How to retrieve documents with a specific source_id from ElasticsearchDocumentStore

[hodge-jai]

Hello,

I am creating a RAG application and was wondering if there was a simple way to pull all the documents from my ElasticsearchDocumentStore instance that have either a specific title or specific source_id. I'm not looking to use it as part of a pipeline, I simply want to retrieve the documents that fit the criteria. I've used a document splitter as part of my indexing pipeline, and would like to allow users a way to see the rest of the original document that was split.

I've tried both using the filter_documents method that is part of the DocumentStore class, as well as the FilterRetriever retriever, but I'm not sure if my query syntax for the filters parameter is correct, because I keep getting an empty array as a return value, implying that it can't find any documents that match the source_id or title that I've supplied even though I know for a fact that the documents exist. I've used various filters values such as {"field": "title", "operator": "==", "value": title} and {"field": "title", "operator": "in", "value": [list_of_titles]} to no avail. A simple {"title": title} filter doesn't not work, either.

I'm sure I'm making a simple mistake here, so any guidance would be appreciated!

[Amnah199]

Hi @hodge-jai. Can you share the code example thats causing this issue?

[hodge-jai]

I ended up modifying the filter to use the source_id instead of the title field to make it work. For some reason it was unable to find the documents with the title I was passing in.