/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/
package hivemall.ftvec.binning;
import hivemall.utils.hadoop.HiveUtils;
import hivemall.utils.hadoop.WritableUtils;
import hivemall.utils.lang.Preconditions;
import org.apache.hadoop.hive.ql.exec.Description;
import org.apache.hadoop.hive.ql.exec.UDFArgumentException;
import org.apache.hadoop.hive.ql.exec.UDFArgumentLengthException;
import org.apache.hadoop.hive.ql.exec.UDFArgumentTypeException;
import org.apache.hadoop.hive.ql.metadata.HiveException;
import org.apache.hadoop.hive.ql.parse.SemanticException;
import org.apache.hadoop.hive.ql.udf.generic.AbstractGenericUDAFResolver;
import org.apache.hadoop.hive.ql.udf.generic.GenericUDAFEvaluator;
import org.apache.hadoop.hive.ql.udf.generic.GenericUDAFParameterInfo;
import org.apache.hadoop.hive.serde2.io.DoubleWritable;
import org.apache.hadoop.hive.serde2.lazybinary.LazyBinaryArray;
import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory;
import org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.StandardListObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.StructField;
import org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.primitive.BooleanObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.primitive.DoubleObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory;
import org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorUtils;
import org.apache.hadoop.hive.serde2.objectinspector.primitive.WritableBooleanObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.primitive.WritableDoubleObjectInspector;
import org.apache.hadoop.io.BooleanWritable;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
import java.util.List;
@Description(name = "build_bins",
value = "_FUNC_(number weight, const int num_of_bins[, const boolean auto_shrink = false])"
+ " - Return quantiles representing bins: array<double>")
public final class BuildBinsUDAF extends AbstractGenericUDAFResolver {
@Override
public GenericUDAFEvaluator getEvaluator(GenericUDAFParameterInfo info)
throws SemanticException {
ObjectInspector[] OIs = info.getParameterObjectInspectors();
if (OIs.length != 2 && OIs.length != 3) {
throw new UDFArgumentLengthException("Specify two or three arguments.");
}
if (!HiveUtils.isNumberOI(OIs[0])) {
throw new UDFArgumentTypeException(0, "Only number type argument is acceptable but "
+ OIs[0].getTypeName() + " was passed as `weight`");
}
if (!HiveUtils.isIntegerOI(OIs[1])) {
throw new UDFArgumentTypeException(1, "Only int type argument is acceptable but "
+ OIs[1].getTypeName() + " was passed as `num_of_bins`");
}
if (OIs.length == 3) {
if (!HiveUtils.isBooleanOI(OIs[2])) {
throw new UDFArgumentTypeException(2,
"Only boolean type argument is acceptable but " + OIs[2].getTypeName()
+ " was passed as `auto_shrink`");
}
}
return new BuildBinsUDAFEvaluator();
}
private static class BuildBinsUDAFEvaluator extends GenericUDAFEvaluator {
// PARTIAL1 and COMPLETE
private PrimitiveObjectInspector weightOI;
// PARTIAL2 and FINAL
private StructObjectInspector structOI;
private StructField autoShrinkField, histogramField, quantilesField;
private BooleanObjectInspector autoShrinkOI;
private StandardListObjectInspector histogramOI;
private DoubleObjectInspector histogramElOI;
private StandardListObjectInspector quantilesOI;
private DoubleObjectInspector quantileOI;
private int nBGBins = 10000; // # of bins for creating histogram (background bins)
private int nBins; // # of bins for result
private boolean autoShrink = false; // default: false
private double[] quantiles; // for reset
@AggregationType(estimable = true)
static final class BuildBinsAggregationBuffer extends AbstractAggregationBuffer {
boolean autoShrink;
NumericHistogram histogram; // histogram used for quantile approximation
double[] quantiles; // the quantiles requested
BuildBinsAggregationBuffer() {}
@Override
public int estimate() {
return (histogram != null ? histogram.lengthFor() : 0) // histogram
+ 20 + 8 * (quantiles != null ? quantiles.length : 0) // quantiles
+ 4; // autoShrink
}
}
@Override
public ObjectInspector init(Mode mode, ObjectInspector[] OIs) throws HiveException {
super.init(mode, OIs);
if (mode == Mode.PARTIAL1 || mode == Mode.COMPLETE) {
weightOI = HiveUtils.asDoubleCompatibleOI(OIs[0]);
// set const values
nBins = HiveUtils.getConstInt(OIs[1]);
if (OIs.length == 3) {
autoShrink = HiveUtils.getConstBoolean(OIs[2]);
}
// check value of `num_of_bins`
if (nBins < 2) {
throw new UDFArgumentException(
"Only greater than or equal to 2 is accepted but " + nBins
+ " was passed as `num_of_bins`.");
}
quantiles = getQuantiles();
} else {
structOI = (StructObjectInspector) OIs[0];
autoShrinkField = structOI.getStructFieldRef("autoShrink");
histogramField = structOI.getStructFieldRef("histogram");
quantilesField = structOI.getStructFieldRef("quantiles");
autoShrinkOI = (WritableBooleanObjectInspector) autoShrinkField.getFieldObjectInspector();
histogramOI = (StandardListObjectInspector) histogramField.getFieldObjectInspector();
quantilesOI = (StandardListObjectInspector) quantilesField.getFieldObjectInspector();
histogramElOI = (WritableDoubleObjectInspector) histogramOI.getListElementObjectInspector();
quantileOI = (WritableDoubleObjectInspector) quantilesOI.getListElementObjectInspector();
}
if (mode == Mode.PARTIAL1 || mode == Mode.PARTIAL2) {
final ArrayList<ObjectInspector> fieldOIs = new ArrayList<ObjectInspector>();
fieldOIs.add(PrimitiveObjectInspectorFactory.writableBooleanObjectInspector);
fieldOIs.add(ObjectInspectorFactory.getStandardListObjectInspector(PrimitiveObjectInspectorFactory.writableDoubleObjectInspector));
fieldOIs.add(ObjectInspectorFactory.getStandardListObjectInspector(PrimitiveObjectInspectorFactory.writableDoubleObjectInspector));
return ObjectInspectorFactory.getStandardStructObjectInspector(
Arrays.asList("autoShrink", "histogram", "quantiles"), fieldOIs);
} else {
return ObjectInspectorFactory.getStandardListObjectInspector(PrimitiveObjectInspectorFactory.writableDoubleObjectInspector);
}
}
private double[] getQuantiles() throws HiveException {
final int nQuantiles = nBins - 1;
final double[] result = new double[nQuantiles];
for (int i = 0; i < nQuantiles; i++) {
result[i] = ((double) (i + 1)) / (nQuantiles + 1);
}
return result;
}
@Override
public AbstractAggregationBuffer getNewAggregationBuffer() throws HiveException {
final BuildBinsAggregationBuffer myAgg = new BuildBinsAggregationBuffer();
myAgg.histogram = new NumericHistogram();
reset(myAgg);
return myAgg;
}
@Override
public void reset(@SuppressWarnings("deprecation") AggregationBuffer agg)
throws HiveException {
final BuildBinsAggregationBuffer myAgg = (BuildBinsAggregationBuffer) agg;
myAgg.autoShrink = autoShrink;
myAgg.histogram.reset();
myAgg.histogram.allocate(nBGBins);
myAgg.quantiles = quantiles;
}
@Override
public void iterate(@SuppressWarnings("deprecation") AggregationBuffer agg,
Object[] parameters) throws HiveException {
Preconditions.checkArgument(parameters.length == 2 || parameters.length == 3);
if (parameters[0] == null || parameters[1] == null) {
return;
}
final BuildBinsAggregationBuffer myAgg = (BuildBinsAggregationBuffer) agg;
// Get and process the current datum
myAgg.histogram.add(PrimitiveObjectInspectorUtils.getDouble(parameters[0], weightOI));
}
@Override
public void merge(@SuppressWarnings("deprecation") AggregationBuffer agg, Object other)
throws HiveException {
if (other == null) {
return;
}
final BuildBinsAggregationBuffer myAgg = (BuildBinsAggregationBuffer) agg;
myAgg.autoShrink = autoShrinkOI.get(structOI.getStructFieldData(other, autoShrinkField));
final List<?> histogram = ((LazyBinaryArray) structOI.getStructFieldData(other,
histogramField)).getList();
myAgg.histogram.merge(histogram, histogramElOI);
final double[] quantiles = HiveUtils.asDoubleArray(
structOI.getStructFieldData(other, quantilesField), quantilesOI, quantileOI);
if (quantiles != null && quantiles.length > 0) {
myAgg.quantiles = quantiles;
}
}
@Override
public Object terminatePartial(@SuppressWarnings("deprecation") AggregationBuffer agg)
throws HiveException {
final BuildBinsAggregationBuffer myAgg = (BuildBinsAggregationBuffer) agg;
final Object[] partialResult = new Object[3];
partialResult[0] = new BooleanWritable(myAgg.autoShrink);
partialResult[1] = myAgg.histogram.serialize();
partialResult[2] = (myAgg.quantiles != null) ? WritableUtils.toWritableList(myAgg.quantiles)
: Collections.singletonList(new DoubleWritable(0));
return partialResult;
}
@Override
public Object terminate(@SuppressWarnings("deprecation") AggregationBuffer agg)
throws HiveException {
final BuildBinsAggregationBuffer myAgg = (BuildBinsAggregationBuffer) agg;
if (myAgg.histogram.getUsedBins() < 1) { // SQL standard - return null for zero elements
return null;
} else {
Preconditions.checkNotNull(myAgg.quantiles);
final List<DoubleWritable> result = new ArrayList<DoubleWritable>();
double prev = Double.NEGATIVE_INFINITY;
result.add(new DoubleWritable(Double.NEGATIVE_INFINITY));
for (int i = 0; i < myAgg.quantiles.length; i++) {
final double val = myAgg.histogram.quantile(myAgg.quantiles[i]);
// check duplication
if (prev == val) {
if (!myAgg.autoShrink) {
throw new HiveException(
"Quantiles were repeated even though `auto_shrink` is false."
+ " Reduce `num_of_bins` or enable `auto_shrink`.");
} // else: skip duplicated quantile
} else {
result.add(new DoubleWritable(val));
prev = val;
}
}
result.add(new DoubleWritable(Double.POSITIVE_INFINITY));
return result;
}
}
}
}