/home/dingraha/projects/pythoncall_import_error/venv/lib/python3.9/site-packages/juliacall/__init__.py", line 218, in <module>
    init()
  File "/home/dingraha/projects/pythoncall_import_error/venv/lib/python3.9/site-packages/juliacall/__init__.py", line 214, in init
    raise Exception('PythonCall.jl did not start properly')
Exception: PythonCall.jl did not start properly

The likely problem is that the "other" Python library (matplotlib, whatever) is loading the system libstdc++.so, which isn't compatible with the libstdc++.so that Julia ships with. Linux distributions with older libstdc++ versions seem more likely to suffer from this issue. The solution is to either:

  • use a Linux distribution with a more recent libstdc++
  • import juliacall before the other Python library, so that Julia's libstdc++ is loaded
  • use a Python from a conda environment, which will have a newer libstdc++ that is compatible with Julia's

Related issues: #255

Can I use JuliaCall to run Julia inside applications with embedded Python?

Yes, it may be possible. A good example of that is having Julia running inside the Python that is running inside Blender, as presented in this Discourse post. From the point that one has JuliaCall running inside Python, if it has access to the terminal, one can even launch a Julia REPL there, and if needed connect with VSCode Julia extension to it. The full Python script to install, launch JuliaCall, and launch a Julia REPL in Blender is here.