spiro.core.config

spiro.core.config module

helper function to set runtime backend.

This module contains the function to define the runtime context which include backend.

Examples:

>>> from spiro.core.config import set_backend
>>> set_backend('tensorflow')
>>> import dgl

set_backend(backend)

Setup backend to the input one.

Returns:

Type	Description
	the name of predefined backend

Source code in core/config.py

def set_backend(backend):
    """Setup backend to the input one.

    Returns:
        the name of predefined backend

    Raises:
        if no valid backend setting is found in either environment variable or config file,
        stop the current run and print error message.

    """
    backend_name = backend
    if backend_name not in supported_backends():
        logger.warning(_BACKEND_NOT_FOUND)
        backend_name = _PYTORCH
    logger.debug("Using Backend: %s" % (backend_name))
    os.environ[_DGL_BACKEND] = backend_name
    global _BACKEND
    _BACKEND = backend_name