feat(clients): add Telys connector (on-device, Algenta-powered vector engine)#823
Closed
angelatgithub wants to merge 4 commits into
Closed
feat(clients): add Telys connector (on-device, Algenta-powered vector engine)#823angelatgithub wants to merge 4 commits into
angelatgithub wants to merge 4 commits into
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: angelatgithub 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 |
Telys (github via PyPI `telys`) is an on-device, Algenta-powered vector-execution
layer that accelerates FILTERED search by physical layout: the partition key becomes
a contiguous slice, so a `label == X` filter reads one block instead of scanning +
masking the whole table.
- backend/clients/telys/{telys.py,config.py,__init__.py}: VectorDB connector over a
running `telys serve` (TCP) via the public telys.client (RemoteTelys). Partitions by
the label field, so the label-filter case (StrEqual) maps to a single contiguous
partition slice; the no-filter case uses one constant partition. Supports NonFilter
+ StrEqual; thread_safe=False (each concurrent runner gets its own connection);
need_normalize_cosine=True (raw collections score by IP == cosine when normalized).
- register Telys in the DB enum + init_cls/config_cls/case_config_cls.
Requires telys >= (PR: remote build_ivf/drop_collection) for optimize()/drop_old.
- clients/telys/cli.py: `vectordbbench telys` command (host/port/access-token/ min-rows/target-recall), registered in cli/vectordbbench.py. - pyproject: `telys` optional-dependency extra (pip install vectordb-bench[telys]). - README: add telys to the supported-clients table. The connector runs over a `telys serve` TCP endpoint; label==key equality maps to Telys's single-key contiguous-partition wedge. Validated end-to-end through the real CaseRunner on Cohere-100K label-filter (recall 1.0, ndcg 1.0).
…build_ivf Reviewer simulation with the published telys 0.1.0b2 surfaced that drop_old cannot reset state without remote drop_collection, causing a schema conflict when a case re-runs or switches partition schema. drop_collection (and build_ivf) ship in the 0.1.0b3 self-host hardening, so pin the extra to it. build_ivf is additionally wrapped so an older server degrades to exact search (still correct) rather than erroring.
Pre-PR review follow-ups: - Reject L2/Euclidean datasets in __init__ (Telys ranks by IP; L2 would silently give wrong recall) instead of the previous silent mis-scoring. - Fix ruff (drop redundant noqa, sort imports) + black formatting -> CI lint passes. - Register DB.Telys in frontend styles (DB_TO_ICON + COLOR_MAP) so the Run-Test UI shows an icon instead of a broken image.
angelatgithub
force-pushed
the
feat/telys-connector
branch
from
July 21, 2026 01:49
1a55487 to
dfd2100
Compare
Author
|
Closing to resubmit cleanly with correct commit authorship (angelatgithub). Reopening shortly. |
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.
What
Adds a VectorDBBench client for Telys — an on-device, Algenta-powered vector-execution layer that accelerates filtered search by physical layout: the partition key is stored as a contiguous slice, so
label == Xreads one block instead of scanning + masking the whole table. The connector talks to a runningtelys serve(TCP) via the publictelysSDK, so it's a fair server-vs-server measurement. It partitions by the label field, so the label-filter case (StrEqual) maps to Telys's single-key contiguous-partition wedge.Changes
backend/clients/telys/{telys.py,config.py,cli.py,__init__.py}—VectorDBconnector + config + CLIbackend/clients/__init__.py(DB enum +init_cls/config_cls/case_config_cls) andcli/vectordbbench.pypyproject.toml:telysextra (pip install vectordb-bench[telys],telys>=0.1.0b3); README supported-clients table updatedSupported filters
NonFilterandStrEqual(label==value — the wedge)NumGE(int-range) is intentionally not declared — TelysEqis equality, not a rangeneed_normalize_cosine=True)Validation
Run end-to-end through the real
CaseRunneragainst a livetelys serve:Run
Notes for maintainers
zilliz.com/vdbbench-leaderboardset?