Corr plot module

plot-base-portable/src/commonMain/kotlin/jetbrains/datalore/plot/base/stat/CorrelationStat.kt

@@ -5,28 +5,27 @@
 
 package jetbrains.datalore.plot.base.stat
 
-import jetbrains.datalore.base.numberFormat.NumberFormat
 import jetbrains.datalore.plot.base.Aes
 import jetbrains.datalore.plot.base.DataFrame
 import jetbrains.datalore.plot.base.StatContext
 import jetbrains.datalore.plot.base.stat.CorrelationUtil.correlationMatrix
 import jetbrains.datalore.plot.base.stat.math3.correlationPearson
 import kotlin.math.abs
 
-
 class CorrelationStat : BaseStat(DEF_MAPPING) {
  var correlationMethod = DEF_CORRELATION_METHOD
  var type = DEF_TYPE
+ var fillDiagonal = DEF_FILL_DIAGONAL
 
  override fun apply(data: DataFrame, statCtx: StatContext, messageConsumer: (s: String) -> Unit): DataFrame {
- if (correlationMethod != Method.PEARSON)
- throw IllegalArgumentException(
-  "Unsupported correlation method: $correlationMethod (only pearson is currently available)"
- )
+ require(correlationMethod == Method.PEARSON) {
+ "Unsupported correlation method: $correlationMethod (only Pearson is currently available)"
+ }
+
 
- val cm = correlationMatrix(data, type, ::correlationPearson)
- val vals = cm.getNumeric(Stats.CORR)
- val abs: List<Double> = vals.map { abs(it!!) }
+ val cm = correlationMatrix(data, type, fillDiagonal, ::correlationPearson)
+ val values = cm.getNumeric(Stats.CORR)
+ val abs: List<Double?> = values.map { it?.let(::abs) }
 
  return cm.builder().putNumeric(Stats.CORR_ABS, abs).build()
  }
@@ -59,6 +58,7 @@ class CorrelationStat : BaseStat(DEF_MAPPING) {
 
  private val DEF_CORRELATION_METHOD = Method.PEARSON
  private val DEF_TYPE = Type.FULL
-
+ const val FILL_DIAGONAL = "fill_diagonal"
+ const val DEF_FILL_DIAGONAL = true
  }
 }

plot-base-portable/src/commonMain/kotlin/jetbrains/datalore/plot/base/stat/CorrelationUtil.kt

@@ -22,19 +22,25 @@ object CorrelationUtil {
  fun correlationMatrix(
  data: DataFrame,
  type: CorrelationStat.Type,
+ fillDiagonal: Boolean,
  corrfn: (DoubleArray, DoubleArray) -> Double
  ): DataFrame {
  val numerics = data.variables().filter { isNumeric(data, it.name) }
 
  val var1: ArrayList<String> = arrayListOf()
  val var2: ArrayList<String> = arrayListOf()
- val corr: ArrayList<Double> = arrayListOf()
+ val corr: ArrayList<Double?> = arrayListOf()
 
  for ((i, vx) in numerics.withIndex()) {
 
  var1.add(vx.label)
  var2.add(vx.label)
- corr.add(1.0)
+
+ if (fillDiagonal) {
+ corr.add(1.0)
+ } else {
+ corr.add(null)
+ }
 
  val xs = data.getNumeric(vx)
 

plot-base-portable/src/commonTest/kotlin/jetbrains/datalore/plot/base/stat/CorrelationTest.kt

@@ -81,20 +81,20 @@ class CorrelationTest {
 
  val labels2series = mapOf("A" to a, "B" to b, "C" to c)
 
- val cm = correlationMatrix(data, CorrelationStat.Type.FULL, ::correlationPearson)
+ val cm = correlationMatrix(data, CorrelationStat.Type.FULL, fillDiagonal = true, corrfn = ::correlationPearson)
 
  assertEquals(cm.rowCount(), a.size.ipow(2).toInt())
 
- val v1 : List<String> = cm[Stats.X] as List<String>
- val v2 : List<String> = cm[Stats.Y] as List<String>
- val cr : List<Double> = cm.getNumeric(Stats.CORR) as List<Double>
+ val v1: List<String> = cm[Stats.X] as List<String>
+ val v2: List<String> = cm[Stats.Y] as List<String>
+ val cr: List<Double> = cm.getNumeric(Stats.CORR) as List<Double>
 
  for ((v12, corr) in v1.zip(v2).zip(cr)) {
  val s1 = labels2series[v12.first]
  val s2 = labels2series[v12.second]
  assertNotNull(s1)
  assertNotNull(s2)
- assertEquals(corr, correlation( s1, s2, ::correlationPearson))
+ assertEquals(corr, correlation(s1, s2, ::correlationPearson))
  }
  }
 }

plot-config-portable/src/commonMain/kotlin/jetbrains/datalore/plot/config/StatProto.kt

@@ -197,6 +197,9 @@ class StatProto {
  }
  }
 
+ stat.fillDiagonal = OptionsAccessor.over(options)
+ .getBoolean(CorrelationStat.FILL_DIAGONAL, def = CorrelationStat.DEF_FILL_DIAGONAL)
+
  return stat
  }
 

python-package/lets_plot/__init__.py

@@ -15,6 +15,7 @@
 from .export import *
 from .frontend_context import *
 from .settings_utils import *
+from .bistro import *
 
 __all__ = (plot.__all__ +
  bistro.__all__ +

python-package/lets_plot/bistro/__init__.py

@@ -2,6 +2,7 @@
 # Use of this source code is governed by the MIT license that can be found in the LICENSE file.
 
 from .im import *
+from .corr import *
 
-__all__ = im.__all__
-
+__all__ = (im.__all__ +
+ corr.__all__)