scikit-bio icon indicating copy to clipboard operation
scikit-bio copied to clipboard

[IO Registry] Limit set of formats to sniff by the intended constructor.

Open ebolyen opened this issue 3 years ago • 0 comments

Noticed by @sjanssen2 in #1825, the result of not providing a format to the skbio.TreeNode.read() resulted in substantially bad performance by the CSV sniffer.

Since there is not a reader in the registry which can convert CSV -> skbio.TreeNode, this sniffer should probably have been skipped.

It would also be ideal to cap the number of bytes any given sniffer actually reads, as a large tree is obviously not the first cell of a CSV.

ebolyen avatar Dec 08 '22 18:12 ebolyen