move eval_filter_grid to operators.filters

remove method from ImageSource

fix imports

no self

import cleanup

docstring

clarity and style fixes

Author: chris-langfield

src/aspire/covariance/covar.py

@@ -11,6 +11,7 @@
 from aspire import config
 from aspire.image import Image
 from aspire.nufft import anufft
+from aspire.operators import evaluate_grid_src
 from aspire.reconstruction import Estimator, FourierKernel, MeanEstimator
 from aspire.utils import (
     ensure,
@@ -47,7 +48,7 @@ def compute_kernel(self):
         _2L = 2 * self.L
 
         kernel = np.zeros((_2L, _2L, _2L, _2L, _2L, _2L), dtype=self.dtype)
-        sq_filters_f = self.src.eval_filter_grid(self.L, power=2)
+        sq_filters_f = evaluate_grid_src(self.src, self.L, power=2)
 
         for i in tqdm(range(0, n, self.batch_size)):
             _range = np.arange(i, min(n, i + self.batch_size))

src/aspire/operators/__init__.py

@@ -13,6 +13,7 @@
     ScalarFilter,
     ScaledFilter,
     ZeroFilter,
+    evaluate_grid_src,
     voltage_to_wavelength,
 )
 from .wemd import wemd_embed, wemd_norm

src/aspire/operators/filters.py

@@ -33,6 +33,30 @@ def wavelength_to_voltage(wavelength):
     ) / (2 * 0.978466)
 
 
+def evaluate_grid_src(src, L, power=1):
+    """
+    Given an ImageSource object, compute the source's unique filters
+    at the filter_indices specified in its metadata.
+    :return: an `L x L x len(src.filter_indices)` array containing the evaluated
+    filters at each gridpoint
+    """
+    grid2d = grid_2d(L, dtype=src.dtype)
+    omega = np.pi * np.vstack((grid2d["x"].flatten(), grid2d["y"].flatten()))
+
+    h = np.empty((omega.shape[-1], len(src.filter_indices)), dtype=src.dtype)
+    for i, filt in enumerate(src.unique_filters):
+        idx_k = np.where(src.filter_indices == i)[0]
+        if len(idx_k) > 0:
+            filter_values = filt.evaluate(omega)
+            if power != 1:
+                filter_values **= power
+            h[:, idx_k] = np.column_stack((filter_values,) * len(idx_k))
+
+    h = np.reshape(h, grid2d["x"].shape + (len(src.filter_indices),))
+
+    return h
+
+
 # TODO: filters should probably be dtyped...
 class Filter:
     def __init__(self, dim=None, radial=False):
@@ -116,20 +140,6 @@ def evaluate_grid(self, L, dtype=np.float32, *args, **kwargs):
 
         return h
 
-    def evaluate_grid_src(self, src, L, power=1):
-        grid2d = grid_2d(L, dtype=src.dtype)
-        omega = np.pi * np.vstack((grid2d["x"].flatten(), grid2d["y"].flatten()))
-        h = np.empty((omega.shape[-1], len(src.filter_indices)), dtype=src.dtype)
-        for i, filt, in enumerate(src.unique_filters):
-            idx_k = np.where(src.filter_indices == i)[0]
-            if len(idx_k) > 0:
-                filter_values = filt.evaluate(omega)
-                if power != 1:
-                    filter_values **= power
-                h[:, idx_k] = np.column_stack((filter_values,)*len(idx_k))
-        h = np.reshape(h, grid2d["x"].shape + (len(src.filter_indices),))
-        return h
-
     def dual(self):
         return DualFilter(self)
 

src/aspire/reconstruction/mean.py

@@ -4,6 +4,7 @@
 from scipy.fftpack import fft2
 
 from aspire.nufft import anufft
+from aspire.operators import evaluate_grid_src
 from aspire.reconstruction import Estimator, FourierKernel
 from aspire.utils.fft import mdim_ifftshift
 from aspire.utils.matlab_compat import m_flatten, m_reshape
@@ -16,7 +17,7 @@ class MeanEstimator(Estimator):
     def compute_kernel(self):
         _2L = 2 * self.L
         kernel = np.zeros((_2L, _2L, _2L), dtype=self.dtype)
-        sq_filters_f = self.src.eval_filter_grid(self.L, power=2)
+        sq_filters_f = evaluate_grid_src(self.src, self.L, power=2)
 
         for i in range(0, self.n, self.batch_size):
             _range = np.arange(i, min(self.n, i + self.batch_size), dtype=int)

src/aspire/source/image.py

@@ -358,23 +358,6 @@ def eval_filters(self, im_orig, start=0, num=np.inf, indices=None):
 
         return im
 
-    def eval_filter_grid(self, L, power=1):
-        grid2d = grid_2d(L, dtype=self.dtype)
-        omega = np.pi * np.vstack((grid2d["x"].flatten(), grid2d["y"].flatten()))
-
-        h = np.empty((omega.shape[-1], len(self.filter_indices)), dtype=self.dtype)
-        for i, filt in enumerate(self.unique_filters):
-            idx_k = np.where(self.filter_indices == i)[0]
-            if len(idx_k) > 0:
-                filter_values = filt.evaluate(omega)
-                if power != 1:
-                    filter_values **= power
-                h[:, idx_k] = np.column_stack((filter_values,) * len(idx_k))
-
-        h = np.reshape(h, grid2d["x"].shape + (len(self.filter_indices),))
-
-        return h
-
     def cache(self):
         logger.info("Caching source images")
         self._cached_im = self.images(start=0, num=np.inf)