Misleading API function names download_file and upload_file

[thomasborgen]

Is your feature request related to a problem? Please describe.

download_file and upload_file function names are misleading.

The problem with download_file is that while it does consume a storage bucket blob, it does not write to a file, it just returns bytes.
The problem with upload_file is that from the name one would expect the input to the function to be a filename, but it is actually bytes.

Describe the solution you'd like

Clearer api.

upload - input data: str | bytes
upload_from_file - input filename: str

download - returns str | bytes
download_to_file - return filename: str or absolute path to the file.

or have string and bytes spesific functions as well. but that might be a bit overkill.

[thomasborgen]

Maybe its not that straight forward.
We decided not to call the blobs in storage bucket for blobs, but for files. Which kind of led to the names upload_file and download file.

Not a 100% sure how to move forward. My issue when i looked into changing it was this:

def download(storage_bucket_name, filename): ...

def download_to_file(storage_bucket_name, filename, local_filename): ...

The argument naming gets a bit weird 😓

So then maybe we should change the argument names aswell. But if we do that we should create a consistent naming standard and we should probably just follow what google does... Calling a blob a blob is probably just the easiest way forward where a blob is data in a storage bucket and a file is data stored in a conventional way.