Added __version__ variable back in for wheels packages.

docs/conf.py

@@ -1,4 +1,3 @@
-
 # -- Project information -----------------------------------------------------
 from pkg_resources import DistributionNotFound, get_distribution
 

src/mpol/__init__.py

@@ -1 +1,8 @@
+import importlib.metadata
+
+try:
+    __version__ = importlib.metadata.version("MPoL")
+except importlib.metadata.PackageNotFoundError:
+    __version__ = "unknown"
+
 zenodo_record = 10064221

src/mpol/coordinates.py

@@ -104,7 +104,7 @@ class GridCoords:
         length-4 list of (left, right, bottom, top) expected by routines
         like ``matplotlib.pyplot.imshow`` in the ``extent`` parameter assuming
         ``origin='lower'``. Units of [:math:`\lambda`]
-    vis_ext_Mlam : list 
+    vis_ext_Mlam : list
         like vis_ext, but in units of [:math:`\mathrm{M}\lambda`].
     """
 
@@ -273,7 +273,7 @@ def ground_q_centers_2D(self) -> npt.NDArray[np.floating[Any]]:
     @cached_property
     def sky_phi_centers_2D(self) -> npt.NDArray[np.floating[Any]]:
         # https://en.wikipedia.org/wiki/Atan2
-        return np.arctan2( # type: ignore
+        return np.arctan2(  # type: ignore
             self.ground_v_centers_2D, self.ground_u_centers_2D
         )  # (pi, pi]
 

src/mpol/crossval.py

@@ -56,7 +56,7 @@
 #         Number of k-folds to use in cross-validation
 #     split_method : str, default='dartboard'
 #         Method to split full dataset into train/test subsets
-#     dartboard_q_edges, dartboard_phi_edges : list of float, default=None, 
+#     dartboard_q_edges, dartboard_phi_edges : list of float, default=None,
 #   unit=[klambda]
 #         Radial and azimuthal bin edges of the cells used to split the dataset
 #         if `split_method`==`dartboard` (see `datasets.Dartboard`)
@@ -510,7 +510,7 @@ def __init__(
         self.cell_list = self.dartboard.get_nonzero_cell_indices(qs, phis)
 
         # indices of cells in the smallest q bin that also have data
-        small_q_idx = [i for i, l in enumerate(self.cell_list) if l[0] == 0] #type: ignore
+        small_q_idx = [i for i, l in enumerate(self.cell_list) if l[0] == 0]  # type: ignore
         # cells in the smallest q bin
         self.small_q = self.cell_list[: len(small_q_idx)]
 
@@ -554,7 +554,7 @@ def from_dartboard_properties(
                 seed (int): (optional) numpy random seed to use for the permutation,
                 for reproducibility
             verbose (bool): whether to print notification messages
-    """
+        """
         dartboard = Dartboard(gridded_dataset.coords, q_edges, phi_edges)
         return cls(gridded_dataset, k, dartboard, seed, verbose)
 

src/mpol/datasets.py

@@ -19,7 +19,7 @@ class GriddedDataset(torch.nn.Module):
         If providing this, cannot provide ``cell_size`` or ``npix``.
     vis_gridded : :class:`torch.Tensor` of :class:`torch.complex128`
         the gridded visibility data stored in a "packed" format (pre-shifted for fft)
-    weight_gridded : :class:`torch.Tensor` 
+    weight_gridded : :class:`torch.Tensor`
         the weights corresponding to the gridded visibility data,
         also in a packed format
     mask : :class:`torch.Tensor` of :class:`torch.bool`
@@ -117,9 +117,9 @@ def forward(self, modelVisibilityCube: torch.Tensor) -> torch.Tensor:
                 across cube dimensions.
         """
 
-        assert (
-            modelVisibilityCube.size()[0] == self.mask.size()[0]
-        ), "vis and dataset mask do not have the same number of channels."
+        assert modelVisibilityCube.size()[0] == self.mask.size()[0], (
+            "vis and dataset mask do not have the same number of channels."
+        )
 
         # As of Pytorch 1.7.0, complex numbers are partially supported.
         # However, masked_select does not yet work (with gradients)
@@ -226,8 +226,10 @@ def get_polar_histogram(
 
         histogram: NDArray
         # make a polar histogram
-        histogram, *_ = np.histogram2d( # type:ignore
-            qs, phis, bins=[self.q_edges.tolist(), self.phi_edges.tolist()] # type:ignore
+        histogram, *_ = np.histogram2d(  # type:ignore
+            qs,
+            phis,
+            bins=[self.q_edges.tolist(), self.phi_edges.tolist()],  # type:ignore
         )
 
         return histogram

src/mpol/exceptions.py

@@ -1,21 +1,16 @@
 from __future__ import annotations
 
 
-class CellSizeError(Exception):
-    ...
+class CellSizeError(Exception): ...
 
 
-class WrongDimensionError(Exception):
-    ...
+class WrongDimensionError(Exception): ...
 
 
-class DataError(Exception):
-    ...
+class DataError(Exception): ...
 
 
-class DimensionMismatchError(Exception):
-    ...
+class DimensionMismatchError(Exception): ...
 
 
-class ThresholdExceededError(Exception):
-    ...
+class ThresholdExceededError(Exception): ...

src/mpol/fourier.py

@@ -275,14 +275,14 @@ def forward(
 
         Parameters
         ----------
-        packed_cube : :class:`torch.Tensor` 
+        packed_cube : :class:`torch.Tensor`
             shape ``(nchan, npix, npix)``). The cube
             should be a "prepacked" image cube, for example,
             from :meth:`mpol.images.ImageCube.forward`
-        uu : :class:`torch.Tensor` 
+        uu : :class:`torch.Tensor`
             2D array of the u (East-West) spatial frequency coordinate
             [:math:`\lambda`] of shape ``(nchan, npix)``
-        vv : :class:`torch.Tensor` 
+        vv : :class:`torch.Tensor`
             2D array of the v (North-South) spatial frequency coordinate
             [:math:`\lambda`] (must be the same shape as uu)
         sparse_matrices : bool
@@ -506,7 +506,7 @@ def forward(self, packed_cube):
         ``uu`` and ``vv`` points.
 
         Args:
-            packed_cube : :class:`torch.Tensor` 
+            packed_cube : :class:`torch.Tensor`
                 shape ``(nchan, npix, npix)``). The cube
                 should be a "prepacked" image cube, for example,
                 from :meth:`mpol.images.ImageCube.forward`
@@ -530,7 +530,6 @@ def forward(self, packed_cube):
         # convert the cube to a complex type, since this is required by TorchKbNufft
         complexed = shifted + 0j
 
-
         # Consider how the similarity of the spatial frequency samples should be
         # treated. We already took care of this on the k_traj side, since we set
         # the shapes. But this also needs to be taken care of on the image side.

src/mpol/geometry.py

@@ -1,5 +1,4 @@
-"""The geometry package provides routines for projecting and de-projecting sky images.
-"""
+"""The geometry package provides routines for projecting and de-projecting sky images."""
 
 import numpy as np
 import torch

src/mpol/images.py

@@ -83,9 +83,9 @@ def __init__(
             self.base_cube = nn.Parameter(base_cube, requires_grad=True)
 
         if pixel_mapping is None:
-            self.pixel_mapping: Callable[
-                [torch.Tensor], torch.Tensor
-            ] = torch.nn.Softplus()
+            self.pixel_mapping: Callable[[torch.Tensor], torch.Tensor] = (
+                torch.nn.Softplus()
+            )
         else:
             self.pixel_mapping = pixel_mapping
 
@@ -193,10 +193,10 @@ def forward(self, cube: torch.Tensor) -> torch.Tensor:
 
 class GaussConvImage(nn.Module):
     r"""
-    This convolutional layer will convolve the input cube with a 2D Gaussian kernel. 
+    This convolutional layer will convolve the input cube with a 2D Gaussian kernel.
     The filter is the same for all channels in the input cube.
     Because the operation is carried out in the image domain, note that it may become
-    computationally prohibitive for large kernel sizes. In that case, 
+    computationally prohibitive for large kernel sizes. In that case,
     :class:`mpol.images.GaussConvFourier` may be preferred.
 
     Parameters
@@ -329,6 +329,7 @@ def forward(self, sky_cube: torch.Tensor) -> torch.Tensor:
         convolved_sky = self.m(sky_cube)
         return convolved_sky
 
+
 class GaussConvFourier(nn.Module):
     r"""
     This layer will convolve the input cube with a (potentially non-circular) Gaussian
@@ -349,19 +350,18 @@ class GaussConvFourier(nn.Module):
     """
 
     def __init__(
-        self,
-        coords: GridCoords,
-        FWHM_maj: float,
-        FWHM_min: float,
-        Omega: float = 0) -> None:
+        self, coords: GridCoords, FWHM_maj: float, FWHM_min: float, Omega: float = 0
+    ) -> None:
         super().__init__()
 
         self.coords = coords
         self.FWHM_maj = FWHM_maj
-        self.FWHM_min = FWHM_min 
+        self.FWHM_min = FWHM_min
         self.Omega = Omega
 
-        taper_2D = uv_gaussian_taper(self.coords, self.FWHM_maj, self.FWHM_min, self.Omega)
+        taper_2D = uv_gaussian_taper(
+            self.coords, self.FWHM_maj, self.FWHM_min, self.Omega
+        )
 
         # store taper to register so it transfers to GPU
         self.register_buffer("taper_2D", torch.tensor(taper_2D, dtype=torch.float32))
@@ -382,10 +382,10 @@ def forward(self, packed_cube, thresh=1e-6):
             The convolved cube in packed format.
         """
         nchan, npix_m, npix_l = packed_cube.size()
-        assert (
-            (npix_m == self.coords.npix) and (npix_l == self.coords.npix)
-        ), "packed_cube {:} does not have the same pixel dimensions as indicated by coords {:}".format(
-            packed_cube.size(), self.coords.npix
+        assert (npix_m == self.coords.npix) and (npix_l == self.coords.npix), (
+            "packed_cube {:} does not have the same pixel dimensions as indicated by coords {:}".format(
+                packed_cube.size(), self.coords.npix
+            )
         )
 
         # in FFT packed format
@@ -400,10 +400,10 @@ def forward(self, packed_cube, thresh=1e-6):
         convolved_packed_cube = torch.fft.ifftn(tapered_vis, dim=(1, 2))
 
         # assert imaginaries are effectively zero, otherwise something went wrong
-        assert (
-            torch.max(convolved_packed_cube.imag) < thresh
-        ), "Round-tripped image contains max imaginary value {:} > {:} threshold, something may be amiss.".format(
-            torch.max(convolved_packed_cube.imag), thresh
+        assert torch.max(convolved_packed_cube.imag) < thresh, (
+            "Round-tripped image contains max imaginary value {:} > {:} threshold, something may be amiss.".format(
+                torch.max(convolved_packed_cube.imag), thresh
+            )
         )
 
         r_cube: torch.Tensor = convolved_packed_cube.real
@@ -580,4 +580,4 @@ def uv_gaussian_taper(
 
     # a flux-conserving taper must have an amplitude of 1 at the origin.
 
-    return taper_2D
+    return taper_2D

src/mpol/input_output.py

@@ -18,17 +18,16 @@ def __init__(self, filename, channel=0):
         self._fits_file = filename
         self._channel = channel
 
-
     def get_extent(self, header):
         """Get extent (in RA and Dec, units of [arcsec]) of image"""
 
         # get the coordinate labels
         nx = header["NAXIS1"]
         ny = header["NAXIS2"]
 
-        assert (
-            nx % 2 == 0 and ny % 2 == 0
-        ), f"Image dimensions x {nx} and y {ny} must be even."
+        assert nx % 2 == 0 and ny % 2 == 0, (
+            f"Image dimensions x {nx} and y {ny} must be even."
+        )
 
         # RA coordinates
         CDELT1 = 3600 * header["CDELT1"]  # arcsec (converted from decimal deg)
@@ -53,7 +52,6 @@ def get_extent(self, header):
 
         return RA, DEC, ext
 
-
     def get_beam(self, hdu_list, header):
         """Get the major and minor widths [arcsec], and position angle, of a
         clean beam"""
@@ -72,7 +70,6 @@ def get_beam(self, hdu_list, header):
 
         return BMAJ, BMIN, BPA
 
-
     def get_image(self, beam=True):
         """Load a .fits image and return as a numpy array. Also return image
         extent and optionally (`beam`) the clean beam dimensions"""