addressed comments

python/hail/docs/representation/index.rst

@@ -14,7 +14,7 @@ representation
 
     hail.representation.AltAllele
     hail.representation.Call
-    hail.representation.GenomeReference
+    hail.representation.ReferenceGenome
     hail.representation.Genotype
     hail.representation.Interval
     hail.representation.Locus

python/hail/expr.py

@@ -571,7 +571,7 @@ class TVariant(Type):
     """
 
     def __init__(self):
-        jtype = scala_object(Env.hail().expr, 'TVariant').apply(Env.hail().variant.GenomeReference.GRCh37())
+        jtype = scala_object(Env.hail().expr, 'TVariant').apply(Env.hail().variant.ReferenceGenome.GRCh37())
         super(TVariant, self).__init__(jtype)
 
     @classmethod
@@ -721,7 +721,7 @@ class TLocus(Type):
     """
 
     def __init__(self):
-        jtype = scala_object(Env.hail().expr, 'TLocus').apply(Env.hail().variant.GenomeReference.GRCh37())
+        jtype = scala_object(Env.hail().expr, 'TLocus').apply(Env.hail().variant.ReferenceGenome.GRCh37())
         super(TLocus, self).__init__(jtype)
 
     @classmethod
@@ -763,7 +763,7 @@ class TInterval(Type):
     """
 
     def __init__(self):
-        jtype = scala_object(Env.hail().expr, 'TInterval').apply(Env.hail().variant.GenomeReference.GRCh37())
+        jtype = scala_object(Env.hail().expr, 'TInterval').apply(Env.hail().variant.ReferenceGenome.GRCh37())
         super(TInterval, self).__init__(jtype)
 
     @classmethod

python/hail/representation/__init__.py

@@ -3,7 +3,7 @@
 from hail.representation.genotype import Genotype, Call
 from hail.representation.annotations import Struct
 from hail.representation.pedigree import Trio, Pedigree
-from hail.representation.genomeref import GenomeReference
+from hail.representation.genomeref import ReferenceGenome
 
 __all__ = ['Variant',
            'Locus',
@@ -14,4 +14,4 @@
            'Call',
            'Pedigree',
            'Trio',
-           'GenomeReference']
+           'ReferenceGenome']

python/hail/representation/genomeref.py

@@ -5,33 +5,33 @@
 from hail.history import *
 
 
-class GenomeReference(HistoryMixin):
-    """An object that represents a genome reference.
+class ReferenceGenome(HistoryMixin):
+    """An object that represents a `reference genome <https://en.wikipedia.org/wiki/Reference_genome>`_.
 
-    :param str name: Name of reference
+    :param str name: Name of reference.
 
-    :param contigs: Contig names
+    :param contigs: Contig names.
     :type contigs: list of str
 
-    :param lengths: Dict of contig names to contig lengths
+    :param lengths: Dict of contig names to contig lengths.
     :type lengths: dict of str to int
 
-    :param x_contigs: Contigs to be treated as X chromosomes
+    :param x_contigs: Contigs to be treated as X chromosomes.
     :type x_contigs: str or list of str
 
-    :param y_contigs: Contigs to be treated as Y chromosomes
+    :param y_contigs: Contigs to be treated as Y chromosomes.
     :type y_contigs: str or list of str
 
-    :param mt_contigs: Contigs to be treated as mitochondrial chromosomes
+    :param mt_contigs: Contigs to be treated as mitochondrial DNA.
     :type mt_contigs: str or list of str
 
-    :param par: List of intervals representing pseudoautosomal regions
+    :param par: List of intervals representing pseudoautosomal regions.
     :type par: list of :class:`.Interval`
 
     >>> contigs = ["1", "X", "Y", "MT"]
     >>> lengths = {"1": 249250621, "X": 155270560, "Y": 59373566, "MT": 16569}
     >>> par = [Interval.parse("X:60001-2699521")]
-    >>> my_gr = GenomeReference("my_gr", contigs, lengths, "X", "Y", "MT", par)
+    >>> my_ref = ReferenceGenome("my_ref", contigs, lengths, "X", "Y", "MT", par)
     """
 
     @handle_py4j
@@ -50,7 +50,7 @@ def __init__(self, name, contigs, lengths, x_contigs=[], y_contigs=[], mt_contig
         mt_contigs = wrap_to_list(mt_contigs)
         par_jrep = [interval._jrep for interval in par]
 
-        jrep = (Env.hail().variant.GenomeReference
+        jrep = (Env.hail().variant.ReferenceGenome
                 .apply(name,
                        contigs,
                        lengths,
@@ -68,14 +68,14 @@ def __init__(self, name, contigs, lengths, x_contigs=[], y_contigs=[], mt_contig
         self._mt_contigs = mt_contigs
         self._par = par
 
-        super(GenomeReference, self).__init__()
+        super(ReferenceGenome, self).__init__()
 
     @handle_py4j
     def __str__(self):
         return self._jrep.toString()
 
     def __repr__(self):
-        return 'GenomeReference(name=%s, contigs=%s, lengths=%s, x_contigs=%s, y_contigs=%s, mt_contigs=%s, par=%s)' % \
+        return 'ReferenceGenome(name=%s, contigs=%s, lengths=%s, x_contigs=%s, y_contigs=%s, mt_contigs=%s, par=%s)' % \
                (self.name, self.contigs, self.lengths, self.x_contigs, self.y_contigs, self.mt_contigs, self.par)
 
     @handle_py4j
@@ -88,67 +88,68 @@ def __hash__(self):
 
     @property
     def name(self):
-        """Name of genome reference
+        """Name of reference genome.
 
         :rtype: str
         """
         return self._name
 
     @property
     def contigs(self):
-        """Contig names
+        """Contig names.
 
         :rtype: list of str
         """
         return self._contigs
 
     @property
     def lengths(self):
-        """Map of contig name to contig length
+        """Dict of contig name to contig length.
 
         :rtype: dict of str to int
         """
         return self._lengths
 
     @property
     def x_contigs(self):
-        """X contigs
+        """X contigs.
 
         :rtype: list of str
         """
         return self._x_contigs
 
     @property
     def y_contigs(self):
-        """Y contigs
+        """Y contigs.
 
         :rtype: list of str
         """
         return self._y_contigs
 
     @property
     def mt_contigs(self):
-        """Mitochondrial contigs
+        """Mitochondrial contigs.
 
         :rtype: list of str
         """
         return self._mt_contigs
 
     @property
     def par(self):
-        """Pseudoautosomal regions
+        """Pseudoautosomal regions.
 
         :rtype: list of :class:`.Interval`
         """
         return self._par
 
     @typecheck_method(contig=strlike)
     def contig_length(self, contig):
-        """Contig length
+        """Contig length.
 
-        :param contig: Contig to get length of
+        :param contig: Contig
         :type contig: str
 
+        :return: Length of contig
         :rtype: int
         """
         return self._jrep.contigLength(contig)
@@ -157,43 +158,43 @@ def contig_length(self, contig):
     @record_classmethod
     @handle_py4j
     def GRCh37(cls):
-        """Genome reference for GRCh37
+        """Reference genome for GRCh37.
 
-        Data from `<ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/b37/human_g1k_v37.dict>`_
+        Data from `GATK resource bundle <ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/b37/human_g1k_v37.dict>`_.
 
-        >>> gr37 = GenomeReference.GRCh37()
+        >>> grch37 = ReferenceGenome.GRCh37()
 
-        :rtype: :class:`.GenomeReference`
+        :rtype: :class:`.ReferenceGenome`
         """
-        return GenomeReference._from_java(Env.hail().variant.GenomeReference.GRCh37())
+        return ReferenceGenome._from_java(Env.hail().variant.ReferenceGenome.GRCh37())
 
     @classmethod
     @record_classmethod
     @handle_py4j
     def GRCh38(cls):
-        """Genome reference for GRCh38
+        """Reference genome for GRCh38.
 
-        Data from `<ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/hg38/Homo_sapiens_assembly38.dict>`_
+        Data from `GATK resource bundle <ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/hg38/Homo_sapiens_assembly38.dict>`_.
 
-        >>> gr38 = GenomeReference.GRCh38()
+        >>> grch38 = ReferenceGenome.GRCh38()
 
-        :rtype: :class:`.GenomeReference`
+        :rtype: :class:`.ReferenceGenome`
         """
-        return GenomeReference._from_java(Env.hail().variant.GenomeReference.GRCh38())
+        return ReferenceGenome._from_java(Env.hail().variant.ReferenceGenome.GRCh38())
 
     @handle_py4j
     def _init_from_java(self, jrep):
         self._jrep = jrep
 
     @classmethod
     def _from_java(cls, jrep):
-        gr = GenomeReference.__new__(cls)
-        gr._init_from_java(jrep)
-        gr._name = jrep.name()
-        gr._contigs = [str(x) for x in jrep.contigs()]
-        gr._lengths = {str(x._1()): int(x._2()) for x in jiterable_to_list(jrep.lengths())}
-        gr._x_contigs = [str(x) for x in jiterable_to_list(jrep.xContigs())]
-        gr._y_contigs = [str(x) for x in jiterable_to_list(jrep.yContigs())]
-        gr._mt_contigs = [str(x) for x in jiterable_to_list(jrep.mtContigs())]
-        gr._par = [Interval._from_java(x) for x in jrep.par()]
-        return gr
+        rg = ReferenceGenome.__new__(cls)
+        rg._init_from_java(jrep)
+        rg._name = jrep.name()
+        rg._contigs = [str(x) for x in jrep.contigs()]
+        rg._lengths = {str(x._1()): int(x._2()) for x in jiterable_to_list(jrep.lengths())}
+        rg._x_contigs = [str(x) for x in jiterable_to_list(jrep.xContigs())]
+        rg._y_contigs = [str(x) for x in jiterable_to_list(jrep.yContigs())]
+        rg._mt_contigs = [str(x) for x in jiterable_to_list(jrep.mtContigs())]
+        rg._par = [Interval._from_java(x) for x in jrep.par()]
+        return rg

python/hail/tests/tests.py

@@ -730,14 +730,14 @@ def test_representation(self):
         self.assertFalse(c_nocall.is_het_non_ref())
         self.assertFalse(c_nocall.is_het_ref())
 
-        gr = GenomeReference.GRCh37()
-        self.assertEqual(gr.name, "GRCh37")
-        self.assertEqual(gr.contigs[0], "1")
-        self.assertListEqual(gr.x_contigs, ["X"])
-        self.assertListEqual(gr.y_contigs, ["Y"])
-        self.assertListEqual(gr.mt_contigs, ["MT"])
-        self.assertEqual(gr.par[0], Interval.parse("X:60001-2699521"))
-        self.assertEqual(gr.contig_length("1"), 249250621)
+        rg = ReferenceGenome.GRCh37()
+        self.assertEqual(rg.name, "GRCh37")
+        self.assertEqual(rg.contigs[0], "1")
+        self.assertListEqual(rg.x_contigs, ["X"])
+        self.assertListEqual(rg.y_contigs, ["Y"])
+        self.assertListEqual(rg.mt_contigs, ["MT"])
+        self.assertEqual(rg.par[0], Interval.parse("X:60001-2699521"))
+        self.assertEqual(rg.contig_length("1"), 249250621)
 
         name = "test"
         contigs = ["1", "X", "Y", "MT"]
@@ -747,15 +747,15 @@ def test_representation(self):
         mt_contigs = ["MT"]
         par = [Interval(Locus("X", 5), Locus("X", 1000))]
 
-        gr2 = GenomeReference(name, contigs, lengths, x_contigs, y_contigs, mt_contigs, par)
-        self.assertEqual(gr2.name, name)
-        self.assertListEqual(gr2.contigs, contigs)
-        self.assertListEqual(gr2.x_contigs, x_contigs)
-        self.assertListEqual(gr2.y_contigs, y_contigs)
-        self.assertListEqual(gr2.mt_contigs, mt_contigs)
-        self.assertEqual(gr2.par, par)
-        self.assertEqual(gr2.contig_length("1"), 10000)
-        self.assertDictEqual(gr2.lengths, lengths)
+        rg2 = ReferenceGenome(name, contigs, lengths, x_contigs, y_contigs, mt_contigs, par)
+        self.assertEqual(rg2.name, name)
+        self.assertListEqual(rg2.contigs, contigs)
+        self.assertListEqual(rg2.x_contigs, x_contigs)
+        self.assertListEqual(rg2.y_contigs, y_contigs)
+        self.assertListEqual(rg2.mt_contigs, mt_contigs)
+        self.assertEqual(rg2.par, par)
+        self.assertEqual(rg2.contig_length("1"), 10000)
+        self.assertDictEqual(rg2.lengths, lengths)
 
     def test_types(self):
         self.assertEqual(TInt32(), TInt32())

src/main/scala/is/hail/HailContext.scala

@@ -11,7 +11,7 @@ import is.hail.io.vcf._
 import is.hail.keytable.KeyTable
 import is.hail.stats.{BaldingNicholsModel, Distribution, UniformDist}
 import is.hail.utils.{log, _}
-import is.hail.variant.{GenericDataset, GenomeReference, Genotype, Locus, VSMFileMetadata, VSMSubgen, Variant, VariantDataset, VariantSampleMatrix}
+import is.hail.variant.{GenericDataset, ReferenceGenome, Genotype, Locus, VSMFileMetadata, VSMSubgen, Variant, VariantDataset, VariantSampleMatrix}
 import org.apache.hadoop
 import org.apache.log4j.{ConsoleAppender, LogManager, PatternLayout, PropertyConfigurator}
 import org.apache.spark.deploy.SparkHadoopUtil
@@ -296,7 +296,7 @@ class HailContext private(val sc: SparkContext,
 
     val signature = TStruct("rsid" -> TString, "varid" -> TString)
 
-    val rdd = sc.union(results.map(_.rdd)).toOrderedRDD(TVariant(GenomeReference.GRCh37).orderedKey, classTag[(Annotation, Iterable[Annotation])])
+    val rdd = sc.union(results.map(_.rdd)).toOrderedRDD(TVariant(ReferenceGenome.GRCh37).orderedKey, classTag[(Annotation, Iterable[Annotation])])
 
     new GenericDataset(this,
       VSMFileMetadata(samples,
@@ -575,7 +575,7 @@ class HailContext private(val sc: SparkContext,
 
   def eval(expr: String): (Annotation, Type) = {
     val ec = EvalContext(
-      "v" -> TVariant(GenomeReference.GRCh37),
+      "v" -> TVariant(ReferenceGenome.GRCh37),
       "s" -> TString,
       "g" -> TGenotype,
       "sa" -> TStruct(

src/main/scala/is/hail/annotations/MemoryBlock.scala

@@ -662,8 +662,8 @@ class RegionValueBuilder(region: MemoryBuffer) {
         case t: TInterval =>
           val i = a.asInstanceOf[Interval[Locus]]
           startStruct()
-          addAnnotation(TLocus(t.gr), i.start)
-          addAnnotation(TLocus(t.gr), i.end)
+          addAnnotation(TLocus(t.rg), i.start)
+          addAnnotation(TLocus(t.rg), i.end)
           endStruct()
       }
   }

src/main/scala/is/hail/annotations/UnsafeRow.scala

@@ -6,7 +6,7 @@ import com.esotericsoftware.kryo.io.{Input, Output}
 import com.esotericsoftware.kryo.{Kryo, KryoSerializable}
 import is.hail.expr._
 import is.hail.utils.Interval
-import is.hail.variant.{AltAllele, GenericGenotype, GenomeReference, Locus, Variant}
+import is.hail.variant.{AltAllele, GenericGenotype, ReferenceGenome, Locus, Variant}
 import org.apache.spark.broadcast.Broadcast
 import org.apache.spark.sql.Row
 
@@ -76,7 +76,7 @@ object UnsafeRow {
     new String(readBinary(region, offset))
 
   def readLocus(region: MemoryBuffer, offset: Long): Locus = {
-    val ft = TLocus(GenomeReference.GRCh37).fundamentalType.asInstanceOf[TStruct]
+    val ft = TLocus(ReferenceGenome.GRCh37).fundamentalType.asInstanceOf[TStruct]
     Locus(
       readString(region, offset + ft.byteOffsets(0)),
       region.loadInt(offset + ft.byteOffsets(1)))

src/main/scala/is/hail/expr/AnnotationImpex.scala

@@ -2,7 +2,7 @@ package is.hail.expr
 
 import is.hail.annotations.Annotation
 import is.hail.utils.{Interval, _}
-import is.hail.variant.{AltAllele, GenomeReference, Genotype, Locus, Variant}
+import is.hail.variant.{AltAllele, ReferenceGenome, Genotype, Locus, Variant}
 import org.apache.spark.sql.Row
 import org.apache.spark.sql.types._
 import org.json4s._
@@ -97,7 +97,7 @@ object SparkAnnotationImpex extends AnnotationImpex[DataType, Any] {
           Locus(r.getAs[String](0), r.getAs[Int](1))
         case x: TInterval =>
           val r = a.asInstanceOf[Row]
-          Interval(importAnnotation(r.get(0), TLocus(x.gr)).asInstanceOf[Locus], importAnnotation(r.get(1), TLocus(x.gr)).asInstanceOf[Locus])
+          Interval(importAnnotation(r.get(0), TLocus(x.rg)).asInstanceOf[Locus], importAnnotation(r.get(1), TLocus(x.rg)).asInstanceOf[Locus])
         case TStruct(fields) =>
           if (fields.isEmpty)
             if (a.asInstanceOf[Boolean]) Annotation.empty else null
@@ -217,10 +217,10 @@ case class JSONExtractInterval(start: Locus, end: Locus) {
 
 case class JSONExtractContig(name: String, length: Int)
 
-case class JSONExtractGenomeReference(name: String, contigs: Array[JSONExtractContig], xContigs: Set[String],
+case class JSONExtractReferenceGenome(name: String, contigs: Array[JSONExtractContig], xContigs: Set[String],
   yContigs: Set[String], mtContigs: Set[String], par: Array[JSONExtractInterval]) {
 
-  def toGenomeReference: GenomeReference = GenomeReference(name, contigs.map(_.name),
+  def toReferenceGenome: ReferenceGenome = ReferenceGenome(name, contigs.map(_.name),
     contigs.map(c => (c.name, c.length)).toMap, xContigs, yContigs, mtContigs, par.map(_.toInterval))
 }