Function _find_imported_submodules

cloudpickle/cloudpickle.py:338–384

Find currently imported submodules used by a function. Submodules used by a function need to be detected and referenced for the function to work correctly at depickling time. Because submodules can be referenced as attribute of their parent package (``package.submodule``), we need a

(code, top_level_dependencies)

def _find_imported_submodules(code, top_level_dependencies):
    """Find currently imported submodules used by a function.

    Submodules used by a function need to be detected and referenced for the
    function to work correctly at depickling time. Because submodules can be
    referenced as attribute of their parent package (``package.submodule``), we
    need a special introspection technique that does not rely on GLOBAL-related
    opcodes to find references of them in a code object.

    Example:
    ```
    import concurrent.futures
    import cloudpickle
    def func():
        x = concurrent.futures.ThreadPoolExecutor
    if __name__ == '__main__':
        cloudpickle.dumps(func)
    ```
    The globals extracted by cloudpickle in the function's state include the
    concurrent package, but not its submodule (here, concurrent.futures), which
    is the module used by func. Find_imported_submodules will detect the usage
    of concurrent.futures. Saving this module alongside with func will ensure
    that calling func once depickled does not fail due to concurrent.futures
    not being imported
    """

    subimports = []
    # check if any known dependency is an imported package
    for x in top_level_dependencies:
        if (
            isinstance(x, types.ModuleType)
            and hasattr(x, "__package__")
            and x.__package__
        ):
            # check if the package has any currently loaded sub-imports
            prefix = x.__name__ + "."
            # A concurrent thread could mutate sys.modules,
            # make sure we iterate over a copy to avoid exceptions
            for name in list(sys.modules):
                # Older versions of pytest will add a "None" module to
                # sys.modules.
                if name is not None and name.startswith(prefix):
                    # check whether the function can address the sub-module
                    tokens = set(name[len(prefix) :].split("."))
                    if not tokens - set(code.co_names):
                        subimports.append(sys.modules[name])
    return subimports


# relevant opcodes