VirtualiZarr icon indicating copy to clipboard operation
VirtualiZarr copied to clipboard

Virtualize ZIP archives?

Open TomNicholas opened this issue 6 months ago • 1 comments

This article is fascinating:

ZIPs facilitate random access and cheap updates as part of their design.

Random access is achieved by providing a central directory, which contains a listing of all files contained in the archive (as well as their metadata and offsets). Once the central directory has been read, entry offsets can be used to seek to the portion of the file that contains a file of interest.

This very much makes it sound like we should be able to virtualize anything that was zipped! (As long as it was only zipped once - I've heard of double-zipped archives and that's insanity).

I'm not really sure what the API for this would look like though. IIUC ZIP can be applied to anything, so perhaps virtualizarr would need some kind of "Unzipper" that could be composed with any Parser?

TomNicholas avatar May 26 '25 03:05 TomNicholas

yml SOzip: https://github.com/sozip/sozip-spec

mdsumner avatar May 26 '25 04:05 mdsumner