[
https://issues.apache.org/jira/browse/TIKA-4896?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Tim Allison updated TIKA-4896:
------------------------------
Description:
Various vendors that support the OpenAI API also have their own keys for their
custom functionality. To the degree possible, we should make this as easy as
possible across vendors.
This ticket focuses on the main OpenAI API family vendors: JIna, Azure and
LiteLLM.
Direct calls to Gemini and Bedrock's Nova 2 would require their own custom
engines, which we can do if there's demand, on a separate ticket.
For now, users can get _some_ functionality from both via LiteLLM.
For fans of technical debt... start your watches and stay tuned for how soon
this PR needs to be updated!!!
was:
Various vendors that support the OpenAI API also have their own keys for their
custom functionality. To the degree possible, we should make this as easy as
possible across vendors.
This ticket focuses on the main OpenAI API family vendors: JIna, Azure and
LiteLLM.
Gemini and Bedrock's Nova 2 would require their own custom engines, which we
can do if there's demand, on a separate ticket.
> Add request parameters for openai family endpoints
> --------------------------------------------------
>
> Key: TIKA-4896
> URL: https://issues.apache.org/jira/browse/TIKA-4896
> Project: Tika
> Issue Type: Task
> Reporter: Tim Allison
> Priority: Minor
>
> Various vendors that support the OpenAI API also have their own keys for
> their custom functionality. To the degree possible, we should make this as
> easy as possible across vendors.
> This ticket focuses on the main OpenAI API family vendors: JIna, Azure and
> LiteLLM.
>
> Direct calls to Gemini and Bedrock's Nova 2 would require their own custom
> engines, which we can do if there's demand, on a separate ticket.
>
> For now, users can get _some_ functionality from both via LiteLLM.
>
> For fans of technical debt... start your watches and stay tuned for how soon
> this PR needs to be updated!!!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
