Python's zipfile module supports writing empty folders to zip archive.

when trying with zipfly i get:

File "/.../zipfly.py", line 211, in generator
    with open( path[self.filesystem], 'rb' ) as e:
IsADirectoryError: [Errno 21] Is a directory: '/path/to/dir'

Is there an undocumented way for writing empty folders?

Hi,

I see that only uncompressed ZIP files are can be created with zipfly. What is the reason for that? I don't see any compression level specific code in ZipFly.generator, except file size estimation (which doesn't work for ZIP64 achives).

So, is compression just an unimplemented feature? Or are there any aspects which don't allow using compression for streaming made this way?

I understand that compression is often useless in streaming scenarios but it's not true in our case. We're looking for a replacement for zipstream-new which seems to be dead. ZipFly looks nice for us...

Running Django 4.1 in async ASGI mode (channels & websockets needed)

Zipfly functionability breaks as it tries to compile the entire zip to memory before sending zip to client, filling up memory and causing a crash

It looks like the last version on PyPI has been published 2 years ago and there have been many changes in the code since.

It would be helpful if you could release the latest changes.

PS since the project is already using GH action for testing, it may be helpful to set up GH actions for release as well.

https://packaging.python.org/en/latest/guides/publishing-package-distribution-releases-using-github-actions-ci-cd-workflows/

https://github.com/marketplace/actions/pypi-publish

Hi, thank you for the inspirational code.

I'm wondering if it's possible to adapt your code and make it compatible with S3 storage.

I use django-storages and boto3 client returns a streaming body (already open fp) and all metadata.

I need a zipfile (or any other archive) to be created from the set of files during the GET request (not great but it has to be).

So to save the memory I have 2 options.

1. Use zipfly as is and download files into a temporary location (and remove it after the operation)
2. Or better solution that doesn't require intermediate storage so I could pipe the content of streaming body into zipfly and return that as a streaming response.

streaming body (many of them) -> zipfly(zip file) -> streaming response

Do you think that option two is possible? If so, could you please point out what pieces of code I should focus on to adapt zipfly?

Thank you

It is a common use case to store large files remotely and not next to the code. See e.g. for django https://github.com/jschneier/django-storages. If many (large) files must be shipped synchronuosly as a zip file, it saves memory and storage to pass them through the web worker as a stream without saving anything to disk. To archive that, file buffers as input may be supported.

I would appreciate if we could add this functionality. This pull request contains a tested initial attempt. Feel free to improve!

This pull request is based on https://github.com/BuzonIO/zipfly/pull/75 and https://github.com/BuzonIO/zipfly/pull/76

Hi, thanks for the great work. I have some bytes generators and I want to make a zip on the fly out of them. Is there an option to do this? Sorry for my poor python knowledge. Thanks :)