Skip to content

feat(vector-search): migrate to databricks-ai-search SDK, add filtering reference#197

Open
jiayi-wu-3150 wants to merge 1 commit into
databricks:mainfrom
jiayi-wu-3150:update/ai-search-sdk-migration
Open

feat(vector-search): migrate to databricks-ai-search SDK, add filtering reference#197
jiayi-wu-3150 wants to merge 1 commit into
databricks:mainfrom
jiayi-wu-3150:update/ai-search-sdk-migration

Conversation

@jiayi-wu-3150

Copy link
Copy Markdown

Summary

  • Migrate all Python SDK code from the deprecated WorkspaceClient.vector_search_* / VectorSearchClient APIs to AISearchClient from the databricks-ai-search package (SDK migration reference)
  • Add references/filtering.md — new comprehensive filter operator reference for both Standard (dict syntax) and Storage-Optimized (SQL string syntax) endpoints
  • Add DatabricksReranker section to search-modes.md (~10% quality gain, ~1.5s overhead)
  • Add Performance & Capacity section to troubleshooting-and-operations.md with SLA targets, sizing guidance, OAuth vs PAT performance, and debug_level=1 latency breakdown
  • Update product name references to "AI Search (formerly Vector Search)" throughout

Files changed

File Change
SKILL.md AISearchClient flat-param API; unified filters= parameter; Installation section; filtering.md reference added
references/index-types.md create_delta_sync_index/create_direct_access_index flat params; index.upsert(list)/index.delete(primary_keys) replacing inputs_json/schema_json
references/search-modes.md index.similarity_search(); lowercase query_type values ("hybrid" not "HYBRID"); filters_json removed; DatabricksReranker section added
references/troubleshooting-and-operations.md client.get_endpoint/get_index, index.sync/describe; Performance & Capacity section; 429 error row added
references/filtering.md New file — full operator reference for Standard (dict) and Storage-Optimized (SQL string)

Notes

  • CLI commands (databricks vector-search-endpoints, databricks vector-search-indexes) are unchanged — official CLI docs have not been updated yet
  • references/end-to-end-rag.md is not changed in this PR (CLI convention question pending)
  • Skill folder rename (databricks-vector-searchdatabricks-ai-search) is intentionally excluded — that requires metaplugin/plugin.meta.json updates and manifest regeneration, and will be filed as a separate PR

Test plan

This pull request and its description were written by Isaac.

…ng reference

Replace deprecated databricks-sdk WorkspaceClient vector_search_* APIs and
databricks-vectorsearch VectorSearchClient with the current AISearchClient
from databricks-ai-search across all skill files.

Key changes:
- SKILL.md: AISearchClient flat-param API, unified filters= parameter, add
  Installation section and filtering.md reference
- references/index-types.md: flat create_delta_sync_index/create_direct_access_index,
  index.upsert(list)/index.delete(primary_keys) instead of inputs_json/schema_json
- references/search-modes.md: index.similarity_search(), lowercase query_type
  values ("hybrid" not "HYBRID"), drop filters_json, add DatabricksReranker section
- references/troubleshooting-and-operations.md: client.get_endpoint/get_index,
  index.sync/describe, add Performance & Capacity section with SLA targets and
  debug_level guidance
- references/filtering.md: new file with full operator reference for both
  Standard (dict) and Storage-Optimized (SQL string) filter syntax

Co-authored-by: Isaac
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant