def __init__(self, appName="PyHail", master=None, local='local[*]',
log='hail.log', quiet=False, append=False, parquet_compression='uncompressed',
block_size=1, branching_factor=50, tmp_dir='/tmp'):
from pyspark import SparkContext
SparkContext._ensure_initialized()
self.gateway = SparkContext._gateway
self.jvm = SparkContext._jvm
self.jsc = scala_package_object(self.jvm.org.broadinstitute.hail.driver).configureAndCreateSparkContext(
appName, joption(self.jvm, master), local,
log, quiet, append, parquet_compression,
block_size, branching_factor, tmp_dir)
self.sc = SparkContext(gateway=self.gateway, jsc=self.jvm.JavaSparkContext(self.jsc))
self.jsql_context = scala_package_object(self.jvm.org.broadinstitute.hail.driver).createSQLContext(self.jsc)
self.sql_context = SQLContext(self.sc, self.jsql_context)
def count(self, genotypes=False):
"""Return number of samples, varaints and genotypes.
:param bool genotypes: If True, return number of called
genotypes and genotype call rate.
"""
return (scala_package_object(self.hc.jvm.org.broadinstitute.hail.driver)
.count(self.jvds, genotypes)
.toJavaMap())
def count(self, genotypes=False):
"""Return number of samples, varaints and genotypes.
:param bool genotypes: If True, return number of called
genotypes and genotype call rate.
"""
try:
return (scala_package_object(self.hc.jvm.org.broadinstitute.hail.driver)
.count(self.jvds, genotypes)
.toJavaMap())
except Py4JJavaError as e:
self._raise_py4j_exception(e)
def make_keytable(self, variant_condition, genotype_condition, key_names):
"""Make a KeyTable with one row per variant.
Per sample field names in the result are formed by concatening
the sample ID with the genotype_condition left hand side with
dot (.). If the left hand side is empty::
`` = expr
then the dot (.) is ommited.
**Example**
Consider a ``VariantDataset`` ``vds`` with 2 variants and 3 samples::
Variant FORMAT A B C
1:1:A:T GT:GQ 0/1:99 ./. 0/0:99
1:2:G:C GT:GQ 0/1:89 0/1:99 1/1:93
Then::
>>> vds = hc.import_vcf('data/sample.vcf')
>>> vds.make_keytable('v = v', 'gt = g.gt', gq = g.gq', [])
returns a ``KeyTable`` with schema::
v: Variant
A.gt: Int
B.gt: Int
C.gt: Int
A.gq: Int
B.gq: Int
C.gq: Int
in particular, the values would be::
v A.gt B.gt C.gt A.gq B.gq C.gq
1:1:A:T 1 NA 0 99 NA 99
1:2:G:C 1 1 2 89 99 93
:param variant_condition: Variant annotation expressions.
:type variant_condition: str or list of str
:param genotype_condition: Genotype annotation expressions.
:type genotype_condition: str or list of str
:param key_names: list of key columns
:type key_names: list of str
:rtype: KeyTable
"""
if isinstance(variant_condition, list):
variant_condition = ','.join(variant_condition)
if isinstance(genotype_condition, list):
genotype_condition = ','.join(genotype_condition)
jkt = (scala_package_object(self.hc.jvm.org.broadinstitute.hail.driver)
.makeKT(self.jvds, variant_condition, genotype_condition,
jarray(self.hc.gateway, self.hc.jvm.java.lang.String, key_names)))
return KeyTable(self.hc, jkt)
def _raise_py4j_exception(self, e):
msg = scala_package_object(self.jvm.org.broadinstitute.hail.utils).getMinimalMessage(e.java_exception)
raise FatalError(msg, e.java_exception)
