P9 (WIP): compiled kernels (rust/simurg-core, PyO3) with bit-identical pure-Python fallback

Rust ports of the hot paths found by profiling: fused per-locus GenCall calling
and scoring, BAF/LRR, libm maps with CPython math semantics, both nearest-neighbour
searches of the normalization fit, and the BPM LocusEntry parser. Address
coverage is memoised per IDAT layout. SIMURG_PURE_PYTHON=1 forces the pure paths.
On the Pi: genotype_sample 3.4 s -> 1.1 s, 36-sample GSA run 131 s -> 46 s, BPM
read 4x faster; 36 GSA + 6 Omni2.5 GTCs and QC tables byte-identical to the pure
path; tests/unit/test_core_kernels.py checks every kernel on edge cases.
bench/profile_genotype.py, bench/scaling.py, cloud task p9 (thread scaling).

src/simurg/_core.py

@@ -0,0 +1,24 @@
+"""Optional compiled kernels (``rust/simurg-core``, plan §18 P9).
+
+``core`` is the ``simurg_core`` extension module when it is installed, else ``None``; every caller keeps a pure
+Python/NumPy path with bit-identical results. ``SIMURG_PURE_PYTHON=1`` forces the pure paths (tests, debugging).
+"""
+
+from __future__ import annotations
+
+import os
+from types import ModuleType
+
+core: ModuleType | None
+if os.environ.get("SIMURG_PURE_PYTHON", "") not in ("", "0"):
+    core = None
+else:
+    try:
+        import simurg_core as core  # type: ignore[no-redef]
+    except ImportError:  # pragma: no cover - depends on the installation
+        core = None
+
+
+def backend() -> str:
+    """``"rust <version>"`` or ``"python"`` (recorded in run manifests)."""
+    return f"rust {core.__version__}" if core is not None else "python"