Add functionality for 3-D CCSN simulations.

ml4gw/waveforms/__init__.py

@@ -1,3 +1,4 @@
+from .ccsn import Pols_from_SDQM_3d
 from .phenom_d import IMRPhenomD
 from .phenom_p import IMRPhenomPv2
 from .ringdown import Ringdown

ml4gw/waveforms/ccsn.py

@@ -0,0 +1,80 @@
+import torch
+
+from ml4gw.types import ScalarTensor, TimeSeriesTensor
+
+
+class Pols_from_SDQM_3d(torch.nn.Module):
+
+    # Reference:
+    # https://academic.oup.com/ptps/article/doi/10.1143/PTPS.128.183/1930275
+
+    """
+    Callable class for generating cross and plus polarizations from
+    3D simulated second-derivative quadrupole moment.
+
+    Args:
+        sample_rate: Sample rate of waveform
+        duration: Duration of waveform
+    """
+
+    def __init__(self):
+
+        super().__init__()
+
+    def forward(
+        self,
+        sqdm: TimeSeriesTensor,
+        theta: ScalarTensor,
+        phi: ScalarTensor,
+    ):
+        """
+        Generate polarizations waveform based on the orientation.
+        The sphere cordinated follows physics convention.
+        See the following link for more detail.
+
+        link = ('https://en.wikipedia.org/wiki/Spherical_coordinate_system#
+        /media/File:3D_Spherical.svg')
+
+        Args:
+            sqdm:
+                Second-derivative quadrupole moment.
+            theta:
+                Polar angle. Range [0, pi]
+            phi:
+                Azimuthal angle. Range [0, 2 * pi)
+        Returns:
+            Tensors of cross and plus polarizations
+        """
+
+        ori_matrix_c = torch.zeros([len(theta), 3, 3])
+        ori_matrix_p = torch.zeros([len(theta), 3, 3])
+
+        # ori_matrix_c
+        ori_matrix_c[:, 0, 0] = -2 * (
+            torch.cos(theta) * torch.sin(phi) * torch.cos(phi)
+        )
+        ori_matrix_c[:, 0, 1] = 2 * torch.cos(theta) * torch.cos(2 * phi)
+        ori_matrix_c[:, 1, 1] = (
+            2 * torch.cos(theta) * torch.sin(phi) * torch.cos(phi)
+        )
+        ori_matrix_c[:, 1, 2] = -2 * (torch.sin(theta) * torch.cos(phi))
+        ori_matrix_c[:, 2, 0] = 2 * torch.sin(theta) * torch.sin(phi)
+
+        # ori_matrix_p
+        ori_matrix_p[:, 0, 0] = (
+            torch.cos(theta) ** 2 * torch.cos(phi) ** 2 - torch.sin(phi) ** 2
+        )
+        ori_matrix_p[:, 0, 1] = torch.cos(theta) ** 2 * torch.sin(
+            2 * phi
+        ) - torch.sin(2 * phi)
+        ori_matrix_p[:, 1, 1] = (
+            torch.cos(theta) ** 2 * torch.sin(phi) ** 2 - torch.cos(phi) ** 2
+        )
+        ori_matrix_p[:, 1, 2] = -(torch.sin(2 * theta) * torch.sin(phi))
+        ori_matrix_p[:, 2, 0] = -(torch.sin(2 * theta) * torch.cos(phi))
+        ori_matrix_p[:, 2, 2] = torch.sin(theta) ** 2
+
+        h_cross = torch.einsum("kji,nij->nk", sqdm, ori_matrix_c)
+        h_plus = torch.einsum("kji,nij->nk", sqdm, ori_matrix_p)
+
+        return h_cross, h_plus