Parent Doc Retriever Limited to In Memory Store only .

[jaideep11061982]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

Parent document retriever seems unusable in case we have large no of document to be embedded for recurrent query .
Multiretriever gets document for child doc from Memory . If suppose i have 50000 documents then I cant use In Memory store to store the docs . I need to first vectorize it .

Motivation

This is a useful feature needs to be made scalable so it can meet Enterprise Doc store needs.

Proposal (If applicable)

POssible approach would be update the text meta data with parent content directly.