tag_groups = mfv.autogen_tag_groups()
machines = ['win7x32','win7x64']
#seperate by machine
for machine in machines :
for tags in tag_groups :
#group name auto generated by tags and machine
group_name = "{0}_{1}_{2}_{3}_{4}".format(tags[0][1],tags[1][1],tags[2][1],tags[3][1],machine)
print "Creating automated stats for {0}".format(group_name)
task_ids = None
hashes = None
#get subset
subset_vectors = mfv.select_vectors(tags, [machine], task_ids, hashes)
subset_vectors = mfv.normalize_vectors(subset_vectors, max_values_vector)
#get superset - keep only fileytpe tags and machine
superset_vectors = mfv.select_vectors([tags[1]],[machine],None,None)
superset_vectors = mfv.normalize_vectors(superset_vectors, max_values_vector)
#find archetype vector and extract feature keys
archetype, archetype_stddev = mfv.get_archetype(subset_vectors, superset_vectors)
feature_keys = []
for key, value in archetype.features.iteritems() :
feature_keys.append(key)
#find groups stats and plot archetype features for group
mfv.stats_summary(subset_vectors,superset_vectors)
title = "Scatter Plot of Family {0}".format(group_name)
test_task_id = int(sys.argv[1])
#retrieve features from the database
cmd = "SELECT label,mean_features,stddev_features,max_value_features FROM archetypes"
cursor.execute(cmd)
results = cursor.fetchall()
for result in results :
#parse data from results
archetype_label = result[0]
mean_features = json.loads(result[1])
stddev_features = json.loads(result[2])
max_value_features = json.loads(result[3])
#add features to FeatureVectors
mean_vector = mfv.FeatureVector(None,None,None,mean_features)
stddev_vector = mfv.FeatureVector(None,None,None,stddev_features)
max_values_vector = mfv.FeatureVector(None,None,None,max_value_features)
#retrieve and normalize the test_vector
test_vector = mfv.select_vectors(None,None,[test_task_id],None)[0]
test_vectors = mfv.normalize_vectors([test_vector],max_values_vector)
#add labels to vectors
mean_vector.label = archetype_label+"_mean"
stddev_vector.label = archetype_label+"_stddev"
title = "Comparison of Task {0} to Archetype : {1}".format(test_task_id,archetype_label)
filename = "comparisions/{0}/{1}".format(test_task_id, archetype_label)
print mfv.plotly_scatter(test_vectors, mean_vector, stddev_vector, None, filename, title)
#retrieve features from the database
cmd = "SELECT label,mean_features,stddev_features,max_value_features FROM archetypes"
cursor.execute(cmd)
results = cursor.fetchall()
for result in results:
#parse data from results
archetype_label = result[0]
mean_features = json.loads(result[1])
stddev_features = json.loads(result[2])
max_value_features = json.loads(result[3])
#add features to FeatureVectors
mean_vector = mfv.FeatureVector(None, None, None, mean_features)
stddev_vector = mfv.FeatureVector(None, None, None, stddev_features)
max_values_vector = mfv.FeatureVector(None, None, None, max_value_features)
#retrieve and normalize the test_vector
test_vector = mfv.select_vectors(None, None, [test_task_id], None)[0]
test_vectors = mfv.normalize_vectors([test_vector], max_values_vector)
#add labels to vectors
mean_vector.label = archetype_label + "_mean"
stddev_vector.label = archetype_label + "_stddev"
title = "Comparison of Task {0} to Archetype : {1}".format(
test_task_id, archetype_label)
filename = "comparisions/{0}/{1}".format(test_task_id, archetype_label)
print mfv.plotly_scatter(test_vectors, mean_vector, stddev_vector, None,
filename, title)
