Integration of azure vector search with azure document intelligence

Great work with integrated data chunking and embedding.
Asking from a RAG context (complex documents, e.g. financial statement) , great work with integrated data chunking and embedding, how does the current framework works with the skills in document intelligence where you have to convert financial tables into json and feeds into LLM, given current framework is using chunk id as the key. Or team is expect this to resolve with prompt flow and sk or other methods? Any sample or best practise?

No file or test is named. Start by reviewing the repository's Azure vector-search samples and compare their chunk-id approach with the requested Azure Document Intelligence table-to-JSON workflow; done would be a documented sample or best-practice answer that explains how the pieces work together.

Written by the indexing model from the issue text.