juliacall

Common errors

• JuliaError: Couldn't find Julia. Ensure Julia is installed and on your PATH, or set the JULIA_BIN environment variable.

  cause The juliacall library could not locate the Julia executable on your system.
  fix Install Julia from https://julialang.org/downloads/. Add Julia's `bin` directory to your system's PATH, or set the `JULIA_BIN` environment variable to the full path of the Julia executable (e.g., `export JULIA_BIN=/opt/julia-1.9.3/bin/julia`).

• ModuleNotFoundError: No module named 'juliacall'

  cause The juliacall Python package is not installed in your current Python environment.
  fix Run `pip install juliacall` in your terminal to install the package.

• AttributeError: 'Julia' object has no attribute 'my_julia_function'

  cause The Julia function or variable you are trying to access has not been defined in the current Julia session or is not in the scope of the `Julia` or `Main` object you are using.
  fix Ensure the Julia code defining `my_julia_function` (e.g., via `jl.seval("my_julia_function(x) = x^2")`) has been executed *before* attempting to call it. Remember that `Julia()` instances are isolated, while `Main` shares the global Julia environment.

Warnings

• gotcha juliacall requires a separate Julia installation on your system. It is not automatically installed via pip and will not function without a correctly configured Julia runtime.
  Fix: Install Julia from https://julialang.org/downloads/. Ensure Julia's `bin` directory is on your system's PATH, or set the `JULIA_BIN` environment variable to the full path of the Julia executable (e.g., `/usr/local/julia-1.8.5/bin/julia`).
• gotcha Instantiating `juliacall.Julia()` creates a new, isolated Julia session, which can be slow and memory-intensive upon first call. For accessing an existing Julia environment or global state, `juliacall.Main` is often more appropriate and performs better after the initial load.
  Fix: For global state or reusing an existing Julia session, import `Main` instead of `Julia`: `from juliacall import Main; jl = Main`. If you need isolated environments, instantiate `Julia()` once and reuse the object.
• gotcha Automatic type conversions between Python and Julia can sometimes lead to unexpected behavior or performance overhead, especially for complex or mutable data structures. For example, Python lists might convert to Julia `Vector{Any}` which can be less performant than a specific `Vector{Int}`.
  Fix: Review the `juliacall` documentation on type conversion to understand the mappings. For critical performance or complex data, consider explicit type hints or using data structures that convert predictably, like NumPy arrays (which map to Julia arrays).

Install

• pip install juliacall Install juliacall

Imports

• Julia

  from juliacall import Julia

  Instantiates a new, isolated Julia session.
• Main

  from juliacall import Main

  Provides access to the main module of the current Julia session.

Quickstart

This quickstart demonstrates how to initialize a Julia session, evaluate Julia code, access Julia variables, and call both built-in and custom Julia functions from Python using juliacall.

from juliacall import Julia
import os

# For robust demos, ensure Julia is found (optional, but good practice if not on PATH)
# os.environ['JULIA_BIN'] = '/path/to/your/julia/bin/julia' # Uncomment and modify if needed

try:
    jl = Julia() # Initializes a Julia session
    
    # Evaluate Julia code
    jl.seval("x = 10; y = 20")
    print(f"Julia variable x: {jl.x}")
    print(f"Julia variable y: {jl.y}")
    
    # Call a Julia function
    result = jl.cos(jl.pi) # Access Julia's pi and cos function
    print(f"cos(pi) from Julia: {result}")

    # Define and call a custom Julia function
    jl.seval("my_add(a, b) = a + b")
    custom_result = jl.my_add(5, 7)
    print(f"Custom Julia function my_add(5, 7): {custom_result}")

except Exception as e:
    print(f"An error occurred: {e}")
    print("Please ensure Julia is installed and discoverable (e.g., on PATH or via JULIA_BIN env var).")