add heightmap:resize(int, int, str)

src/maths/Heightmap.cpp

@@ -0,0 +1,77 @@
+#include "Heightmap.hpp"
+
+#include <cmath>
+#include <stdexcept>
+#include <glm/glm.hpp>
+
+static inline float smootherstep(float x) {
+    return glm::smoothstep(std::floor(x), std::floor(x)+1, x);
+}
+
+static inline float sample_at(
+    const float* buffer,
+    uint width, uint height,
+    uint x, uint y
+) {
+    return buffer[y*width+x];
+}
+
+static inline float sample_at(
+    const float* buffer,
+    uint width, uint height,
+    float x, float y,
+    InterpolationType interp
+) {
+    // std::floor is redundant here because x and y are positive values
+    uint ix = static_cast<uint>(x);
+    uint iy = static_cast<uint>(y);
+    float val = buffer[iy*width+ix];
+    if (interp == InterpolationType::NEAREST) {
+        return val;
+    }
+    float tx = x - ix;
+    float ty = y - iy;
+
+    switch (interp) {
+        case InterpolationType::LINEAR: {
+            float s00 = val;
+            float s10 = sample_at(buffer, width, height, 
+                ix + 1 < width ? ix + 1 : ix, iy);
+            float s01 = sample_at(buffer, width, height, ix, 
+                iy + 1 < height ? iy + 1 : iy);
+            float s11 = sample_at(buffer, width, height, 
+                ix + 1 < width ? ix + 1 : ix, iy + 1 < height ? iy + 1 : iy);
+
+            float a00 = s00;
+            float a10 = s10 - s00;
+            float a01 = s01 - s00;
+            float a11 = s11 - s10 - s01 + s00;
+
+            return a00 + a10*tx + a01*ty + a11*tx*ty;
+        }
+        // TODO: implement CUBIC (Bicubic) interpolation
+        default:
+            throw std::runtime_error("interpolation type is not implemented");
+    }
+    return val;
+}
+
+void Heightmap::resize(
+    uint dstwidth, uint dstheight, InterpolationType interp
+) {
+    std::vector<float> dst;
+    dst.resize(dstwidth*dstheight);
+
+    uint index = 0;
+    for (uint y = 0; y < dstheight; y++) {
+        for (uint x = 0; x < dstwidth; x++, index++) {
+            float sx = static_cast<float>(x) / dstwidth * width;
+            float sy = static_cast<float>(y) / dstheight * height;
+            dst[index] = sample_at(buffer.data(), width, height, sx, sy, interp);
+        }
+    }
+
+    width = dstwidth;
+    height = dstheight;
+    buffer = std::move(dst);
+}