[ML] Adding docs for the unified inference API #118696
Conversation
jonathan-buttner
added
>docs
|
Documentation preview: |
| @@ -63,4 +63,44 @@ Specifies the chunking strategy. | |||
| It could be either `sentence` or `word`. | |||
| end::chunking-settings-strategy[] | |||
|
|
|||
| tag::unified-schema-content-with-examples[] | |||
There was a problem hiding this comment.
I import this a couple times for each of the messages since they all have the same format (string or an array of objects).
| The text content. | ||
| + | ||
| Object representation::: | ||
| `text`:::: |
There was a problem hiding this comment.
We get lucky that each of the messages has content:: so the colons work out here to nest it correctly.
| (Optional, array of objects) | ||
| A list of tools that the model can call. | ||
| + | ||
| .Structure |
There was a problem hiding this comment.
I'm open to other names here? Maybe Format?
| (Required unless `tool_calls` is specified, string or array of objects) | ||
| The contents of the message. | ||
| + | ||
| include::inference-shared.asciidoc[tag=unified-schema-content-with-examples] |
There was a problem hiding this comment.
I personally find it helpful if examples are sprinkled throughout the docs as we're reading instead of just include complete ones at the bottom. I'm open to other suggestions here.
There was a problem hiding this comment.
This is probably a larger issue than this PR, but theres a bug with the example expansion where the arrow always points down 
, but only for the inner expansion box, the outer one is correc.t
There was a problem hiding this comment.
Thanks for pointing that out. I made the docs team aware.
| ------------------------------------------------------------ | ||
| // TEST[skip:TBD] | ||
|
|
||
| <1> Each tool call needs a corresponding Tool message. |
There was a problem hiding this comment.
Maybe this is too OpenAI specific to include?
|
@elasticmachine run elasticsearch-ci/docs |
|
Pinging @elastic/ml-core (Team:ML) |
|
Pinging @elastic/es-docs (Team:Docs) |
|
@elasticmachine run elasticsearch-ci/docs |
jonathan-buttner
added
the
auto-backport
| ==== {api-request-body-title} | ||
|
|
||
| `messages`:: | ||
| (Required, array of objects) A list of objects representing the conversation. |
There was a problem hiding this comment.
What do you think about adding some information about how these messages lists should be generated? Something like
"Requests should generally only add new messages from the user. The other messages ("assistent", "system", or "tool") should generally only be copy-pasted from the response to a previous completion request, such that the messages array is built up over the course of a conversation."
|
|
||
| `model`:: | ||
| (Optional, string) | ||
| The ID of the model to use. |
There was a problem hiding this comment.
| The ID of the model to use. | |
| The ID of the model to use. By default, the model ID set in the inference endpoint is used. |
This PR adds docs for the new unified inference API. I tried to cover the more complicated objects with examples.
https://elasticsearch_bk_118696.docs-preview.app.elstc.co/guide/en/elasticsearch/reference/master/unified-inference-api.html