docs: document approx_mode tuning

[BubbleCal]

Summary

• Document the query-time approx_mode parameter in the vector search ANN index docs.
• List the supported fast, normal, and accurate modes and their speed/recall trade-offs.
• Clarify that normal only has an effect for RQ indexes built with num_bits > 1; for num_bits=1, it uses the same one-bit scoring path as fast.
• Note that approx_mode can be changed per query without rebuilding the index and interacts with refine_factor before reranking.

Validation

• uv run python - <<'PY' ... content check for the new approx_mode docs section
• uv run pytest tests/py/test_search.py (15 passed, 2 skipped)

[mintlify]

Preview deployment for your docs. Learn more about Mintlify Previews.

Project	Status	Preview	Updated (UTC)
lancedb-bcbb4faf	🟢 Ready	View Preview	Jun 25, 2026, 9:02 AM