community: batch request LocalAIEmbeddings

[mkhludnev]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

OpenAI API /embeddings support batch request of an array of texts.
However LocalAIEmbeddings loops texts and request them one by one.

Motivation

It's deadly slow.

Proposal (If applicable)

Send batch request.

[mkhludnev]

see #22666