Fix client usage of the Inference API

[pquentin]

I don't have an OpenAPI, so I can't actually test this. I would love help here.

[gitnotebooks]

Found 4 changed notebooks. Review the changes at https://app.gitnotebooks.com/elastic/elasticsearch-labs/pull/323

[pquentin]

I now have an API key, and need to test the modified notebooks before merging.

[pquentin]

I can confirm that the search notebooks all run fine now. I have not touched the other notebooks that use the Inference API.

This includes the Cohere integration (cc @szabosteve @maxhniebergall)

notebooks/integrations/cohere/cohere-elasticsearch.ipynb
notebooks/integrations/cohere/inference-cohere.ipynb
notebooks/integrations/cohere/updated-cohere-elasticsearch-inference-api.ipynb

And those blogs (cc @lio-p @llermaly):

supporting-blog-content/customer-success-example/ingest.py
supporting-blog-content/building-multilingual-rag-with-elastic-and-mistral/building_multilingual_rag_with_elastic_and_mistral.ipynb

[joemcelroy]

great! thanks for that. Merge when you're ready.