Merge pull request #63 from punch-mission/develop

Merge develop

Author: jmbhughes

CODE_OF_CONDUCT.md

@@ -2,8 +2,9 @@ We welcome and encourage contributions to this project in the form of pull reque
 
 Contributors are expected
 
-to be respectful and constructive; and
-to enforce 1.
+1. to be respectful and constructive; and
+2. to enforce 1. 
+
 This code of conduct applies to all project-related communication that takes place on or at mailing lists, forums, social media, conferences, meetings, and social events.
 
-This code of conduct is from [the HAPI-Server project](https://github.com/hapi-server/client-python/blob/master/CODE_OF_CONDUCT.md). 
+This code of conduct is adapted from [the HAPI-Server project](https://github.com/hapi-server/client-python/blob/master/CODE_OF_CONDUCT.md). 

regularizepsf/corrector.py

@@ -4,8 +4,8 @@
 from pathlib import Path
 from typing import Any, Tuple
 
-import deepdish as dd
 import dill
+import h5py
 import numpy as np
 from numpy.fft import fft2, ifft2, ifftshift
 
@@ -264,11 +264,21 @@ def __getitem__(self, xy: Tuple[int, int]) -> np.ndarray:
             raise UnevaluatedPointError(f"Model not evaluated at {xy}.")
 
     def save(self, path: str) -> None:
-        dd.io.save(path, (self._evaluations, self._target_evaluation))
+        with h5py.File(path, 'w') as f:
+            eval_grp = f.create_group('evaluations')
+            for key, val in self._evaluations.items():
+                eval_grp.create_dataset(f'{key}', data=val)
+            f.create_dataset('target', data=self._target_evaluation)
 
     @classmethod
     def load(cls, path: str) -> ArrayCorrector:
-        evaluations, target_evaluation = dd.io.load(path)
+        with h5py.File(path, 'r') as f:
+            target_evaluation = f['target'][:].copy()
+
+            evaluations = dict()
+            for key, val in f['evaluations'].items():
+                parsed_key = tuple(int(val) for val in key.replace("(", "").replace(")", "").split(","))
+                evaluations[parsed_key] = val[:].copy()
         return cls(evaluations, target_evaluation)
 
     def simulate_observation(self, image: np.ndarray) -> np.ndarray:

regularizepsf/fitter.py

@@ -7,7 +7,7 @@
 from numbers import Real
 from typing import Any, Dict, Generator, List, Optional, Tuple
 
-import deepdish as dd
+import h5py
 import numpy as np
 import sep
 from astropy.io import fits
@@ -172,6 +172,7 @@ def fit(self,
 
         """
 
+    @abc.abstractmethod
     def save(self, path: str) -> None:
         """Save the PatchCollection to a file
 
@@ -184,9 +185,10 @@ def save(self, path: str) -> None:
         -------
         None
         """
-        dd.io.save(path, self.patches)
+
 
     @classmethod
+    @abc.abstractmethod
     def load(cls, path: str) -> PatchCollectionABC:
         """Load a PatchCollection from a file
 
@@ -200,7 +202,6 @@ def load(cls, path: str) -> PatchCollectionABC:
         PatchCollectionABC
             the new patch collection
         """
-        return cls(dd.io.load(path))
 
     def keys(self) -> List:
         """Gets identifiers for all patches"""
@@ -476,7 +477,7 @@ def average(self, corners: np.ndarray, patch_size: int, psf_size: int,  # noqa:
 
         for identifier, patch in self.patches.items():
             # Normalize the patch
-            patch = patch / np.max(patch)
+            patch = patch / patch[psf_size//2, psf_size//2]
 
             # Determine which average region it belongs to
             center_x = identifier.x + self.size // 2
@@ -569,3 +570,41 @@ def to_array_corrector(self, target_evaluation: np.array) -> ArrayCorrector:
 
         return ArrayCorrector(evaluation_dictionary, target_evaluation)
 
+    def save(self, path: str) -> None:
+        """Save the CoordinatePatchCollection to a file
+
+        Parameters
+        ----------
+        path : str
+            where to save the patch collection
+
+        Returns
+        -------
+        None
+        """
+        with h5py.File(path, 'w') as f:
+            patch_grp = f.create_group('patches')
+            for key, val in self.patches.items():
+                patch_grp.create_dataset(f"({key.image_index, key.x, key.y})", data=val)
+
+    @classmethod
+    def load(cls, path: str) -> PatchCollectionABC:
+        """Load a PatchCollection from a file
+
+        Parameters
+        ----------
+        path : str
+            file path to load from
+
+        Returns
+        -------
+        PatchCollectionABC
+            the new patch collection
+        """
+        patches = dict()
+        with h5py.File(path, "r") as f:
+            for key, val in f['patches'].items():
+                parsed_key = tuple(int(val) for val in key.replace("(", "").replace(")", "").split(","))
+                coord_id = CoordinateIdentifier(image_index=parsed_key[0], x=parsed_key[1], y=parsed_key[2])
+                patches[coord_id] = val[:].copy()
+        return cls(patches)

requirements.txt

@@ -1,9 +1,9 @@
-numpy==1.23.4
+numpy>=1.25.2
 dill==0.3.6
-deepdish==0.3.7
+h5py>=3.9.0
 lmfit==1.2.2
 cython==3.0.0
-astropy=5.3.1
+astropy==5.3.1
 scipy>=1.10.0
 scikit-image==0.19.3
 sep==1.2.1

requirements_dev.txt

@@ -1,6 +1,6 @@
-numpy==1.23.4
+numpy>=1.25.2
 dill==0.3.6
-deepdish==0.3.7
+h5py>=3.9.0
 lmfit==1.2.2
 cython==3.0.0
 astropy==5.3.1

setup.py

@@ -11,7 +11,7 @@
 
 setup(
     name='regularizepsf',
-    version='0.2.1',
+    version='0.2.2',
     description='Point spread function modeling and regularization',
     long_description=long_description,
     long_description_content_type='text/markdown',
@@ -21,7 +21,7 @@
     author='J. Marcus Hughes',
     author_email='[email protected]',
     ext_modules=cythonize(ext_modules, annotate=True, compiler_directives={'language_level': 3}),
-    install_requires=["numpy", "dill", "deepdish", "lmfit", "sep", "cython", "astropy", "scipy", "scikit-image", "matplotlib"],
+    install_requires=["numpy", "dill", "h5py", "lmfit", "sep", "cython", "astropy", "scipy", "scikit-image", "matplotlib"],
     package_data={"regularizepsf": ["helper.pyx"]},
     setup_requires=["cython"],
     extras_require={"test": ['pytest', 'coverage', 'pytest-runner', 'pytest-mpl']}

tests/test_corrector.py

@@ -55,22 +55,6 @@ def padded_100by100_image_psf_10_with_pattern():
     img_padded = np.pad(img, padding_shape, mode='constant')
     return img_padded
 
-#
-# @pytest.mark.parametrize("coord, value",
-#                          [((0, 0), 2),
-#                           ((10, 10), 1),
-#                           ((-10, -10), 0)])
-# def test_get_padded_img_section(coord, value, padded_100by100_image_psf_10_with_pattern):
-#     img_i = get_padded_img_section(padded_100by100_image_psf_10_with_pattern, coord[0], coord[1], 10)
-#     assert np.all(img_i == np.zeros((10, 10)) + value)
-#
-#
-# def test_set_padded_img_section(padded_100by100_image_psf_10_with_pattern):
-#     test_img = np.pad(np.ones((100, 100)), ((20, 20), (20, 20)), mode='constant')
-#     for coord, value in [((0, 0), 2), ((10, 10), 1), ((-10, -10), 0)]:
-#         set_padded_img_section(test_img, coord[0], coord[1], 10, np.zeros((10, 10))+value)
-#     assert np.all(test_img == padded_100by100_image_psf_10_with_pattern)
-
 
 def test_create_array_corrector():
     example = ArrayCorrector({(0, 0): np.zeros((10, 10))},
@@ -210,6 +194,8 @@ def test_save_load_array_corrector(tmp_path):
     assert os.path.isfile(fname)
     loaded = example.load(fname)
     assert isinstance(loaded, ArrayCorrector)
+    assert np.all(loaded._target_evaluation == np.ones((100, 100)))
+    assert np.all(loaded._evaluations[(0,0)] == np.ones((100, 100)))
 
 
 def test_array_corrector_simulate_observation_with_zero_stars():

tests/test_fitter.py

@@ -81,7 +81,7 @@ def test_coordinate_patch_average():
              })
     for patch in collection.values():
         # Make the normalization of each patch a no-op
-        patch[-1, -1] = 1
+        patch[5, 5] = 1
 
     averaged_collection = collection.average(
             np.array([[0, 0]]), 10, 10, mode='median')