I need to retrieve all the labeled utterances (aka examples) for a particular intent.
It appears the only call in the LUIS Authoring API to do this is GET review labeled examples:
/luis/api/v2.0/apps/:appId/versions/:versionId/examples?skip=0&take=500
/luis/api/v2.0/apps/:appId/versions/:versionId/examples?skip=500&take=500
This returns all utterances for all intents. I have 880 labeled utterances, and due to the verbose reply, it's a 2.5MB file. This makes it slow.
The LUIS web UI uses a filtered call:
/luis/webapi/v2.0/apps/:appId/versions/0.1/models/:modelId/reviewLabels
. The resulting file is usually 10-50kB. However, there is no documentation around this call (notice the webapi
in the path rather than just api
).
So: is there a supported method for retrieving a filtered list of utterances?
https://westus.api.cognitive.microsoft.com/luis/api/v2.0/apps/:appId/versions/:versionId/export
) produces a much smaller file, but it doesn't provide any ids. Without ids, it's useless from a programmatic perspective. – Chris McGrath/examples/
, each example has a unique numeric ID./export
doesn't provide that. – Chris McGrath/examples
in batches of 500 from this Programmatic API – Kunal Mukherjee