Package rdkit :: Package ML :: Package Cluster :: Module ClusterUtils
[hide private]
[frames] | no frames]

Module ClusterUtils

source code

utility functions for clustering

Functions [hide private]
 
GetNodeList(cluster)
returns an ordered list of all nodes below cluster
source code
 
GetNodesDownToCentroids(cluster, above=1)
returns an ordered list of all nodes below cluster
source code
 
FindClusterCentroidFromDists(cluster, dists)
find the point in a cluster which has the smallest summed Euclidean distance to all others
source code
 
_BreadthFirstSplit(cluster, n)
*Internal Use Only*
source code
 
_HeightFirstSplit(cluster, n)
*Internal Use Only*
source code
 
SplitIntoNClusters(cluster, n, breadthFirst=1)
splits a cluster tree into a set of branches
source code
Variables [hide private]
  __package__ = None
hash(x)
Function Details [hide private]

GetNodeList(cluster)

source code 
returns an ordered list of all nodes below cluster

the ordering is done using the lengths of the child nodes

 **Arguments**

   - cluster: the cluster in question

 **Returns**

   - a list of the leaves below this cluster

FindClusterCentroidFromDists(cluster, dists)

source code 
find the point in a cluster which has the smallest summed 
  Euclidean distance to all others

**Arguments**

  - cluster: the cluster to work with

  - dists: the distance matrix to use for the points

**Returns**

  - the index of the centroid point

SplitIntoNClusters(cluster, n, breadthFirst=1)

source code 
splits a cluster tree into a set of branches

**Arguments**

  - cluster: the root of the cluster tree

  - n: the number of clusters to include in the split

  - breadthFirst: toggles breadth first (vs depth first) cleavage
    of the cluster tree.

**Returns**

  - a list of sub clusters