Newer
Older
/*
* Copyright (C) 2024, UChicago Argonne, LLC
* Licensed under the 3-clause BSD license. See accompanying LICENSE.txt file
* in the top-level directory.
*/
package Octeres

Eric Pershey
committed
import Octeres.DataUDF.{LiteBitmaskSchema, intersection_any, intersection_any_v0, intersection_any_v1, liteBitmaskStructToRow, logical_or, validateLiteBitmaskSlotsLike}

Eric Pershey
committed
import org.apache.spark.sql.catalyst.expressions.GenericRowWithSchema
import org.apache.spark.sql.functions._
import org.apache.spark.sql.types._
import org.apache.spark.sql.{Encoders, Row, SparkSession}
import org.scalatest.funsuite.AnyFunSuite
import org.scalatest.prop.TableDrivenPropertyChecks._
import org.scalatest.matchers.should.Matchers._

Eric Pershey
committed
object ExampleUDF {
// very simple UDF function as an example.
case class SimpleAddStruct(a: Int, b: Int)
val SimpleAddSchema: StructType = Encoders.product[SimpleAddStruct].schema
def simpleAdd(a: Int, b: Int): Int = a + b
}
class DataUDFTest extends AnyFunSuite {

Eric Pershey
committed
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
test("test_add") {
val result = ExampleUDF.simpleAdd(1, 2)
assert(result == 3)
}
test("test_simple") {
val sparkSession = SparkSession.builder().appName("DataUDFTest").master("local").getOrCreate()
import sparkSession.implicits._
val deck = Seq(
Row(1, 2),
)
var df = sparkSession.createDataFrame(sparkSession.sparkContext.parallelize(deck), schema = ExampleUDF.SimpleAddSchema)
sparkSession.udf.register("SimpleAdd", ExampleUDF.simpleAdd(_: Int, _: Int), IntegerType)
df = df.withColumn("c", expr("SimpleAdd(a, b)"))
df.show()
}
test("test_logical_or_00") {
println("test_logical_or_00")
val bitmask_length = 8
val a_bitmask = new GenericRowWithSchema(Array(bitmask_length, 3, List(List(0, 2)), 0, 2), LiteBitmaskSchema)
val b_bitmask = new GenericRowWithSchema(Array(bitmask_length, 3, List(List(1, 3)), 1, 3), LiteBitmaskSchema)
println(s"a_bitmask=${a_bitmask}")
println(s"b_bitmask=${b_bitmask}")
println(DataUDF.logical_or(a_bitmask, b_bitmask))
println(DataUDF.logical_or(null, b_bitmask))
println(DataUDF.logical_or(a_bitmask, null))
}
test("test_logical_or_01") {
println("test_logical_or_01")
val bitmask_length = 8
val a_bitmask = new GenericRowWithSchema(Array(bitmask_length, 3, List(List(0, 2)), 0, 2), LiteBitmaskSchema)
val b_bitmask = new GenericRowWithSchema(Array(bitmask_length, 0, List(), -1, -1), LiteBitmaskSchema)
println(s"a_bitmask=${a_bitmask}")
println(s"b_bitmask=${b_bitmask}")
println(DataUDF.logical_or(a_bitmask, b_bitmask))
println(DataUDF.logical_or(null, b_bitmask))
println(DataUDF.logical_or(a_bitmask, null))
println(DataUDF.logical_or(null, null))
println(DataUDF.logical_or(a_bitmask, a_bitmask))
println(DataUDF.logical_or(b_bitmask, b_bitmask))
}
test("test_bitmask") {
println("test_bitmask")
val bitmask_length = 8
// val a_bitmask = new GenericRowWithSchema(Array(bitmask_length, 3, Array(Array(0, 2)), 0, 2), LiteBitmaskSchema)
// val b_bitmask = new GenericRowWithSchema(Array(bitmask_length, 3, Array(Array(1, 3)), 1, 3), LiteBitmaskSchema)
val a_bitmask = new GenericRowWithSchema(Array(bitmask_length, 3, List(List(0, 2)), 0, 2), LiteBitmaskSchema)
val b_bitmask = new GenericRowWithSchema(Array(bitmask_length, 3, List(List(1, 3)), 1, 3), LiteBitmaskSchema)
val c_bitmask = new GenericRowWithSchema(Array(bitmask_length, 5, List(List(0, 4)), 0, 4), LiteBitmaskSchema)
val d_bitmask = new GenericRowWithSchema(Array(bitmask_length, 0, List(), -1, -1), LiteBitmaskSchema)
val e_bitmask = new GenericRowWithSchema(Array(bitmask_length, 6, List(List(0, 5)), 0, 5), LiteBitmaskSchema)
val schema = StructType(List(
StructField("a", IntegerType, nullable = true),
StructField("b", IntegerType, nullable = true),
StructField("bitmask", LiteBitmaskSchema, nullable = true)
))
val deck = Seq(
Row(0, 1, null),
Row(1, 2, a_bitmask),
Row(2, 4, b_bitmask),
Row(3, 8, c_bitmask),
Row(4, 16, d_bitmask),
Row(5, 32, e_bitmask),
)
val sparkSession = SparkSession.builder().appName("DataUDFTest").master("local").getOrCreate()
val df = sparkSession.createDataFrame(sparkSession.sparkContext.parallelize(deck), schema = schema)

Eric Pershey
committed
DataUDF.registerAll(sparkSession)
df.createOrReplaceTempView("temp_view_a")
df.createOrReplaceTempView("temp_view_b")
df.show(32, truncate=false)
sparkSession.sql("select bitmask.intervals from temp_view_a").show(32, truncate=false)
var df2 = sparkSession.sql("Select * from temp_view_a")
// https://spark.apache.org/docs/latest/sql-ref-datatypes.html
df2.show(32, truncate=false)
df2 = df2.withColumn("r_bitmask", expr("logical_or(bitmask, bitmask)"))
df2.show(32, truncate=false)
var df3 = sparkSession.sql("select tva.a, tva.b, " +
"tva.bitmask as a_bitmask, " +
"tvb.bitmask as b_bitmask, " +
"logical_or(tva.bitmask, tvb.bitmask) as c_bitmask " +
// "intersection_any(tva.bitmask, tvb.bitmask) " +
"from temp_view_a tva join temp_view_b tvb"
)
// BROKEN |2 |4 |{8, 3, [[1, 3]], 1, 3}|{8, 6, [[0, 5]], 0, 5}|{8, 0, [], -1, -1} |

Eric Pershey
committed
df3.show(32, truncate=false)
df3.printSchema()
}

Eric Pershey
committed
test("test_bitmask_logical_or") {
val bitmaskTable = Table(
("bitmask0", "bitmask1", "bitmask_or"),
(
DataUDF.getLiteBitmaskRow(8, 3, List(List(0, 2)), 0, 2),
DataUDF.getLiteBitmaskRow(8, 3, List(List(1, 3)), 1, 3),
DataUDF.getLiteBitmaskRow(8, 4, List(List(0, 3)), 0, 3)
), (
DataUDF.getLiteBitmaskRow(8, 5, List(List(0, 4)), 0, 4),
DataUDF.getLiteBitmaskRow(8, 0, List(), -1, -1),
DataUDF.getLiteBitmaskRow(8, 5, List(List(0, 4)), 0, 4)
), (
DataUDF.getLiteBitmaskRow(16, 0, List(), -1, -1),
DataUDF.getLiteBitmaskRow(16, 0, List(), -1, -1),
DataUDF.getLiteBitmaskRow(16, 0, List(), -1, -1)
), (
DataUDF.getLiteBitmaskRow(8, 4, List(List(0, 1), List(3, 4)), 0, 4),
DataUDF.getLiteBitmaskRow(8, 2, List(List(1, 2)), 1, 2),
DataUDF.getLiteBitmaskRow(8, 5, List(List(0, 4)), 0, 4)

Eric Pershey
committed
// ), ( // this will be a problem 8 is the max.
// DataUDF.getLiteBitmaskRow(8, 4, List(List(1, 2), List(5, 6)), 1, 6),
// DataUDF.getLiteBitmaskRow(8, 4, List(List(3, 4), List(7, 8)), 3, 4),
// DataUDF.getLiteBitmaskRow(8, 8, List(List(1, 8)), 1, 8)

Eric Pershey
committed
DataUDF.getLiteBitmaskRow(16, 4, List(List(1, 2), List(5, 6)), 1, 6),
DataUDF.getLiteBitmaskRow(16, 4, List(List(3, 4), List(7, 8)), 3, 8),
DataUDF.getLiteBitmaskRow(16, 8, List(List(1, 8)), 1, 8)
)
)
forAll(bitmaskTable) { (bitmask0, bitmask1, bitmask_or: Row) => {
println(s"bitmask0: ${bitmask0}")
println(s"bitmask1: ${bitmask1}")
println(s"bitmask_or: ${bitmask_or}")
validateLiteBitmaskSlotsLike(bitmask0)
validateLiteBitmaskSlotsLike(bitmask1)
validateLiteBitmaskSlotsLike(bitmask_or)
val bitmask_r = logical_or(bitmask0, bitmask1)
val bitmask_rr = liteBitmaskStructToRow(bitmask_r)
println(s"bitmask_rr: ${bitmask_rr}")
validateLiteBitmaskSlotsLike(bitmask_rr)
bitmask_rr should equal(bitmask_or)

Eric Pershey
committed
println(intersection_any(bitmask0, bitmask1))

Eric Pershey
committed
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
}
test("test_getIntervalBitsGenerator") {
val list_gen = DataUDF.getIntervalBitsGenerator(List(List(1, 2), List(4, 5)), 0, -2)
// println(list_gen.take(1))
// println(list_gen.take(3))
// println(list_gen.take(20))
val bitmaskTable = Table(
("intervals", "bit_idx_first", "bit_idx_last"),
(List(List(1, 2), List(4, 5)), 0, -1),
(List(List(2, 4), List(8, 10)), 4, 9),
(List(List(1, 2), List(5, 6)), 3, 6),
(List(List(3, 4), List(7, 8)), 3, 6)
)
forAll(bitmaskTable) { (intervals, bit_idx_first, bit_idx_last) => {
{
println(intervals)
val ii = new DataUDF.IntervalIterator(intervals, bit_idx_first, bit_idx_last)
ii.printState()
while (!ii.isEmpty) {
println(s"i: ${ii.get()}")
}
}
}}
}
test("test_bitmask_intersection_any") {
val bitmaskTable = Table(
("bitmask0", "bitmask1", "logical_truth"),
(
DataUDF.getLiteBitmaskRow(8, 3, List(List(0, 2)), 0, 2),
DataUDF.getLiteBitmaskRow(8, 3, List(List(1, 3)), 1, 3),
true
), (
DataUDF.getLiteBitmaskRow(8, 5, List(List(0, 4)), 0, 4),
DataUDF.getLiteBitmaskRow(8, 0, List(), -1, -1),
false
), (
DataUDF.getLiteBitmaskRow(16, 0, List(), -1, -1),
DataUDF.getLiteBitmaskRow(16, 0, List(), -1, -1),
false
), (
DataUDF.getLiteBitmaskRow(8, 4, List(List(0, 1), List(3, 4)), 0, 4),
DataUDF.getLiteBitmaskRow(8, 2, List(List(1, 2)), 1, 2),
true
), (
DataUDF.getLiteBitmaskRow(16, 4, List(List(1, 2), List(5, 6)), 1, 6),
DataUDF.getLiteBitmaskRow(16, 4, List(List(3, 4), List(7, 8)), 3, 8),
false
)
)

Eric Pershey
committed
forAll(bitmaskTable) { (bitmask0, bitmask1, logical_truth: Boolean) => {
println(s"bitmask0: ${bitmask0}")
println(s"bitmask1: ${bitmask1}")
validateLiteBitmaskSlotsLike(bitmask0)
validateLiteBitmaskSlotsLike(bitmask1)
val truth = intersection_any(bitmask0, bitmask1)
println(s"truth: ${truth}")
truth should equal(logical_truth)
intersection_any_v0(bitmask0, bitmask1) should equal(intersection_any_v1(bitmask0, bitmask1))
}
}