mojo-python-interop

To call Python libraries from Mojo, import the target Python module, create Python values, and invoke APIs directly. Use the py= keyword to wrap Mojo primitives and handle Python exceptions as Mojo errors using the raises convention.

Exception handling for Python interop in Mojo uses a raises-based convention to catch Python exceptions as Mojo errors. This standardizes error propagation, ensuring failures in Python expressions or module invocations surface reliably in Mojo code.

Yes, Mojo primitives require type conversion using the py= keyword when passed to Python APIs, except for Bool types which use positional arguments. You can also implement ConvertibleToPython for automatic conversions or call to_python_object() explicitly.

To build a Python extension module in Mojo, use PythonModuleBuilder patterns to construct a shared library. Export a PyInit_<module_name> entry point and register your Mojo functions and types so Python can import and execute them natively.

Yes, PythonObject supports direct attribute access, calls, operators, indexing, slicing, iteration, and len operations. This allows you to manipulate Python collections and values natively in Mojo without triggering unnecessary type conversions.