add a few more nodes

backend/nodes/level_rotate.py

@@ -0,0 +1,75 @@
+"""Level Rotate — level by physically rotating the data plane."""
+
+from __future__ import annotations
+
+import numpy as np
+from scipy.ndimage import map_coordinates
+
+from backend.node_registry import register_node
+from backend.data_types import DataField
+
+
+@register_node(display_name="Level Rotate")
+class LevelRotate:
+    @classmethod
+    def INPUT_TYPES(cls):
+        return {
+            "required": {
+                "field": ("DATA_FIELD",),
+            }
+        }
+
+    OUTPUTS = (
+        ('DATA_FIELD', 'leveled'),
+    )
+    FUNCTION = "process"
+
+    DESCRIPTION = (
+        "Level by physically rotating the data plane. Fits a best-fit plane, "
+        "converts its slopes to tilt angles, then rotates the surface by "
+        "those angles using interpolation rather than algebraic subtraction."
+    )
+
+    KEYWORDS = ("rotate", "tilt", "level", "plane")
+
+    def process(self, field: DataField) -> tuple:
+        data = np.asarray(field.data, dtype=np.float64)
+        yres, xres = data.shape
+
+        # Fit plane: z = a + bx*x + by*y (x,y in pixel coords)
+        yy, xx = np.mgrid[:yres, :xres].astype(np.float64)
+        A = np.column_stack([np.ones(yres * xres), xx.ravel(), yy.ravel()])
+        coeffs, _, _, _ = np.linalg.lstsq(A, data.ravel(), rcond=None)
+        _, bx, by = coeffs
+
+        # Convert pixel slopes to tilt angles
+        alpha_x = np.arctan(bx)
+        alpha_y = np.arctan(by)
+
+        # Build rotation: for each output pixel, find where it came from
+        cx = (xres - 1) / 2.0
+        cy = (yres - 1) / 2.0
+
+        cos_x = np.cos(alpha_x)
+        cos_y = np.cos(alpha_y)
+
+        # Source coordinates after removing tilt
+        src_x = xx.copy()
+        src_y = yy.copy()
+        src_z = data.copy()
+
+        # Rotate about x-axis (corrects y-tilt)
+        dy = yy - cy
+        src_y_rot = cy + dy * cos_y
+        src_z = src_z - dy * np.sin(alpha_y)
+
+        # Rotate about y-axis (corrects x-tilt)
+        dx = xx - cx
+        src_x_rot = cx + dx * cos_x
+        src_z = src_z - dx * np.sin(alpha_x)
+
+        # Resample with the adjusted z values
+        result = map_coordinates(src_z, [src_y_rot, src_x_rot], order=1,
+                                 mode='nearest')
+
+        return (field.replace(data=result),)