use of org.apache.spark.sql.types.StructType in project incubator-systemml by apache.
the class MLContextUtil method doesDataFrameLookLikeMatrix.
/**
* Examine the DataFrame schema to determine whether the data appears to be
* a matrix.
*
* @param df
* the DataFrame
* @return {@code true} if the DataFrame appears to be a matrix,
* {@code false} otherwise
*/
public static boolean doesDataFrameLookLikeMatrix(Dataset<Row> df) {
StructType schema = df.schema();
StructField[] fields = schema.fields();
if (fields == null) {
return true;
}
for (StructField field : fields) {
DataType dataType = field.dataType();
if ((dataType != DataTypes.DoubleType) && (dataType != DataTypes.IntegerType) && (dataType != DataTypes.LongType) && (!(dataType instanceof org.apache.spark.ml.linalg.VectorUDT)) && (!(dataType instanceof org.apache.spark.mllib.linalg.VectorUDT))) {
// }
return false;
}
}
return true;
}
use of org.apache.spark.sql.types.StructType in project incubator-systemml by apache.
the class FrameTest method testFrameGeneral.
private void testFrameGeneral(InputInfo iinfo, OutputInfo oinfo, boolean bFromDataFrame, boolean bToDataFrame) throws IOException, DMLException, ParseException {
boolean oldConfig = DMLScript.USE_LOCAL_SPARK_CONFIG;
DMLScript.USE_LOCAL_SPARK_CONFIG = true;
RUNTIME_PLATFORM oldRT = DMLScript.rtplatform;
DMLScript.rtplatform = RUNTIME_PLATFORM.HYBRID_SPARK;
int rowstart = 234, rowend = 1478, colstart = 125, colend = 568;
int bRows = rowend - rowstart + 1, bCols = colend - colstart + 1;
int rowstartC = 124, rowendC = 1178, colstartC = 143, colendC = 368;
int cRows = rowendC - rowstartC + 1, cCols = colendC - colstartC + 1;
HashMap<String, ValueType[]> outputSchema = new HashMap<String, ValueType[]>();
HashMap<String, MatrixCharacteristics> outputMC = new HashMap<String, MatrixCharacteristics>();
TestConfiguration config = getTestConfiguration(TEST_NAME);
loadTestConfiguration(config);
List<String> proArgs = new ArrayList<String>();
proArgs.add(input("A"));
proArgs.add(Integer.toString(rows));
proArgs.add(Integer.toString(cols));
proArgs.add(input("B"));
proArgs.add(Integer.toString(bRows));
proArgs.add(Integer.toString(bCols));
proArgs.add(Integer.toString(rowstart));
proArgs.add(Integer.toString(rowend));
proArgs.add(Integer.toString(colstart));
proArgs.add(Integer.toString(colend));
proArgs.add(output("A"));
proArgs.add(Integer.toString(rowstartC));
proArgs.add(Integer.toString(rowendC));
proArgs.add(Integer.toString(colstartC));
proArgs.add(Integer.toString(colendC));
proArgs.add(output("C"));
fullDMLScriptName = SCRIPT_DIR + TEST_DIR + TEST_NAME + ".dml";
ValueType[] schema = schemaMixedLarge;
// initialize the frame data.
List<ValueType> lschema = Arrays.asList(schema);
fullRScriptName = SCRIPT_DIR + TEST_DIR + TEST_NAME + ".R";
rCmd = "Rscript" + " " + fullRScriptName + " " + inputDir() + " " + rowstart + " " + rowend + " " + colstart + " " + colend + " " + expectedDir() + " " + rowstartC + " " + rowendC + " " + colstartC + " " + colendC;
double sparsity = sparsity1;
double[][] A = getRandomMatrix(rows, cols, min, max, sparsity, 1111);
writeInputFrameWithMTD("A", A, true, schema, oinfo);
sparsity = sparsity2;
double[][] B = getRandomMatrix((int) (bRows), (int) (bCols), min, max, sparsity, 2345);
ValueType[] schemaB = new ValueType[bCols];
for (int i = 0; i < bCols; ++i) schemaB[i] = schema[colstart - 1 + i];
List<ValueType> lschemaB = Arrays.asList(schemaB);
writeInputFrameWithMTD("B", B, true, schemaB, oinfo);
ValueType[] schemaC = new ValueType[colendC - colstartC + 1];
for (int i = 0; i < cCols; ++i) schemaC[i] = schema[colstartC - 1 + i];
Dataset<Row> dfA = null, dfB = null;
if (bFromDataFrame) {
// Create DataFrame for input A
StructType dfSchemaA = FrameRDDConverterUtils.convertFrameSchemaToDFSchema(schema, false);
JavaRDD<Row> rowRDDA = FrameRDDConverterUtils.csvToRowRDD(sc, input("A"), DataExpression.DEFAULT_DELIM_DELIMITER, schema);
dfA = spark.createDataFrame(rowRDDA, dfSchemaA);
// Create DataFrame for input B
StructType dfSchemaB = FrameRDDConverterUtils.convertFrameSchemaToDFSchema(schemaB, false);
JavaRDD<Row> rowRDDB = FrameRDDConverterUtils.csvToRowRDD(sc, input("B"), DataExpression.DEFAULT_DELIM_DELIMITER, schemaB);
dfB = spark.createDataFrame(rowRDDB, dfSchemaB);
}
try {
Script script = ScriptFactory.dmlFromFile(fullDMLScriptName);
String format = "csv";
if (oinfo == OutputInfo.TextCellOutputInfo)
format = "text";
if (bFromDataFrame) {
script.in("A", dfA);
} else {
JavaRDD<String> aIn = sc.textFile(input("A"));
FrameSchema fs = new FrameSchema(lschema);
FrameFormat ff = (format.equals("text")) ? FrameFormat.IJV : FrameFormat.CSV;
FrameMetadata fm = new FrameMetadata(ff, fs, rows, cols);
script.in("A", aIn, fm);
}
if (bFromDataFrame) {
script.in("B", dfB);
} else {
JavaRDD<String> bIn = sc.textFile(input("B"));
FrameSchema fs = new FrameSchema(lschemaB);
FrameFormat ff = (format.equals("text")) ? FrameFormat.IJV : FrameFormat.CSV;
FrameMetadata fm = new FrameMetadata(ff, fs, bRows, bCols);
script.in("B", bIn, fm);
}
// Output one frame to HDFS and get one as RDD //TODO HDFS input/output to do
script.out("A", "C");
// set positional argument values
for (int argNum = 1; argNum <= proArgs.size(); argNum++) {
script.in("$" + argNum, proArgs.get(argNum - 1));
}
MLResults results = ml.execute(script);
format = "csv";
if (iinfo == InputInfo.TextCellInputInfo)
format = "text";
String fName = output("AB");
try {
MapReduceTool.deleteFileIfExistOnHDFS(fName);
} catch (IOException e) {
throw new DMLRuntimeException("Error: While deleting file on HDFS");
}
if (!bToDataFrame) {
if (format.equals("text")) {
JavaRDD<String> javaRDDStringIJV = results.getJavaRDDStringIJV("A");
javaRDDStringIJV.saveAsTextFile(fName);
} else {
JavaRDD<String> javaRDDStringCSV = results.getJavaRDDStringCSV("A");
javaRDDStringCSV.saveAsTextFile(fName);
}
} else {
Dataset<Row> df = results.getDataFrame("A");
// Convert back DataFrame to binary block for comparison using original binary to converted DF and back to binary
MatrixCharacteristics mc = new MatrixCharacteristics(rows, cols, -1, -1, -1);
JavaPairRDD<LongWritable, FrameBlock> rddOut = FrameRDDConverterUtils.dataFrameToBinaryBlock(sc, df, mc, bFromDataFrame).mapToPair(new LongFrameToLongWritableFrameFunction());
rddOut.saveAsHadoopFile(output("AB"), LongWritable.class, FrameBlock.class, OutputInfo.BinaryBlockOutputInfo.outputFormatClass);
}
fName = output("C");
try {
MapReduceTool.deleteFileIfExistOnHDFS(fName);
} catch (IOException e) {
throw new DMLRuntimeException("Error: While deleting file on HDFS");
}
if (!bToDataFrame) {
if (format.equals("text")) {
JavaRDD<String> javaRDDStringIJV = results.getJavaRDDStringIJV("C");
javaRDDStringIJV.saveAsTextFile(fName);
} else {
JavaRDD<String> javaRDDStringCSV = results.getJavaRDDStringCSV("C");
javaRDDStringCSV.saveAsTextFile(fName);
}
} else {
Dataset<Row> df = results.getDataFrame("C");
// Convert back DataFrame to binary block for comparison using original binary to converted DF and back to binary
MatrixCharacteristics mc = new MatrixCharacteristics(cRows, cCols, -1, -1, -1);
JavaPairRDD<LongWritable, FrameBlock> rddOut = FrameRDDConverterUtils.dataFrameToBinaryBlock(sc, df, mc, bFromDataFrame).mapToPair(new LongFrameToLongWritableFrameFunction());
rddOut.saveAsHadoopFile(fName, LongWritable.class, FrameBlock.class, OutputInfo.BinaryBlockOutputInfo.outputFormatClass);
}
runRScript(true);
outputSchema.put("AB", schema);
outputMC.put("AB", new MatrixCharacteristics(rows, cols, -1, -1));
outputSchema.put("C", schemaC);
outputMC.put("C", new MatrixCharacteristics(cRows, cCols, -1, -1));
for (String file : config.getOutputFiles()) {
MatrixCharacteristics md = outputMC.get(file);
FrameBlock frameBlock = readDMLFrameFromHDFS(file, iinfo, md);
FrameBlock frameRBlock = readRFrameFromHDFS(file + ".csv", InputInfo.CSVInputInfo, md);
ValueType[] schemaOut = outputSchema.get(file);
verifyFrameData(frameBlock, frameRBlock, schemaOut);
System.out.println("File " + file + " processed successfully.");
}
System.out.println("Frame MLContext test completed successfully.");
} finally {
DMLScript.rtplatform = oldRT;
DMLScript.USE_LOCAL_SPARK_CONFIG = oldConfig;
}
}
use of org.apache.spark.sql.types.StructType in project incubator-systemml by apache.
the class MLContextTest method testDataFrameSumPYDMLDoublesWithNoIDColumn.
@Test
public void testDataFrameSumPYDMLDoublesWithNoIDColumn() {
System.out.println("MLContextTest - DataFrame sum PYDML, doubles with no ID column");
List<String> list = new ArrayList<String>();
list.add("10,20,30");
list.add("40,50,60");
list.add("70,80,90");
JavaRDD<String> javaRddString = sc.parallelize(list);
JavaRDD<Row> javaRddRow = javaRddString.map(new CommaSeparatedValueStringToDoubleArrayRow());
List<StructField> fields = new ArrayList<StructField>();
fields.add(DataTypes.createStructField("C1", DataTypes.DoubleType, true));
fields.add(DataTypes.createStructField("C2", DataTypes.DoubleType, true));
fields.add(DataTypes.createStructField("C3", DataTypes.DoubleType, true));
StructType schema = DataTypes.createStructType(fields);
Dataset<Row> dataFrame = spark.createDataFrame(javaRddRow, schema);
MatrixMetadata mm = new MatrixMetadata(MatrixFormat.DF_DOUBLES);
Script script = pydml("print('sum: ' + sum(M))").in("M", dataFrame, mm);
setExpectedStdOut("sum: 450.0");
ml.execute(script);
}
use of org.apache.spark.sql.types.StructType in project incubator-systemml by apache.
the class MLContextTest method testDataFrameSumPYDMLMllibVectorWithIDColumn.
@Test
public void testDataFrameSumPYDMLMllibVectorWithIDColumn() {
System.out.println("MLContextTest - DataFrame sum PYDML, mllib vector with ID column");
List<Tuple2<Double, org.apache.spark.mllib.linalg.Vector>> list = new ArrayList<Tuple2<Double, org.apache.spark.mllib.linalg.Vector>>();
list.add(new Tuple2<Double, org.apache.spark.mllib.linalg.Vector>(1.0, org.apache.spark.mllib.linalg.Vectors.dense(1.0, 2.0, 3.0)));
list.add(new Tuple2<Double, org.apache.spark.mllib.linalg.Vector>(2.0, org.apache.spark.mllib.linalg.Vectors.dense(4.0, 5.0, 6.0)));
list.add(new Tuple2<Double, org.apache.spark.mllib.linalg.Vector>(3.0, org.apache.spark.mllib.linalg.Vectors.dense(7.0, 8.0, 9.0)));
JavaRDD<Tuple2<Double, org.apache.spark.mllib.linalg.Vector>> javaRddTuple = sc.parallelize(list);
JavaRDD<Row> javaRddRow = javaRddTuple.map(new DoubleMllibVectorRow());
List<StructField> fields = new ArrayList<StructField>();
fields.add(DataTypes.createStructField(RDDConverterUtils.DF_ID_COLUMN, DataTypes.DoubleType, true));
fields.add(DataTypes.createStructField("C1", new org.apache.spark.mllib.linalg.VectorUDT(), true));
StructType schema = DataTypes.createStructType(fields);
Dataset<Row> dataFrame = spark.createDataFrame(javaRddRow, schema);
MatrixMetadata mm = new MatrixMetadata(MatrixFormat.DF_VECTOR_WITH_INDEX);
Script script = pydml("print('sum: ' + sum(M))").in("M", dataFrame, mm);
setExpectedStdOut("sum: 45.0");
ml.execute(script);
}
use of org.apache.spark.sql.types.StructType in project incubator-systemml by apache.
the class MLContextTest method testInputMatrixBlockPYDML.
@Test
public void testInputMatrixBlockPYDML() {
System.out.println("MLContextTest - input MatrixBlock PYDML");
List<String> list = new ArrayList<String>();
list.add("10,20,30");
list.add("40,50,60");
list.add("70,80,90");
JavaRDD<String> javaRddString = sc.parallelize(list);
JavaRDD<Row> javaRddRow = javaRddString.map(new CommaSeparatedValueStringToRow());
List<StructField> fields = new ArrayList<StructField>();
fields.add(DataTypes.createStructField("C1", DataTypes.StringType, true));
fields.add(DataTypes.createStructField("C2", DataTypes.StringType, true));
fields.add(DataTypes.createStructField("C3", DataTypes.StringType, true));
StructType schema = DataTypes.createStructType(fields);
Dataset<Row> dataFrame = spark.createDataFrame(javaRddRow, schema);
Matrix m = new Matrix(dataFrame);
MatrixBlock matrixBlock = m.toMatrixBlock();
Script script = pydml("avg = avg(M)").in("M", matrixBlock).out("avg");
double avg = ml.execute(script).getDouble("avg");
Assert.assertEquals(50.0, avg, 0.0);
}
Aggregations