Minor update

examples/neighbors/pamm.py

@@ -1,18 +1,22 @@
-# %%
-# Probabilistic Analysis of Molecular Motifs (PAMM)
-# -------------------------------------------------
-#
+#!/usr/bin/env python
+# coding: utf-8
+
+"""
+Probabilistic Analysis of Molecular Motifs (PAMM)
+================================================
+
+Probabilistic analysis of molecular motifs (`PAMM <https://doi.org/10.1063/1.4900655>`_)
+is a method identifying molecular patterns based on an analysis of the probability
+distribution of fragments observed in an atomistic simulation. With the help of sparse
+KDE, it can be easily conducted.
+
+Here we define some functions to help us. `quick_shift_refinement` is used to refine the
+clusters generated by `QuickShift` by merging outlier clusters into their nearest
+neighbours. `generate_probability_model` is to interpret the quick shift results into
+a probability model. `cluster_distribution_3D` is to plot the probability model
+of the H-bond motif.
+"""
 
-# %%
-# Probabilistic analysis of molecular motifs is a method identifying molecular patterns
-# based on an analysis of the probability distribution of fragments observed in an
-# atomistic simulation. With the help of sparse KDE, it can be easily conducted. Here
-# we define some functions to help us. `quick_shift_refinement` is used to refine the
-# clusters generated by `QuickShift` by merging outlier clusters into their nearest
-# neighbours. `generate_probability_model` is to interpret the quick shift results into
-# a probability model. `cluster_distribution_3D` is to plot the probability model
-# of the H-bond motif.
-#
 
 # %%
 from typing import Callable, Union

src/skmatter/neighbors/_sparsekde.py

@@ -20,10 +20,7 @@ class SparseKDE(BaseEstimator):
     This class is used to build a sparse kernel density estimator.
     It takes a set of descriptors and a set of weights as input,
     and fit the KDE model on the sampled point (e.g. the grid point
-    selected by FPS). First, the probability density is estimated for
-    each sampled point. Then, quick shift clustering is applied to the
-    grid points. Finally, a kernel density estimator is built based on
-    the clustering results.
+    selected by FPS).
 
     .. note::
         Currently only the Gaussian kernel is supported.
@@ -55,6 +52,8 @@ class SparseKDE(BaseEstimator):
     fpoints : float, default=0.15
         The fractional number of points in the voronoi cell of each grid points. Use
         this when each cell has a similar number of points.
+    kernel : str, default=gaussian
+        The kernel used here. Now only the gaussian kernel is available.
     verbose : bool, default=False
         Whether to print progress.