fix(lancedb): keep FTS migration nonblocking

[SiyaoZheng]

Summary

Keep the one-time FTS schema migration from performing synchronous marker-file I/O inside its async startup path.

migrate_fts_indexes() previously called pathlib.Path.exists(), read_text(), and write_text() directly. On a slow or busy filesystem these calls block the event loop during service startup. This change wraps the marker in anyio.Path and awaits the file operations while preserving the existing migration/version semantics.

Area

• Architecture method
• Benchmark
• Use case
• Documentation
• Developer experience
• CI, build, or release

Verification

uv run pytest tests/unit/test_core/test_persistence/test_lancedb/test_fts_behavior.py tests/unit/test_core/test_persistence/test_lancedb/test_repository.py -q
44 passed

make ci
1500 unit tests passed; 78 integration tests passed (6 deselected);
lint, architecture contracts, OpenAPI drift, package build, and wheel smoke test passed

Checklist

• I kept the change scoped to the relevant area.
• I am opening this from a separate branch, not pushing directly to main.
• No documentation change is required because the migration contract is unchanged.
• I added or updated tests when the change affects behavior.
• I did not commit secrets, .env files, dependency folders, or generated output.
• Active relative links in Markdown files resolve.

Notes for Reviewers

The tests verify async marker reads and writes for both the migration and already-migrated paths.

By submitting this pull request, I agree that my contribution is licensed under the Apache License 2.0.