fix(mcp-server): use double-checked locking to prevent race condition in _init()

[zong0728]

Fixes #169

_init() checked if model is None without any synchronisation. Under concurrent tool calls, multiple threads could observe model as None simultaneously and each create a separate SentenceTransformer, wasting memory and causing non-deterministic initialisation behaviour.

Root Cause

The check-then-act on the global has no lock. Two threads arriving before the first load completes both pass the if model is None guard and each run SentenceTransformer(EMBEDDING_MODEL), doubling memory usage.

Fix

Add a module-level threading.Lock() and apply double-checked locking: fast unsynchronised check on the hot path, safe initialisation under the lock. Only one thread ever runs the constructor.

Testing

• Single-threaded path behaviour unchanged
• Concurrent calls can no longer double-initialise the model

Checklist

• Commits are signed off (DCO)
• Fixes bug(mcp-server): race condition in _init() allows multiple threads to initialize model and client simultaneously #169
• No behaviour change on single-threaded path
• Hardcoded credentials untouched (tracked separately in bug(mcp-server): Placeholder <YOUR_NAMESPACE> in Milvus connection URI #91 / PR fix: Replace hardcoded Milvus credentials with env vars (fixes #91)  #135)

[google-oss-prow]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign franciscojavierarceo for approval. For more information see the Kubernetes Code Review Process.

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