Merge ddf06cb into eac3c80

.github/workflows/publish.yml

@@ -17,10 +17,9 @@ jobs:
     - name: Install build dependencies
       run: |
         python -m pip install --upgrade pip
-        python -m pip install setuptools wheel numpy cython
+        python -m pip install setuptools wheel
     - name: Build wheel
       run: |
-        python setup.py build_ext --inplace
         python setup.py sdist bdist_wheel
     - name: Test wheel
       run: |

doc/environment.yml

@@ -6,6 +6,5 @@ dependencies:
   - rdkit>=2016.03.4
   - numpy>=1.11.3
   - scipy>=0.18.0
-  - cython>=0.25.2
   - sdaxen_python_utilities>=0.1.4
   - sphinxcontrib-programoutput

e3fp/fingerprint/metrics/array_metrics.py

@@ -10,7 +10,7 @@
 import numpy as np
 import scipy
 from scipy.sparse import csr_matrix, issparse, vstack
-from ._fast import soergel as fast_soergel
+from e3fp.util import maybe_jit
 
 
 def tanimoto(X, Y=None):
@@ -58,15 +58,22 @@ def soergel(X, Y=None):
     -------
     soergel : array of shape (`n_fprints_X`, `n_fprints_Y`)
 
+    Notes
+    --------
+    If Numba is available, this function is jit-compiled and much more efficient.
+
     See Also
     --------
     tanimoto: A fast version of this function for binary data.
     pearson: Pearson correlation, also appropriate for non-binary data.
     cosine, dice
     """
     X, Y = _check_array_pair(X, Y)
-    return fast_soergel(X, Y, sparse=issparse(X))
-
+    S = np.empty((X.shape[0], Y.shape[0]), dtype=np.float64)
+    if issparse(X):
+        return _sparse_soergel(X.data, X.indices, X.indptr,
+                               Y.data, Y.indices, Y.indptr, S)
+    return _dense_soergel(X, Y, S)
 
 def dice(X, Y=None):
     """Compute the Dice coefficients between `X` and `Y`.
@@ -212,3 +219,81 @@ def _sparse_cosine(X, Y):
     XY = (X * Y.T).toarray()
     with np.errstate(divide="ignore"):  # handle 0 in denominator
         return np.nan_to_num(XY / (Xnorm * Ynorm.T))
+
+@maybe_jit(nopython=True, nogil=True, cache=True)
+def _dense_soergel(X, Y, S):
+    for ix in range(S.shape[0]):
+        for iy in range(S.shape[1]):
+            sum_abs_diff = 0
+            sum_max = 0
+            for j in range(X.shape[1]):
+                diff = X[ix, j] - Y[iy, j]
+                if diff > 0:
+                    sum_abs_diff += diff
+                    sum_max += X[ix, j]
+                else:
+                    sum_abs_diff -= diff
+                    sum_max += Y[iy, j]
+
+            if sum_max == 0:
+                S[ix, iy] = 0
+                continue
+            S[ix, iy] = 1 - sum_abs_diff / sum_max
+    return S
+
+@maybe_jit(nopython=True, nogil=True, cache=True)
+def _sparse_soergel(Xdata, Xindices, Xindptr, Ydata, Yindices, Yindptr, S):
+    for ix in range(S.shape[0]):
+        if Xindptr[ix] == Xindptr[ix + 1]:
+            for iy in range(S.shape[1]):  # no X values in row
+                S[ix, iy] = 0
+            continue
+        jxindmax = Xindptr[ix + 1] - 1
+        for iy in range(S.shape[1]):
+            if Yindptr[iy] == Yindptr[iy + 1]:  # no Y values in row
+                S[ix, iy] = 0
+                continue
+
+            sum_abs_diff = 0
+            sum_max = 0
+            # Implementation of the final step of merge sort
+            jyindmax = Yindptr[iy + 1] - 1
+            jx = Xindptr[ix]
+            jy = Yindptr[iy]
+            while jx <= jxindmax and jy <= jyindmax:
+                jxind = Xindices[jx]
+                jyind = Yindices[jy]   
+                if jxind < jyind:
+                    sum_max += Xdata[jx]
+                    sum_abs_diff += Xdata[jx]
+                    jx += 1
+                elif jyind < jxind:
+                    sum_max += Ydata[jy]
+                    sum_abs_diff += Ydata[jy]
+                    jy += 1
+                else:
+                    diff = Xdata[jx] - Ydata[jy]
+                    if diff > 0:
+                        sum_abs_diff += diff
+                        sum_max += Xdata[jx]
+                    else:
+                        sum_abs_diff -= diff
+                        sum_max += Ydata[jy]
+                    jx += 1
+                    jy += 1
+
+            while jx <= jxindmax:
+                sum_max += Xdata[jx]
+                sum_abs_diff += Xdata[jx]
+                jx += 1
+
+            while jy <= jyindmax:
+                sum_max += Ydata[jy]
+                sum_abs_diff += Ydata[jy]
+                jy += 1
+
+            if sum_max == 0:
+                S[ix, iy] = 0
+                continue
+            S[ix, iy] = 1 - sum_abs_diff / sum_max
+    return S

e3fp/util.py

@@ -33,6 +33,25 @@ class E3FPEfficiencyWarning(E3FPWarning, RuntimeWarning):
     """A warning class for a potentially inefficient process."""
 
 
+def maybe_jit(*args, **kwargs):
+    """Decorator to jit a function using Numba if available.
+    
+    Usage is identical to `numba.jit`.
+    """
+    def wrapper(func):
+        try:
+            import numba
+            has_numba = True
+        except ImportError:
+            has_numba = False
+
+        if has_numba:
+            return numba.jit(*args, **kwargs)(func)
+        else:
+            return func
+    return wrapper
+
+
 class deprecated(object):
     """Decorator to mark a function as deprecated.
 

environment.yml

@@ -6,7 +6,6 @@ dependencies:
   - rdkit>=2016.03.4
   - numpy>=1.11.3
   - scipy>=0.18.0
-  - cython>=0.25.2
   - setuptools
   - sdaxen_python_utilities>=0.1.4
   # optional

setup.py

@@ -2,8 +2,6 @@
 
 from setuptools import setup
 from setuptools.command.build_ext import build_ext
-from setuptools.extension import Extension
-from Cython.Build import cythonize
 import numpy as np
 from e3fp import version
 
@@ -14,7 +12,6 @@
     "scipy>=0.18.0",
     "numpy>=1.11.3",
     "mmh3>=2.3.1",
-    "cython>=0.25.2",
     "sdaxen_python_utilities>=0.1.4",
 ]
 if ON_RTD:  # ReadTheDocs can't handle C libraries
@@ -28,7 +25,6 @@
     "Programming Language :: Python :: 3.6",
     "Programming Language :: Python :: 3.7",
     "Programming Language :: Python :: 3.8",
-    "Programming Language :: Cython",
     "License :: OSI Approved :: GNU Lesser General Public License v3 (LGPLv3)",
     "Operating System :: OS Independent",
     "Development Status :: 4 - Beta",
@@ -43,13 +39,6 @@ def get_readme():
     with open("README.rst") as f:
         return f.read()
 
-ext_modules = [
-    Extension(
-        "e3fp.fingerprint.metrics._fast",
-        sources=["e3fp/fingerprint/metrics/_fast.pyx"],
-        include_dirs=[np.get_include()],
-    )
-]
 
 setup(
     name="e3fp",
@@ -74,5 +63,4 @@ def get_readme():
     include_package_data=True,
     test_suite="nose.collector",
     tests_require=test_requirements,
-    ext_modules=cythonize(ext_modules),
 )