Maybe a silly question.. but how does archiving/taring a pex file and supporting files work in practice? What I mean is.. if I have a tar that contains: • pex • index.html If I want to serve the index.html from the pex file.. would it be the path to whereever the tar was extracted?

We have resource directories that we can load via

importlib

, the directory itself doesn't have

__init__.py

in it, although it is nested within a top-level one that does. I think PEXes ~always run as actual files on disk these days, so using

open

+

__file__

can likely work too. I don't know of a way to identify the location of the original PEX file, although there might be some goodies hiding in https://pex.readthedocs.io/en/v2.1.138/api/vars.html

Got it. I’ll try again. Almost certainly a typo. Shrug. Another thing I don’t understand does importlib unzip things in a context (then clean up when it closes?) Essentially we load a bunch of static files into flask. Today people do this with ../../path/to/staticDir kind of thing.

Do you mean

<http://importlib.resources.as|importlib.resources.as>_file

? It seems to have documentation that suggests you are right... that said, by the time user code within a PEX is running, it's always normal files on disk; the unzipping happens as part of PEX start-up

Hmm okay where is where I'm at. I have this function

def __get_resource_dir_path(package: str, directory: str, file: str = None) -> str:
    """Get a posix path to a resource directory or file within the directory."""
    directories = directory.split('/')
    resource = importlib.resources.files(package)

    path = resource.joinpath(*directories)

    if file:
        path = path.joinpath(file)
        if not path.exists():
            raise FileNotFoundError(f"The file '{file}' does not exist in the directory path '{directory}'")
    else:
        if not path.is_dir():
            print(resource.joinpath())
            raise NotADirectoryError(f"The directory path '{directory}' does not exist")

    return path.as_posix()

and this BUILD file

python_sources(
    dependencies=[
        "//frontend:python-sources",
    ]
)

pex_binary(
    name="binary",
    entry_point="main.py",
)

pants run path/to/main.py

-> works

pants run path/to:binary

-> throws a NotADirectoryError 🤔 How does run

path/to/module.py

differ from run

path/to:binary

?