def __compute_cluster_score_means(self, network_score):
"""compute the score means on the given network score"""
result = {}
for cluster in xrange(1, self.num_clusters() + 1):
cluster_scores = [network_score[cluster][gene]
if gene in network_score[cluster] else 0.0
for gene in self.rows_for_cluster(cluster)]
result[cluster] = util.trim_mean(cluster_scores, 0.05)
return result
def compute_iteration_scores(network_iteration_scores):
"""called 'cluster.ns' in the original cMonkey"""
result = {}
for cluster in network_iteration_scores:
cluster_scores = []
for _, score in network_iteration_scores[cluster].items():
cluster_scores.append(score)
result[cluster] = util.trim_mean(cluster_scores, 0.05)
return result
def log_subresult(self, score_function, matrix):
"""output an accumulated subresult to the log"""
scores = []
mvalues = matrix.values
for cluster in xrange(1, matrix.num_columns + 1):
cluster_rows = self.membership.rows_for_cluster(cluster)
for row in xrange(matrix.num_rows):
if matrix.row_names[row] in cluster_rows:
scores.append(mvalues[row][cluster - 1])
logging.debug("function '%s', trim mean score: %f", score_function.id,
util.trim_mean(scores, 0.05))
def __compute_cluster_score_means(self, network_score):
"""compute the score means on the given network score"""
result = {}
for cluster in xrange(1, self.num_clusters() + 1):
cluster_scores = [
network_score[cluster][gene]
if gene in network_score[cluster] else 0.0
for gene in self.rows_for_cluster(cluster)
]
result[cluster] = util.trim_mean(cluster_scores, 0.05)
return result
def log_subresult(self, score_function, matrix):
"""output an accumulated subresult to the log"""
scores = []
mvalues = matrix.values
for cluster in xrange(1, matrix.num_columns + 1):
cluster_rows = self.membership.rows_for_cluster(cluster)
for row in xrange(matrix.num_rows):
if matrix.row_names[row] in cluster_rows:
scores.append(mvalues[row][cluster - 1])
logging.info("function '%s', trim mean score: %f",
score_function.name(),
util.trim_mean(scores, 0.05))
def __compute_cluster_score_means(self, network_score):
"""compute the score means on the given network score"""
result = {}
for cluster in xrange(1, self.num_clusters() + 1):
cluster_scores = []
for gene in sorted(self.rows_for_cluster(cluster)):
if gene in network_score[cluster].keys():
cluster_scores.append(network_score[cluster][gene])
else:
cluster_scores.append(0.0)
result[cluster] = util.trim_mean(cluster_scores, 0.05)
return result
def __compute_cluster_score_means(self, network_score):
"""compute the score means on the given network score"""
result = {}
for cluster in xrange(1, self.num_clusters() + 1):
## TODO: we can rewrite this as a list comprehension. Check whether that
## will be faster
cluster_scores = []
for gene in sorted(self.rows_for_cluster(cluster)):
if gene in network_score[cluster].keys():
cluster_scores.append(network_score[cluster][gene])
else:
cluster_scores.append(0.0)
result[cluster] = util.trim_mean(cluster_scores, 0.05)
return result
def test_trim_mean_real(self):
values = [0.0, 0.0, -8.7520618359684352, -8.7520618359684352, 0.0, 0.0,
0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
self.assertAlmostEqual(-1.4586770, util.trim_mean(values, 0.05))
def test_trim_mean_no_values(self):
self.assertEqual(0, util.trim_mean([], 0.05))
def test_trim_mean_median(self):
self.assertAlmostEqual(3.5, util.trim_mean([.1, .2, 3, 4, 5, 6], 0.5))
def test_trim_mean_nonmedian(self):
self.assertAlmostEqual(
40.625,
util.trim_mean([2, 4, 6, 7, 11, 21, 81, 90, 105, 121], 0.1))
