feat: add AliyunMilvus DISKANN support#795
Open
frankleaf wants to merge 6 commits into
Open
Conversation
…ch params Add an AliyunMilvus client (modeled on the VolcanoMilvus DISKANN client) that exposes DISKANN build knobs and three search-time knobs: rerank_topk_multiplier, early_termination_threshold and cross_segment_rerank. The three search params are optional and only injected into the per-query search params when explicitly specified (None = not sent, server keeps its own default). 0 is preserved as a meaningful value; a negative/"DEFAULT" sentinel from the UI is normalized to None. Wires up CLI command `aliyunmilvusdiskann`, the DB registry, and the Streamlit web UI (dbCaseConfigs + CaseConfigParamType entries). Co-authored-by: Cursor <cursoragent@cursor.com>
…arams Drop the Volcano-style scaffolding that is not needed for this use case: remove the dedicated AliyunMilvus client class and db config, and all the extra index-build knobs (max_degree/store_strategy/quant_type/ distance_strategy/enable_thp/enable_prefetch/build_search_list/legacy). AliyunMilvus now reuses the upstream Milvus client and MilvusConfig as-is; the only difference is AliyunMilvusDISKANNConfig, which extends the existing DISKANNConfig with the three opt-in search params and injects them into the search params only when explicitly specified. Co-authored-by: Cursor <cursoragent@cursor.com>
…init__ - Remove the empty aliyun_milvus/__init__.py: like milvus/ (and most client dirs) it is an implicit namespace package; the marker was unnecessary. - Wrap the case config with _with_partition_key, mirroring the MilvusDISKANN command, so AliyunMilvus is literally "Milvus DISKANN + 3 opt-in search params" with identical behavior otherwise. Co-authored-by: Cursor <cursoragent@cursor.com>
- Add tests/test_aliyun_milvus.py covering: reuse of the upstream Milvus client/config, opt-in search params (omitted by default, injected only when set), 0 treated as a real value, UI sentinel (-1 / "DEFAULT") -> unset, frontend exposure in Performance (not Load), and DISKANN-only case mapping. - README: add aliyunmilvus to the default (pymilvus) client row. Co-authored-by: Cursor <cursoragent@cursor.com>
Add parametrized cases for cross_segment_rerank string/int coercion so config.py reaches 100% coverage. Tidy README default-client wording. Co-authored-by: Cursor <cursoragent@cursor.com>
Explicit --no-cross-segment-rerank must send cross_segment_rerank=False (disable), not omit it. Omission is the only "unset" path. Co-authored-by: Cursor <cursoragent@cursor.com>
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: frankleaf The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
Author
|
/assign @XuanYang-cn |
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
2 participants
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.
Summary
Add AliyunMilvus as a new VectorDBBench target and only extends the existing Milvus DISKANN config with three opt-in search-time parameters.
What's added
New AliyunMilvusDiskANN type. It reuses the upstream Milvus client and MilvusConfig unchanged (no dedicated client class, no dedicated DB config).
New CLI command: vectordbbench aliyunmilvusdiskann ...
AliyunMilvusDISKANNConfig extends the existing Milvus DISKANNConfig and adds three opt-in search-time params:
Supported index
Only DISKANN is exposed for AliyunMilvus. Other index types are not silently mapped to upstream Milvus configs.
Example
Omitting any of --rerank-topk-multiplier / --early-termination-threshold / --cross-segment-rerank (or --no-cross-segment-rerank) means that param is not sent, so the run behaves exactly like milvusdiskann.