You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@geaflow.apache.org by "kitalkuyo-gita (via GitHub)" <gi...@apache.org> on 2026/08/24 03:13:34 UTC

[I] [geaflow/ai-memory] Add `EntityVectorIndex` wrapper around graph entity embeddings [geaflow]

kitalkuyo-gita opened a new issue, #853:
URL: https://github.com/apache/geaflow/issues/853

   Priority: P0  
   Difficulty: Intermediate
   
   Context: Graph entity embedding exists, but Phase 1 needs explicit entity-vector semantics for fuzzy entity anchoring.
   
   Scope:
   
   - Wrap current `EmbeddingIndexStore` behavior behind `EntityVectorIndex`.
   - Add entity ID, entity type, model metadata, and vector version.
   - Add compatibility tests with existing `EmbeddingOperator`.
   
   Constraints:
   
   - Do not change retrieval ranking yet.
   - Link existing embedding correctness fix issue/PR where applicable.
   
   Acceptance Criteria:
   
   - Entity vector lookup can distinguish vertex and edge entities.
   - Dimension/model mismatch is detected before search.
   
   Suggested paths:
   
   - `geaflow-ai/src/main/java/org/apache/geaflow/ai/index/entity`
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscribe@geaflow.apache.org.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@geaflow.apache.org
For additional commands, e-mail: issues-help@geaflow.apache.org


Re: [I] [geaflow/ai-memory] Add `EntityVectorIndex` wrapper around graph entity embeddings [geaflow]

Posted by "Mugetsu-bankai (via GitHub)" <gi...@apache.org>.
Mugetsu-bankai commented on issue #853:
URL: https://github.com/apache/geaflow/issues/853#issuecomment-5791569663

   Hi @kitalkuyo-gita, I’d like to take #853 (GM-AI-P1-019). My implementation plan is:
   1. Add EntityVectorIndex under geaflow-ai/.../index/entity, preserving the existing EmbeddingIndexStore and IndexStore#getEntityIndex(GraphEntity) behavior so EmbeddingOperator remains compatible.
   2. Introduce typed entity-vector records and lookup results containing a stable entity ID, vertex/edge kind, model metadata, vector version, and score. Vertex and edge IDs will have distinct, unambiguous encodings; multiple embeddings for one entity will remain identifiable.
   3. Align the new lookup/storage path with the VectorStore contract from #851 / PR #855. Validate model and dimension before searching, and restrict results to entity vectors. Legacy JSONL entries have no model metadata, so I’ll preserve their existing lookup behavior but require explicit metadata binding or migration before using them in the validated lookup path.
   4. Add offline tests with deterministic embeddings for vertex/edge distinction, ID collisions, model and dimension mismatches, multiple vectors per entity, and compatibility with the existing EmbeddingOperator. Its ranking and current global vertex-only behavior will remain unchanged in this issue.
   5. Reference the embedding correctness fix in #830 / PR #832 and rebase on the final #851 SPI contract before opening the PR.
   Please let me know if you prefer a different policy for legacy records without model metadata.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscribe@geaflow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@geaflow.apache.org
For additional commands, e-mail: issues-help@geaflow.apache.org