Questions regarding indexing documents such as pdf or word etc

[AsuraKev]

Hi there,

Based on the documentation site, in order to index a document we must first divide it into chunks and index individual chunk. We can then distinctively query them so that it returns as a search result.

My question is: Do we need to rebuild a pdf doc based on the index chunks? or should we be storing the actual file somewhere like Azure blob storage and index the file link url against each chunk? This way the purpose of the indexed chunk is to allow search by document content

thanks