Standardize dspy.Retrieve on Document class #8683
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Closes #8678
Currently,
dspy.Retrievecoerces everyrmmodule to a simple list of passages, which is not ideal for every enterprise use-case. This PR formalizes the databricks Document dataclass as the standard for output from a retriever.I did my best with this PR to remain backwards compatible, making this return type opt-in by default (See dspy/retrievers/retrieve.py). I updated each of the built-in retriever classes to return a list of these documents without any additional metadata, as this is unused in the current day.
My goal here is to be able to add things like document score, source url, etc to the document metadata such that it can empower rich UIs through more information about the source data that was fetched and used during inference. This feels like a potentially heavy-handed approach, so I am happy to discuss alternatives.
Right now, the pain that I am feeling is that the current retrieve base class only returns the text with no additional information available about the documents themselves.