Nice interface; works for simple histograms.

Author: jpivarski

histogrammar/sparksql.py

@@ -14,6 +14,9 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import json
+import types
+
 import histogrammar.primitives.average
 import histogrammar.primitives.bag
 import histogrammar.primitives.bin
@@ -29,44 +32,90 @@
 import histogrammar.primitives.sparselybin
 import histogrammar.primitives.stack
 import histogrammar.primitives.sum
+from histogrammar.defs import Factory
 
 def addMethods(df):
-    def histogrammar(self, h):
-        converter = self.df._sc._jvm.org.dianahep.histogrammar.sparksql.pyspark.AggregatorConverter()
-        agg = h._sparksql(self.df._sc._jvm, converter)
-        result = converter.histogrammar(self.df._jdf, agg)
-        return Factory.fromJson(jsonlib.loads(result.toJsonString()))
+    def hg(self, h):
+        converter = self._sc._jvm.org.dianahep.histogrammar.sparksql.pyspark.AggregatorConverter()
+        agg = h._sparksql(self._sc._jvm, converter)
+        result = converter.histogrammar(self._jdf, agg)
+        return Factory.fromJson(json.loads(result.toJsonString()))
 
     def Average(self, quantity):
         return self.histogrammar(histogrammar.primitives.average.Average(quantity))
 
     def Bag(self, quantity, range):
         return self.histogrammar(histogrammar.primitives.bag.Bag(quantity, range))
 
-    # def Bin(self, num, low, high, quantity, value=histogrammar.primitives.count.Count(), underflow=histogrammar.primitives.count.Count(), overflow=histogrammar.primitives.count.Count(), nanflow=histogrammar.primitives.count.Count()):
-    #     return self.histogrammar()
-
-
-
-
-    # df.histogrammar    = types.MethodType(histogrammar, df)
-
-    # hg.Average         = types.MethodType(Average        , df)
-    # hg.Bag             = types.MethodType(Bag            , df)
-    # hg.Bin             = types.MethodType(Bin            , df)
-    # hg.Categorize      = types.MethodType(Categorize     , df)
-    # hg.CentrallyBin    = types.MethodType(CentrallyBin   , df)
-    # hg.Label           = types.MethodType(Label          , df)
-    # hg.UntypedLabel    = types.MethodType(UntypedLabel   , df)
-    # hg.Index           = types.MethodType(Index          , df)
-    # hg.Branch          = types.MethodType(Branch         , df)
-    # hg.Count           = types.MethodType(Count          , df)
-    # hg.Deviate         = types.MethodType(Deviate        , df)
-    # hg.Fraction        = types.MethodType(Fraction       , df)
-    # hg.IrregularlyBin  = types.MethodType(IrregularlyBin , df)
-    # hg.Minimize        = types.MethodType(Minimize       , df)
-    # hg.Maximize        = types.MethodType(Maximize       , df)
-    # hg.Select          = types.MethodType(Select         , df)
-    # hg.SparselyBin     = types.MethodType(SparselyBin    , df)
-    # hg.Stack           = types.MethodType(Stack          , df)
-    # hg.Sum             = types.MethodType(Sum            , df)
+    def Bin(self, num, low, high, quantity, value=histogrammar.primitives.count.Count(), underflow=histogrammar.primitives.count.Count(), overflow=histogrammar.primitives.count.Count(), nanflow=histogrammar.primitives.count.Count()):
+        return self.histogrammar(histogrammar.primitives.bin.Bin(num, low, high, quantity, value, underflow, overflow, nanflow))
+
+    def Categorize(self, quantity, value=histogrammar.primitives.count.Count()):
+        return self.histogrammar(histogrammar.primitives.categorize.Categorize(quantity, value))
+
+    def CentrallyBin(self, bins, quantity, value=histogrammar.primitives.count.Count(), nanflow=histogrammar.primitives.count.Count()):
+        return self.histogrammar(histogrammar.primitives.centrallybin.CentrallyBin(bins, quantity, value, nanflow))
+
+    def Label(self, **pairs):
+        return self.histogrammar(histogrammar.primitives.collection.Label(**pairs))
+
+    def UntypedLabel(self, **pairs):
+        return self.histogrammar(histogrammar.primitives.collection.UntypedLabel(**pairs))
+
+    def Index(self, *values):
+        return self.histogrammar(histogrammar.primitives.collection.Index(*values))
+
+    def Branch(self, *values):
+        return self.histogrammar(histogrammar.primitives.collection.Branch(*values))
+
+    def Count(self):    # TODO: handle transform
+        return self.histogrammar(histogrammar.primitives.count.Count())
+
+    def Deviate(self, quantity):
+        return self.histogrammar(histogrammar.primitives.deviate.Deviate(quantity))
+
+    def Fraction(self, quantity, value=histogrammar.primitives.count.Count()):
+        return self.histogrammar(histogrammar.primitives.fraction.Fraction(quantity, value))
+
+    def IrregularlyBin(self, thresholds, quantity, value=histogrammar.primitives.count.Count(), nanflow=histogrammar.primitives.count.Count()):
+        return self.histogrammar(histogrammar.primitives.irregularlybin.IrregularlyBin(thresholds, quantity, value=histogrammar.primitives.count.Count(), nanflow=histogrammar.primitives.count.Count()))
+
+    def Minimize(self, quantity):
+        return self.histogrammar(histogrammar.primitives.minmax.Minimize(quantity))
+
+    def Maximize(self, quantity):
+        return self.histogrammar(histogrammar.primitives.minmax.Maximize(quantity))
+
+    def Select(self, quantity, cut=histogrammar.primitives.count.Count()):
+        return self.histogrammar(histogrammar.primitives.select.Select(quantity, cut))
+
+    def SparselyBin(self, binWidth, quantity, value=histogrammar.primitives.count.Count(), nanflow=histogrammar.primitives.count.Count(), origin=0.0):
+        return self.histogrammar(histogrammar.primitives.sparselybin.SparselyBin(binWidth, quantity, value, nanflow, origin))
+
+    def Stack(self, bins, quantity, value=histogrammar.primitives.count.Count(), nanflow=histogrammar.primitives.count.Count()):
+        return self.histogrammar(histogrammar.primitives.stack.Stack(bins, quantity, value, nanflow))
+
+    def Sum(self, quantity):
+        return self.histogrammar(histogrammar.primitives.sum.Sum(quantity))
+
+    df.histogrammar = types.MethodType(hg, df)
+
+    df.Average         = types.MethodType(Average        , df)
+    df.Bag             = types.MethodType(Bag            , df)
+    df.Bin             = types.MethodType(Bin            , df)
+    df.Categorize      = types.MethodType(Categorize     , df)
+    df.CentrallyBin    = types.MethodType(CentrallyBin   , df)
+    df.Label           = types.MethodType(Label          , df)
+    df.UntypedLabel    = types.MethodType(UntypedLabel   , df)
+    df.Index           = types.MethodType(Index          , df)
+    df.Branch          = types.MethodType(Branch         , df)
+    df.Count           = types.MethodType(Count          , df)
+    df.Deviate         = types.MethodType(Deviate        , df)
+    df.Fraction        = types.MethodType(Fraction       , df)
+    df.IrregularlyBin  = types.MethodType(IrregularlyBin , df)
+    df.Minimize        = types.MethodType(Minimize       , df)
+    df.Maximize        = types.MethodType(Maximize       , df)
+    df.Select          = types.MethodType(Select         , df)
+    df.SparselyBin     = types.MethodType(SparselyBin    , df)
+    df.Stack           = types.MethodType(Stack          , df)
+    df.Sum             = types.MethodType(Sum            , df)