Add MemoryHub memory provider - #34
Open
rdwj wants to merge 3 commits into
Open
Conversation
Adapter: clean 130-line provider derived from our internal 503-line version. Strips ablation config to 4 env vars (URL, key, project, k). Also adds lab notes documenting reproduction protocol, a PR draft template, and the run-repro.sh convenience script. Assisted-by: Claude Code (Opus 4.6)
493/589 correct on PersonaMem 32k with Gemini 3.1 Pro Preview. Hybrid vector + keyword search with cross-encoder reranking. Assisted-by: Claude Code (Opus 4.6)
Assisted-by: Claude Code (Opus 4.6)
|
@rdwj is attempting to deploy a commit to the Vectorize Team on Vercel. A member of the Team first needs to authorize it. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #33
Summary
Adds MemoryHub as a memory provider for AMB. MemoryHub is a Kubernetes-native agent memory component for OpenShift AI. Apache 2.0 licensed, self-hostable.
Architecture: verbatim conversation storage in PostgreSQL + pgvector. Hybrid search via vector + keyword + BM25 with cross-encoder reranking and reciprocal-rank fusion. Embeddings are server-side (Granite via the MemoryHub pipeline).
Results -- PersonaMem 32k
Answer LLM:
gemini-3.1-pro-preview, judge:gemini-3.5-flash-lite.What's included
src/memory_bench/memory/memoryhub.py-- provider adapter (~130 lines). Connects via thememoryhubPython SDK over streamable-HTTP (MCP protocol).src/memory_bench/memory/__init__.pyand display entry incatalog.json.pyproject.toml--memoryhub>=0.15dependency.outputs/personamem/memoryhub/rag/32k.json.gz-- committed run results.results-manifest.json-- manifest entry.Reproduce
Requires a running MemoryHub instance. The
memoryhubSDK handles MCP session management, authentication, and search.