Skip to content

awsopensearch add normalize for cosine similarity when indexing data #614

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

Closed
norrishuang wants to merge 2 commits into from
Closed

awsopensearch add normalize for cosine similarity when indexing data #614

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d5f3a88
awsopensearch add normalize for cosine similarity, solved isuue of re…
norrishuang Oct 15, 2025
7328bdb
fix format issue
norrishuang Oct 15, 2025
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
6 changes: 6 additions & 0 deletions vectordb_bench/backend/clients/aws_opensearch/aws_opensearch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,6 +64,12 @@ def __init__(
self._update_ef_search_before_search(client)
self._load_graphs_to_memory(client)

def need_normalize_cosine(self) -> bool:
if self.case_config.metric_type.upper() == "COSINE":
log.info("cosine dataset need normalize.")
return True
return False

@alwayslove2013 alwayslove2013 Oct 16, 2025

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

The need_normalize_cosine function is designed for databases that do NOT natively support cosine similarity metrics. When testing cosine-based datasets, this function normalizes all training and query vectors before tests, to make the metric equivalent to L2 (Euclidean) or IP (Inner Product), ensuring accurate recall calculations. This is a workaround solution when cosine similarity is not directly supported.

For databases that already support cosine similarity, we do not recommend enabling this feature. The goal is to maintain consistent benchmarking across different databases and ensure fair testing conditions.


def _create_index(self, client: OpenSearch) -> None:
ef_search_value = self.case_config.ef_search
log.info(f"Creating index with ef_search: {ef_search_value}")
Expand Down