class VariantSparkAPITestCase(VariantSparkPySparkTestCase):
def setUp(self):
self.spark = SparkSession(self.sc)
self.vc = VariantsContext(self.spark)
def test_variants_context_parameter_type(self):
with self.assertRaises(TypeError) as cm:
self.vc.load_label(label_file_path=123, col_name=456)
self.assertEqual('keyword argument label_file_path = 123 doesn\'t match signature str',
str(cm.exception))
def test_importance_analysis_from_vcf(self):
label_data_path = os.path.join(PROJECT_DIR, 'data/chr22-labels.csv')
label = self.vc.load_label(label_file_path=label_data_path, col_name='22_16050678')
feature_data_path = os.path.join(PROJECT_DIR, 'data/chr22_1000.vcf')
features = self.vc.import_vcf(vcf_file_path=feature_data_path)
imp_analysis = features.importance_analysis(label, 200, None, True, 17, 50, 3)
imp_vars = imp_analysis.important_variables(20)
most_imp_var = imp_vars[0][0]
self.assertEqual('22_16050678_C_T', most_imp_var)
df = imp_analysis.variable_importance()
self.assertEqual('22_16050678_C_T',
str(df.orderBy('importance', ascending=False).collect()[0][0]))
oob_error = imp_analysis.oob_error()
self.assertAlmostEqual(0.004578754578754579, oob_error, 4)
def main():
spark = SparkSession.builder\
.config(conf=VariantsContext.spark_conf()) \
.appName("HipsterIndex") \
.getOrCreate()
vs = VariantsContext(spark)
features = vs.import_vcf(os.path.join(PROJECT_DIR, 'data/chr22_1000.vcf'))
labels = vs.load_label(os.path.join(PROJECT_DIR, 'data/chr22-labels.csv'),
'22_16050408')
model = features.importance_analysis(labels,
mtry_fraction=0.1,
seed=13,
n_trees=200)
print("Oob = %s" % model.oob_error())
for entry in model.important_variables(10):
print entry
def setUp(self):
self.spark = SparkSession(self.sc)
self.vc = VariantsContext(self.spark)