# Given:
  #  X: An (NxM) numpy matrix of feature vectors for N objects, each of length M.
  #  L: An (Nx1) numpy array of boolean values describing labels for each object.
  #  IDS: A list of length N containing identifiers for each object.
  #  D: a linkage from scipy.linkage(X) or Fastcluster.linkage(X)

import dtcut as dtcut;
from EnrichmentTest import TEST;

  # Prepare the DTcut algorithm
T    = dtcut.prepare_tree(D);
tree = dtcut.DTCUT(T);

  # Prepare the statistical test
stat_test = TEST.Test(tree, IDS, X, L);
stat_func = lambda i_node: stat_test.test(i_node);

  # Set your parameters
pvalue_thresh = 0.05;
min_set_size  = 20;

  # Test the tree, getting significant node IDS
S = tree.test_tree(stat_func, pvalue_thresh, min_set_size);

  # For each leaf, get a cluster label
  # C[i] = cluster_id.
  # if C[i] = j < 0, the leaf is not in a significant cluster.
  # if C[i] = j > 0, the leaf is in significant cluster j.
C = tree.get_clusters(S);

  # Get information for each cluster
  # Returns
  # * Height of node in dendrogram
  # * p-value of node
  # * leaf members
I = tree.get_clusters_info(S, IDS);