Upgrade to v0.4 breaks Jupyter Notebooks use cases

Code that worked inside of a Jupyter Notebook in v0.3 of MSGraphFS no longer works in v0.4. Here is an example of code that runs without error when executed in a notebook cell in v0.3 but not in v0.4. The code runs successfully in v0.4 when executed as a script, so it seems to be the fact that Jupyter provides its own event loop that is tripping it up.

import os
import msgraphfs
from dsd_sharepointfs import get_access_token_interactive

token = get_access_token_interactive() # retrieves access token via public client flow

client_id = os.getenv("DSD_SHAREPOINT_FS_CLIENT_ID")
tenant_id = os.getenv("DSD_AZURE_TENANT_ID")
drive_id = "b!oc7uSdROqk-OOYAKmTu8MmOA9AXS2kZDs59EF92bPvUkJSHmL6ZzQJDueJUxFuGV"

fs = msgraphfs.MSGDriveFS(
    client_id=client_id,
    drive_id=drive_id,
    oauth2_client_params = {
        "token":token,
    }
)

print(fs.ls("/", detail=False)[:5])

Running inside a notebook cell with v0.4 gives the error

NotImplementedError: Calling sync() from within a running loop.

The same error is obtained when calling fs.open(). The error persists even if I pass in asyncio.get_event_loop() for the loop parameter of the MSGDriveFS constructor.

Reproduce the v0.4 failure in a Jupyter notebook using the MSGDriveFS example, especially fs.ls() and fs.open(), then compare it with running the same code as a script. Start by tracing the asyncio event-loop handling around the sync() call; done means these operations work inside a notebook without the NotImplementedError.