Using indexProjections in data ingestion and improve logging for large files

Using indexProjections now (just one index, indexer and datasource instead of two).
Setting SEARCH API to 2023-10-01-Preview to support indexProjections and support Managed Identity authResourceId parameter.
Changed chunk funcion timeout to 3m50s for alignment with AI Search web API skill max timeout.
Changed chunk_documents_formrec to chunk_documents_docint.
Improved function's logging to help in cases of large files.
Improved documentation.

Azure / gpt-rag-ingestion