Removes optional dependency bottleneck

README.md

@@ -43,17 +43,6 @@ or from this distribution with
 $ python setup.py install
 ```
 
-## Performance: Optional
-
-This library is optionally compatible with [bottleneck](https://github.com/pydata/bottleneck) and may offer performance improvements in some cases.
-Install bottleneck in your project as a peer to astroalign using:
-
-```bash
-pip install bottleneck
-```
-
-`Astroalign` will pick this optional dependency up and use it's performance improved functions for computing transforms.
-
 ## Running Tests
 
 ```bash

astroalign.py

@@ -56,13 +56,6 @@
     "register",
 ]
 
-try:
-    import bottleneck as bn
-except ImportError:
-    HAS_BOTTLENECK = False
-else:
-    HAS_BOTTLENECK = True
-
 import numpy as _np
 
 
@@ -109,26 +102,6 @@ def matrix_transform(*args, **kwargs):
 Default: 5
 """
 
-_default_median = bn.nanmedian if HAS_BOTTLENECK else _np.nanmedian  # pragma: no cover
-"""
-Default median function when/if optional bottleneck is available
-"""
-
-_default_average = bn.nanmean if HAS_BOTTLENECK else _np.nanmean  # pragma: no cover
-"""
-Default mean function when/if optional bottleneck is available
-"""
-
-_default_sum = bn.nansum if HAS_BOTTLENECK else _np.nansum  # pragma: no cover
-"""
-Default sum function when/if optional bottleneck is available
-"""
-
-_default_std = bn.nanstd if HAS_BOTTLENECK else _np.nanstd  # pragma: no cover
-"""
-Default std deviation function when/if optional bottleneck is available
-"""
-
 
 def _invariantfeatures(x1, x2, x3):
     """Given 3 points x1, x2, x3, return the invariant features for the set."""
@@ -270,7 +243,7 @@ def _bw(image):
     """Return a 2D numpy array for an array of arbitrary channels."""
     if image.ndim == 2:
         return image
-    return _default_average(image, axis=-1)
+    return _np.mean(image, axis=-1)
 
 
 def _shape(image):
@@ -471,7 +444,7 @@ def apply_transform(
         output_shape=target_shape,
         order=3,
         mode="constant",
-        cval=_default_median(source_data),
+        cval=_np.median(source_data),
         clip=True,
         preserve_range=True,
     )

docs/index.rst

@@ -29,10 +29,6 @@ You can find a Jupyter notebook example with the main features at http://quatrop
     If your images contain a large number of hot pixels, this may result in an incorrect registration.
     Please refer to the tutorial for how to solve this problem using `CCDProc's cosmic-ray remover <https://ccdproc.readthedocs.io/en/latest/api/ccdproc.cosmicray_lacosmic.html>`_.
 
-.. note::
-    This library is compatible with `bottleneck <https://github.com/pydata/bottleneck>`_ and may offer performance improvements in some cases.
-    `Astroalign` will pick this optional dependency up and use its performance improved ``mean`` and ``median`` functions.
-
 
 Citation
 --------

pyproject.toml

@@ -20,10 +20,6 @@ keywords = [
     "stellar-astronomical-images"
 ]
 
-
-[project.optional-dependencies]
-bottleneck = ["bottleneck"]
-
 [project.urls]
 homepage = "https://astroalign.quatrope.org"
 documentation = "https://astroalign.readthedocs.io/"