Simon Willison on Nostr: Case in point: https://llm.datasette.io has pretty comprehensive test coverage, but ...
Case in point:
https://llm.datasette.io has pretty comprehensive test coverage, but there's one obscure feature (the ability to run it against OpenAI-compatible endpoints via YAML configuration) which isn't easily tested... and that's the feature that's causing me to drag my heels on an upgrade of the core openai Python library dependency right now
Published at
2024-01-26 02:07:46Event JSON
{
"id": "9868d0a64180eaac61a73bd0201a29fc8f4a67ff6fcc30233749bcdd55cdc094",
"pubkey": "8b0be93ed69c30e9a68159fd384fd8308ce4bbf16c39e840e0803dcb6c08720e",
"created_at": 1706234866,
"kind": 1,
"tags": [
[
"e",
"e98d3aaed89ff078765eb23577730ed73880ee975dba044996b4aa26a7fde802",
"wss://relay.mostr.pub",
"reply"
],
[
"proxy",
"https://fedi.simonwillison.net/users/simon/statuses/111819808192832629",
"activitypub"
]
],
"content": "Case in point: https://llm.datasette.io has pretty comprehensive test coverage, but there's one obscure feature (the ability to run it against OpenAI-compatible endpoints via YAML configuration) which isn't easily tested... and that's the feature that's causing me to drag my heels on an upgrade of the core openai Python library dependency right now",
"sig": "dd929899c221f4f40c3ebf3b76f6a54a003c8585f948c60419124b1d9121ff51bdec92e65486cf153e6eaf4bd2912fd0d823020aee56c00bc725275ed1875ddc"
}
