Mismatch issue of Documents shown and Citation number in text fix #3421
Conversation
This file contains 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
When document order presented to LLM differs from order shown to user, wrong doc numbers are cited. Fix: - SearchTool.get_search_result returns now final and initial ranking - initial ranking is passed through a few objects and used for replacement in citation processing Notes: - the citation_num in the CitationInfo() object has not been changed.
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
Updated: - test_citation_substitution (changed description) - test_citation_processing (removed data only relevant for the substitution)
yuhongsun96
reviewed
Dec 11, 2024
| self, context_docs: list[LlmDoc], doc_id_to_rank_map: DocumentIdOrderMapping | ||
| self, | ||
| context_docs: list[LlmDoc], | ||
| doc_id_to_rank_map: DocumentIdOrderMapping, |
There was a problem hiding this comment.
Prefer putting the display_doc_order_list within the DocumentIdOrderMapping
| @@ -224,6 +226,7 @@ def _get_response(self, llm_calls: list[LLMCall]) -> AnswerStream: | |||
| answer_handler = CitationResponseHandler( | |||
| context_docs=search_result, | |||
| doc_id_to_rank_map=map_document_id_order(search_result), | |||
There was a problem hiding this comment.
Would create the single object with both in this function here, so above we can do processed_search_result, displayed_search_results
| search_result = SearchTool.get_search_result(current_llm_call) or [] | ||
| search_result, displayed_search_results_map = SearchTool.get_search_result( | ||
| current_llm_call | ||
| ) or ([], {}) |
There was a problem hiding this comment.
Pref ([], []) for consistency and building the ordering in the call below similar to the other object. For reducing complexity for future readers
yuhongsun96
reviewed
Dec 11, 2024
| @@ -22,12 +22,16 @@ def __init__( | |||
| self, | |||
| context_docs: list[LlmDoc], | |||
| doc_id_to_rank_map: DocumentIdOrderMapping, | |||
| display_doc_order_dict: dict[str, int], | |||
| stop_stream: str | None = STOP_STREAM_PAT, | |||
| ): | |||
| self.context_docs = context_docs | |||
| self.doc_id_to_rank_map = doc_id_to_rank_map | |||
| self.stop_stream = stop_stream | |||
| self.order_mapping = doc_id_to_rank_map.order_mapping | |||
There was a problem hiding this comment.
This attribute would contain both that way we have a single cohesive "state object" we can pass to the function
hagen-danswer
approved these changes
Dec 11, 2024
This was referenced Dec 16, 2024
aronszanto
pushed a commit
to aronszanto/onyx
that referenced
this pull request
Dec 23, 2024
…yx-dot-app#3421) * Mismatch issue of Documents shown and Citation number in text fix When document order presented to LLM differs from order shown to user, wrong doc numbers are cited. Fix: - SearchTool.get_search_result returns now final and initial ranking - initial ranking is passed through a few objects and used for replacement in citation processing Notes: - the citation_num in the CitationInfo() object has not been changed. * PR fixes - linting - removed erroneous tab - added a substitution test case - adjusted original citation extraction use case * Included a key test and * Fixed extra spaces * Updated test documentation Updated: - test_citation_substitution (changed description) - test_citation_processing (removed data only relevant for the substitution)
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.
When document order presented to LLM differs from order shown to user, wrong doc numbers are cited.
Fix:
Notes:
Description
[Provide a brief description of the changes in this PR]
How Has This Been Tested?
[Describe the tests you ran to verify your changes]
Accepted Risk (provide if relevant)
N/A
Related Issue(s) (provide if relevant)
N/A
Mental Checklist:
Backporting (check the box to trigger backport action)
Note: You have to check that the action passes, otherwise resolve the conflicts manually and tag the patches.