Add effective_search_ratio to vectorstore

[tomasonjo]

effective_search_ratio parameter allows you to improve query accuracy by increasing the potential candidate pool from lucene index and then filtering back down to k with Cypher

[tomasonjo]

Vector index returns an ordered list already V pet., 6. dec. 2024, 11:58 je oseba Alex Thomas ***@***.***> napisala:

…

***@***.**** commented on this pull request. ------------------------------ In libs/neo4j/langchain_neo4j/vectorstores/neo4j_vector.py <#18 (comment)> : > "YIELD node, score " + "WITH node, score LIMIT $k " Is an ORDER BY not needed here to ensure the nodes with the top k highest scores are returned? — Reply to this email directly, view it on GitHub <#18 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEYGGTNYJCCJ5Y7DW5TTBZ32EF7NJAVCNFSM6AAAAABTEF4EJCVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDIOBUGM4DIOJTHE> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[alexthomas93]

LGTM @tomasonjo I think the unit tests just need updating. Could you also update the CHANGELOG with the comment below?

[alexthomas93]

@CodiumAI-Agent /update_changelog

[CodiumAI-Agent]

Changelog updates: 🔄

Added

• Introduced effective_search_ratio parameter in Neo4jVector to enhance query accuracy by adjusting the candidate pool size during similarity searches.

to commit the new content to the CHANGELOG.md file, please type:
'/update_changelog --pr_update_changelog.push_changelog_changes=true'

[CodiumAI-Agent]

Changelog updates: 🔄

Next

Added

• Enhanced similarity search functionality in Neo4jVector with the effective_search_ratio parameter, allowing improved control over query accuracy and performance.

to commit the new content to the CHANGELOG.md file, please type:
'/update_changelog --pr_update_changelog.push_changelog_changes=true'