Fast path for hot property getters: stop importing mcrfpy per call; closes #331

Replace the PyImport_ImportModule("mcrfpy") + PyObject_GetAttrString +
full type-call ceremony with direct &mcrfpydef::PyXType references
(safe since all PyTypeObjects are inline C++17 definitions) in every
non-init call site:

- UIDrawable: get_pos/get_origin/get_global_pos now allocate via
  PyVector(...).pyObject(); set_pos/set_origin check the Vector type
  directly.
- GridLayers: ColorLayer at/subscript wrap via PyColor(...).pyObject()
  (also fixes a module refcount leak per read); set/fill/fill_rect,
  draw_fov/apply_perspective color parsing, Entity/Texture/HeightMap
  instance checks, and TileLayer texture get/set use direct type refs.
- UIGrid init layers=, add_layer/remove_layer/layer(), Grid.layers
  getter: direct ColorLayer/TileLayer type refs.
- PyTileSetFile.to_texture, PyLdtkProject.tileset: direct type refs.

Benchmark (tests/benchmarks/issue_331_property_read_bench.py, 200k
reads, release build):
  frame.pos              559 ns -> 65 ns/read  (8.6x)
  frame.origin           571 ns -> 64 ns/read  (8.9x)
  frame.global_position  589 ns -> 66 ns/read  (8.9x)
  ColorLayer.at          608 ns -> 262 ns/read (2.3x)

Test suite: 304/304 pass.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DK5DHgRDNGmpN54vnEHAot

tests/benchmarks/issue_331_property_read_bench.py

@@ -0,0 +1,74 @@
+#!/usr/bin/env python3
+"""
+Issue #331: micro-benchmark for hot property getters.
+
+Times tight-loop reads of properties that (pre-fix) executed
+PyImport_ImportModule("mcrfpy") + PyObject_GetAttrString + full type call
+per read, versus the direct tp_alloc fast path (post-fix).
+
+Direct-execution benchmark: no game loop needed, property reads work
+immediately after object construction.
+
+Usage:
+    ./mcrogueface --headless --exec tests/benchmarks/issue_331_property_read_bench.py
+"""
+
+import mcrfpy
+import sys
+import time
+
+N = 200_000
+
+
+def bench(label, fn):
+    start = time.perf_counter()
+    fn()
+    elapsed = time.perf_counter() - start
+    per_read_ns = elapsed / N * 1e9
+    print(f"{label:24s} {elapsed*1000:8.1f} ms total  {per_read_ns:8.0f} ns/read")
+    return per_read_ns
+
+
+def main():
+    frame = mcrfpy.Frame(pos=(10, 20), size=(100, 100))
+    layer = mcrfpy.ColorLayer(name="bench", z_index=0)
+    grid = mcrfpy.Grid(grid_size=(8, 8), layers=[layer])
+    layer.set((3, 3), (10, 20, 30, 255))
+
+    def read_pos():
+        for _ in range(N):
+            frame.pos
+
+    def read_origin():
+        for _ in range(N):
+            frame.origin
+
+    def read_global_pos():
+        for _ in range(N):
+            frame.global_position
+
+    def read_layer_at():
+        for _ in range(N):
+            layer.at((3, 3))
+
+    print(f"issue #331 property-read benchmark, N={N}")
+    results = {
+        "pos": bench("frame.pos", read_pos),
+        "origin": bench("frame.origin", read_origin),
+        "global_pos": bench("frame.global_position", read_global_pos),
+        "layer_at": bench("ColorLayer.at", read_layer_at),
+    }
+
+    # Sanity: values must be correct regardless of construction path
+    p = frame.pos
+    assert (p.x, p.y) == (10.0, 20.0), f"pos wrong: {p}"
+    c = layer.at((3, 3))
+    assert (c.r, c.g, c.b) == (10, 20, 30), f"color wrong: {c}"
+
+    print("PASS")
+    return results
+
+
+if __name__ == "__main__":
+    main()
+    sys.exit(0)