Skip to content

feat: add query_transformer parameter to BaseRAGQuestionAnswerer #209

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
srishti-1935 wants to merge 1 commit into
base: main
Choose a base branch
from
+11 −1
Open

feat: add query_transformer parameter to BaseRAGQuestionAnswerer #209

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 11 additions & 1 deletion python/pathway/xpacks/llm/question_answering.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -462,6 +462,11 @@ class BaseRAGQuestionAnswerer(SummaryQuestionAnswerer):
search_topk: Top k parameter for the retrieval. Adjusts number of chunks in the context.
rerank_topk: Number of top-scoring documents to retain after reranking, when a reranker is provided.
If ``None``, reranking is disabled. Defaults to ``None``.
query_transformer: Transformer for query rewriting before retrieval.
Can be a ``pw.UDF``, callable, or ``None`` to skip query transformation.
Available options: ``pathway.xpacks.llm.prompts.prompt_query_rewrite``,
``pathway.xpacks.llm.prompts.prompt_query_rewrite_hyde``.
Defaults to ``None``.


Example:
Expand Down Expand Up @@ -523,11 +528,13 @@ def __init__(
search_topk: int = 6,
reranker: pw.UDF | None = None,
rerank_topk: int | None = None,
query_transformer: pw.UDF | Callable | None = None,
) -> None:

self.llm = llm
self.indexer = indexer
self.reranker = reranker
self.query_transformer = query_transformer

if default_llm_name is None:
default_llm_name = llm.model
Expand Down Expand Up @@ -637,7 +644,10 @@ def add_score_to_doc(doc: pw.Json, score: float) -> dict:
@pw.table_transformer
def answer_query(self, pw_ai_queries: pw.Table) -> pw.Table:
"""Answer a question based on the available information."""

if self.query_transformer is not None:
pw_ai_queries += pw_ai_queries.select(
prompt=self.query_transformer(pw.this.prompt)
)
Comment on lines +647 to +650
Copy link
Copy Markdown
Contributor

@szymondudycz szymondudycz Apr 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

pathway.xpacks.llm.prompts.prompt_query_rewrite and `pathway.xpacks.llm.prompts.prompt_query_rewrite_hyde`` are just prompts to be sent to an llm to ask to rewrite a question for the sake of retrieval. Replacing a question to RAG with these prompt makes the RAG completely wrong

pw_ai_results = pw_ai_queries + self.indexer.retrieve_query(
pw_ai_queries.select(
metadata_filter=pw.this.filters,
Expand Down