Sylvain Lesage

See https://github.com/hyparam/hyparquet/commit/6ec836dac57bfab5e096e3cde7aa16c46decbfc7

Parquet stream-conversion fails to embed images/audio files from gated repos

1

it seems to be an issue with datasets not passing the token to embed_table_storage when generating a dataset See https://github.com/huggingface/datasets-server/issues/2010

fixes #3639

In the README 'io.js' section, explain what io.js was?

2

The README 'io.js' section only refers to `https://github.com/iojs/io.js/` for people to understand what io.js is. But it is now redirected to `https://github.com/nodejs/node`, so it took me some time to find...

This issue groups the breaking changes that will be included in hyparquet v2. We invite the users and contributors to join the discussion below. We aim to improve the API...

I think we should pass [`Accept-Encoding`](https://developer.mozilla.org/en-US/docs/Web/HTTP/Reference/Headers/Accept-Encoding) with value: `identity` here: https://github.com/hyparam/hyparquet/blob/2a91b8a11853553e556b8ef3f98dc5a963c16281/src/utils.js#L113 The range requests work on the encoded bytes, ie: if the response is compressed with gzip, our ranges computation...

See https://github.com/hyparam/hyparquet/blob/cc6cc86ba41816a3a9e6d72723a84d1ad19c5f8c/src/convert.js#L11-L28

`byteLengthFromUrl` fallback with `GET` in case `HEAD` is not allowed

1

I got a case with a signed S3 URL that is only allowed for GET requests, and returns 403 on HEAD requests. An alternative is to send a GET request,...

The HTTP Range header lets you pass `bytes=-524288` to get the last 512K bytes of a remote resource. Instead of first fetching the bytelength with a HEAD request, then fetch...