fullwave25 1.0.7__py3-none-any.whl

fullwave/utils/numerical.py

@@ -0,0 +1,111 @@
+"""Utility functions for numerical operations."""
+
+import os
+import random
+
+import numpy as np
+import scipy
+
+
+def seed_everything(seed: int) -> None:
+    """Apply seed."""
+    random.seed(seed)
+    os.environ["PYTHONHASSEED"] = str(seed)
+    np.random.seed(seed)  # noqa: NPY002
+
+
+def normalize_255(image: np.ndarray) -> np.ndarray:
+    """Normalize image data to the 0-255 range."""
+    return (image - image.min()) / (image.max() - image.min() + 1e-6) * 255
+
+
+def matlab_round(value: float) -> int:
+    """MATLAB style rounding of a value."""
+    return np.round(value + 1e-9).astype(int)
+
+
+def matlab_gaussian_filter(image: np.ndarray, sigma: float) -> np.ndarray:
+    """Apply a Gaussian filter similar to MATLAB's implementation."""
+    return scipy.ndimage.gaussian_filter(
+        image.astype(float),
+        sigma=sigma,
+        radius=np.ceil(2 * sigma).astype(int),
+    )
+
+
+def matlab_interp2easy(
+    image: np.ndarray,
+    interpolation_x: float,
+    interpolation_y: float,
+) -> np.ndarray:
+    """Perform interpolation similar to MATLAB's interp2 function."""
+    dxi = 1 / matlab_round(image.shape[0] * interpolation_x - 1)
+    xvec = np.arange(0, 1 + dxi, dxi)
+    xvec = xvec * (image.shape[0] - 1)
+
+    dyi = 1 / matlab_round(image.shape[1] * interpolation_y - 1)
+    yvec = np.arange(0, 1 + dyi, dyi)
+    yvec = yvec * (image.shape[1] - 1)
+    xi, yi = np.meshgrid(xvec, yvec)
+
+    return (
+        scipy.ndimage.map_coordinates(image, [xi.ravel(), yi.ravel()], order=0, mode="nearest")
+        .reshape(yvec.shape[0], xvec.shape[0])
+        .T
+    )
+
+
+def matlab_interp2easy_3d(
+    image: np.ndarray,
+    interpolation_x: float,
+    interpolation_y: float,
+    interpolation_z: float,
+) -> np.ndarray:
+    """Perform interpolation similar to MATLAB's interp2 function."""
+    dxi = 1 / matlab_round(image.shape[0] * interpolation_x - 1)
+    xvec = np.arange(0, 1 + dxi, dxi)
+    xvec = xvec * (image.shape[0] - 1)
+
+    dyi = 1 / matlab_round(image.shape[1] * interpolation_y - 1)
+    yvec = np.arange(0, 1 + dyi, dyi)
+    yvec = yvec * (image.shape[1] - 1)
+
+    dzi = 1 / matlab_round(image.shape[2] * interpolation_z - 1)
+    zvec = np.arange(0, 1 + dzi, dzi)
+    zvec = zvec * (image.shape[2] - 1)
+
+    xi, yi, zi = np.meshgrid(xvec, yvec, zvec, indexing="ij")
+
+    return (
+        scipy.ndimage.map_coordinates(
+            image,
+            [xi.ravel(), yi.ravel(), zi.ravel()],
+            order=1,
+            mode="nearest",
+        ).reshape(xvec.shape[0], yvec.shape[0], zvec.shape[0])
+        # .reshape(yvec.shape[0], xvec.shape[0], zvec.shape[0])
+        # .T
+    )
+
+
+def detect_envelope(image: np.ndarray) -> np.ndarray:
+    """Compute the envelope of a signal using the Hilbert transform."""
+    return np.abs(scipy.signal.hilbert(image, axis=0))
+
+
+def cut_domain(
+    map_data: np.ndarray,
+    i_event: int,
+    num_x: int,
+    num_y: int,
+    n_events: int,
+    beam_spacing: float,
+) -> np.ndarray:
+    """Extract a subdomain from map_data based on event parameters."""
+    orig = int(
+        np.round(map_data.shape[0] / 2)
+        - np.round(num_x / 2)
+        - np.round(((i_event + 1) - (n_events + 1) / 2) * beam_spacing)
+        - 6,
+    )
+    return map_data[orig : orig + num_x, 0:num_y]