fix: make QdrantLocal runnable from the web UI

[serhiizghama]

Fixes #796. QdrantLocal can't be run from the web UI at all — QdrantLocalIndexConfig requires m and ef_construct, but there's no DB.QdrantLocal entry in CASE_CONFIG_MAP, so the UI renders "Auto", passes an empty config, and pydantic raises 2 validation errors ... Field required before the run starts. That matches the reporter's "seems like 'ef_construct' is missing, but there is not such option" — the option genuinely isn't there.

Gave both fields the same defaults the CLI already uses (m=16, ef_construct=200) and registered QdrantLocal in CASE_CONFIG_MAP so m, ef_construct and on_disk are tunable for load, plus hnsw_ef for search — the same four knobs qdrant_local/cli.py exposes. Needed two new CaseConfigParamType members since the existing ef_construction spelling doesn't match the model field.

The CLI path is unchanged: it always passed these explicitly. Verified the empty-config case now builds, that the UI defaults round-trip into index_param()/search_param(), and that tuned values reach the client. New tests fail on main and pass here.

[sre-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: serhiizghama
To complete the pull request process, please assign xuanyang-cn after the PR has been reviewed.
You can assign the PR to them by writing /assign @xuanyang-cn in a comment when ready.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[yhmo]

vectordb_bench/frontend/config/dbCaseConfigs.py line:3299
Low ---- The new m input lets users set m=0 with the help text "set 0 to disable the index", but no such code path exists: QdrantLocalIndexConfig.index_param() always forwards m into HnswConfigDiff(m=...), and Qdrant does not validate m (only ef_construct is constrained to min=4), so create_collection accepts m=0 and builds an HNSW graph with m0 = 2*m = 0 edges — a degenerate index that silently yields broken search results rather than a disabled index. Consider setting min=2, rewording/removing the "set 0 to disable the index" claim, or implementing an actual no-index path.

[serhiizghama]

You're right — that help text was inherited from the local CLI's --m option, and it doesn't hold here. m=0 is Qdrant's multitenancy knob: it disables the global index only when payload_m builds per-tenant ones instead, which qdrant_cloud sets but qdrant_local never does. So in this client 0 just leaves the collection without a vector index.

Bumped the input to min=2 and reworded the help to describe m plainly. Fixed the same sentence in the CLI too, so the two surfaces don't disagree about the same flag. Added a small test on the bound.