Thank you so much for this! I was looking for some implementation of jszip and it works great!

Thanks! It's a pity that zip/gzip is not a native browser feature.

This notebook looks great, Fil! I have a question about the "string" here in the second-to-last line. I've used zip files with Arrow tables before, where what you really want out of the zip file is the binary data, rather than a string. There's a "type" option in the jszip docs – maybe it could be offered as an optional parameter to this function? https://stuk.github.io/jszip/documentation/api_zipobject/async.html#type-option

Sure! Let's create an {options} object. If the zip contains several files (say a text and an image), you'll want to load them all in binary then transform the text yourself? Also do you have a concrete example that we could link to?

Sure, here's a quick demo of loading a small Arrow table using this library: https://observablehq.com/d/13fed54c3689d975.

And good question re: heterogenous zip files... I'm not sure about the best answer there, other than simply returning files[f] and telling the user to call .async with the appropriate type. There's something nice about the concision of the current approach, though.

Yes, it's here to solve this recurring use case: to upload a bunch of csv or txt files and get them back easily. When I need the whole API I will `import {zip}` and read the docs. Next step: xlsx files (?!).

I added your example in this notebook. Thank you!

Thanks! :)

Something wrong here with the arrow file?

Thanks for the notice. Apparently apache-arrow@7 has a different API? I went back to @6 for now.