Add support for zipped files

[nwlandry]

It would be nice to be able to read/write compressed files in the I/O methods, which is not currently possible.

[nwlandry]

So GZIP, ZIP, etc.

[kaiser-dan]

I think this could be managed simply through xopen. It supports .gz, .bz2, .xz, and .zst for both reading and writing. Conveniently, using it should be not much more than replacing all open keywords with xopen.xopen. I've tested it out on reading HIF files and it seems to work very well. The team behind it seems to be pretty on top of its development and updates as well.

If adding xopen as a dependency is ok I can get started on adding it to the I/O soon!

[maximelucas]

Good idea! Why not use xopen. If it's easy and only a few lines of code to code it ourselves, it might be worth not adding a dependency. Any idea how much that would be?

[kaiser-dan]

It should boil down to some conditional on the file format and dispatching the correct decompression library, e.g. gzip for data.hif.gz files. These libraries exist in Python's stdlib and these are what xopen uses.

The main advantage as I see it is simplifying the updated file read logic and the cost of a non-stdlib dependency, albeit effectively it's just a wrapper over several stdlib compression libraries.