Updated notebooks and utils.py to use the latest stable Azure Search API Version, needed for future Index projection and automatic embedding

[maurominella]

I've just updated the notebooks 1, 2, 3, 4 + utils.py to use ⁠the latest Azure Search API version 2023-11-01. I spent a few hours to check all notebooks 1 to 11 work without any problems, including backend and frontend apps. This way we'll be able to create a "notebook 3.5" that creates, in a single indexer run, the index including the chunks projection (in order to have 1 search document per chunk), the chunk embedding and the entities -URLs, emails, keyphrases- associated to that chunk.

[maurominella]

I added the 035-ChunkDocumentsInSingleStep.ipynb file to fill secondary index in a single step and to automatically create embeddings leveraging a skill. Now the indexer fills only one indexer with a single chunk per row (aka search document) together with its embeddings, its key phrases and some entities like persons/urls/emails.
Furthermore, the search is now able to pass raw vectors (externally calculated), but we can also pass simple text that is automatically converted to embeddings and used for the similarity search.
The last cell, finally, shows how to interrogate Azure AI Search TRHOUGH AZURE OPENAI leveraging the COMPLETION EXTENSIONS: a single call TO AZURE OPENAI implicitly calls Azure AI Search.

Please note that this notebook is just for demonostration purposes and has no effect with the rest of the solution accelerator, that keeps using the two indexes created in notebooks #1 / #2 / #3.