Closed lll-lll-lll-lll closed 1 year ago
Hi @lll-lll-lll-lll. Thanks so much for your work on this library.
I recently implemented a download helper method for predictions in this PR to the replicate
CLI.
I used errgroup
there to parallelize the downloads, but other cases may have different requirements that warrant a different approach.
Although it could be helpful for this library to provide a built-in Download
method, I'm not sure there's a way to do it that doesn't bake in a bunch of assumptions that wouldn't hold for some use cases. Whether to download in parallel, whether to download into a directory or a Zip archive, what to do for outputs that aren't an array of URLs... I think there are probably too many decisions than can fit neatly into a single API.
Instead of providing a Download
method, what do you think about providing some example code in the docs?
@mattt thanks for your PR of cli.
That is certainly true. This Download
method may be trying to handle too many different use cases.
That's a good idea! Can I create some example code with client and server? I would like to make the code as simple as possible.
Thanks for showing me how to do it !
I created #14 but close PR. because If an error occurs during the process, another thread will continue execution and execute the http.Get method.
now, I am trying to add the errrgroup pkg and write a process that will stop all threads if an error occurs. Is it ok to create a Download method with the errrgroup pkg added?