add v2 benchmark

benchmark/parity/build.gradle.kts

@@ -0,0 +1,19 @@
+plugins {
+    alias(libs.plugins.kotlinJvm)
+}
+
+kotlin {
+    jvmToolchain(21)
+}
+
+dependencies {
+    testImplementation(project(":"))
+    testImplementation(project(":benchmark:shared"))
+    testImplementation(libs.kotlincsv.v1.jvm)
+    testImplementation(libs.bundles.kotest)
+    testImplementation(libs.kotlin.test.junit5)
+}
+
+tasks.withType<Test>().configureEach {
+    useJUnitPlatform()
+}

benchmark/parity/src/test/kotlin/com/jsoizo/kotlincsv/bench/parity/AssertRowsEqual.kt

@@ -0,0 +1,21 @@
+package com.jsoizo.kotlincsv.bench.parity
+
+import kotlin.test.assertEquals
+
+fun assertRowsEqual(expected: List<List<String>>, actual: List<List<String>>) {
+    assertEquals(expected.size, actual.size, "row count mismatch")
+    for (i in expected.indices) {
+        val e = expected[i]
+        val a = actual[i]
+        if (e.size != a.size) {
+            throw AssertionError("row $i col count mismatch: expected=${e.size} actual=${a.size}\n  expected=$e\n  actual=$a")
+        }
+        for (j in e.indices) {
+            if (e[j] != a[j]) {
+                throw AssertionError("row $i col $j mismatch: expected=${e[j].quoted()} actual=${a[j].quoted()}")
+            }
+        }
+    }
+}
+
+private fun String.quoted(): String = "\"" + replace("\\", "\\\\").replace("\"", "\\\"") + "\""

benchmark/parity/src/test/kotlin/com/jsoizo/kotlincsv/bench/parity/ParityFixtures.kt

@@ -0,0 +1,10 @@
+package com.jsoizo.kotlincsv.bench.parity
+
+import com.jsoizo.kotlincsv.bench.shared.CsvDataGen
+import com.jsoizo.kotlincsv.bench.shared.DatasetSpec
+
+object ParityFixtures {
+    val hard: CsvDataGen.Generated by lazy {
+        CsvDataGen.generate(DatasetSpec.HARD, CsvDataGen.DEFAULT_SEED)
+    }
+}

benchmark/parity/src/test/kotlin/com/jsoizo/kotlincsv/bench/parity/ParityHeaderTest.kt

@@ -0,0 +1,32 @@
+package com.jsoizo.kotlincsv.bench.parity
+
+import com.github.doyaaaaaken.kotlincsv.dsl.csvReader as v1csvReader
+import com.jsoizo.kotlincsv.csvReader as v2csvReader
+import com.jsoizo.kotlincsv.reader.withHeader
+import org.junit.jupiter.api.Test
+import kotlin.test.assertEquals
+
+class ParityHeaderTest {
+
+    @Test
+    fun readAllWithHeader_string_v1_v2_parity() {
+        val headers = (0 until ParityFixtures.hard.rows[0].size)
+            .joinToString(",") { "col$it" }
+        val body = ParityFixtures.hard.csvText
+        val text = headers + "\r\n" + body
+
+        val v1: List<Map<String, String>> = v1csvReader().readAllWithHeader(text)
+        val v2Rows: List<Map<String, String>> =
+            v2csvReader().readAll(text).asSequence().withHeader().toList()
+
+        assertEquals(v1.size, v2Rows.size, "row count mismatch")
+        for (i in v1.indices) {
+            val e = v1[i]
+            val a = v2Rows[i]
+            assertEquals(e.keys.toList(), a.keys.toList(), "row $i header keys mismatch")
+            for (k in e.keys) {
+                assertEquals(e[k], a[k], "row $i key '$k' mismatch")
+            }
+        }
+    }
+}

benchmark/parity/src/test/kotlin/com/jsoizo/kotlincsv/bench/parity/ParityReadTest.kt

@@ -0,0 +1,58 @@
+package com.jsoizo.kotlincsv.bench.parity
+
+import com.github.doyaaaaaken.kotlincsv.dsl.csvReader as v1csvReader
+import com.jsoizo.kotlincsv.csvReader as v2csvReader
+import com.jsoizo.kotlincsv.reader.readAll as v2readAll
+import com.jsoizo.kotlincsv.reader.readAllFromFile as v2readAllFromFile
+import org.junit.jupiter.api.AfterAll
+import org.junit.jupiter.api.BeforeAll
+import org.junit.jupiter.api.Test
+import org.junit.jupiter.api.TestInstance
+import java.io.ByteArrayInputStream
+import java.io.File
+
+@TestInstance(TestInstance.Lifecycle.PER_CLASS)
+class ParityReadTest {
+
+    private lateinit var tempFile: File
+
+    @BeforeAll
+    fun setUp() {
+        tempFile = File.createTempFile("parity-hard-", ".csv").apply {
+            writeBytes(ParityFixtures.hard.csvBytes)
+        }
+    }
+
+    @AfterAll
+    fun tearDown() {
+        tempFile.delete()
+    }
+
+    @Test
+    fun readAll_string_v1_v2_parity() {
+        val v1 = v1csvReader().readAll(ParityFixtures.hard.csvText)
+        val v2 = v2csvReader().readAll(ParityFixtures.hard.csvText)
+        assertRowsEqual(v1, v2)
+        assertRowsEqual(ParityFixtures.hard.rows, v2)
+    }
+
+    @Test
+    fun readAll_inputStream_v1_v2_parity() {
+        val v1 = ByteArrayInputStream(ParityFixtures.hard.csvBytes).use {
+            v1csvReader().readAll(it)
+        }
+        val v2 = ByteArrayInputStream(ParityFixtures.hard.csvBytes).use {
+            v2csvReader().v2readAll(it)
+        }
+        assertRowsEqual(v1, v2)
+        assertRowsEqual(ParityFixtures.hard.rows, v2)
+    }
+
+    @Test
+    fun readAll_file_v1_v2_parity() {
+        val v1 = v1csvReader().readAll(tempFile)
+        val v2 = v2csvReader().v2readAllFromFile(tempFile)
+        assertRowsEqual(v1, v2)
+        assertRowsEqual(ParityFixtures.hard.rows, v2)
+    }
+}

benchmark/parity/src/test/kotlin/com/jsoizo/kotlincsv/bench/parity/ParityWriteTest.kt

@@ -0,0 +1,29 @@
+package com.jsoizo.kotlincsv.bench.parity
+
+import com.github.doyaaaaaken.kotlincsv.dsl.csvReader as v1csvReader
+import com.github.doyaaaaaken.kotlincsv.dsl.csvWriter as v1csvWriter
+import com.jsoizo.kotlincsv.csvReader as v2csvReader
+import com.jsoizo.kotlincsv.csvWriter as v2csvWriter
+import com.jsoizo.kotlincsv.writer.write as v2write
+import org.junit.jupiter.api.Test
+import java.io.ByteArrayOutputStream
+
+class ParityWriteTest {
+
+    @Test
+    fun writeAll_outputStream_then_v1_reparse_matches_input() {
+        val rows = ParityFixtures.hard.rows
+
+        val v1Out = ByteArrayOutputStream()
+        v1csvWriter().writeAll(rows, v1Out)
+        val v1Reparsed = v1csvReader().readAll(v1Out.toString(Charsets.UTF_8))
+
+        val v2Out = ByteArrayOutputStream()
+        v2csvWriter().v2write(rows.asSequence(), v2Out, "UTF-8")
+        val v2Reparsed = v2csvReader().readAll(v2Out.toString(Charsets.UTF_8))
+
+        assertRowsEqual(rows, v1Reparsed)
+        assertRowsEqual(rows, v2Reparsed)
+        assertRowsEqual(v1Reparsed, v2Reparsed)
+    }
+}

benchmark/shared/build.gradle.kts

@@ -0,0 +1,7 @@
+plugins {
+    alias(libs.plugins.kotlinJvm)
+}
+
+kotlin {
+    jvmToolchain(21)
+}

benchmark/shared/src/main/kotlin/com/jsoizo/kotlincsv/bench/shared/CsvDataGen.kt

@@ -0,0 +1,108 @@
+package com.jsoizo.kotlincsv.bench.shared
+
+import kotlin.random.Random
+
+object CsvDataGen {
+    const val DEFAULT_SEED: Long = 42L
+
+    private val asciiPrintable: CharArray =
+        (('a'..'z') + ('A'..'Z') + ('0'..'9') + " .-_/".toList()).toCharArray()
+
+    private val hiragana: CharArray = ('ぁ'..'ん').toList().toCharArray()
+
+    data class Generated(
+        val rows: List<List<String>>,
+        val csvText: String,
+        val csvBytes: ByteArray,
+        val stats: DataStats,
+    )
+
+    fun generate(spec: DatasetSpec, seed: Long = DEFAULT_SEED): Generated {
+        val random = Random(seed)
+        val rows = ArrayList<List<String>>(spec.rows)
+        var totalLen = 0L
+        var maxLen = 0
+        var quoteCells = 0
+        var commaCells = 0
+        var newlineCells = 0
+        var multiByteCells = 0
+        for (r in 0 until spec.rows) {
+            val row = ArrayList<String>(spec.cols)
+            for (c in 0 until spec.cols) {
+                val len = 4 + random.nextInt(12) // 4..15
+                val sb = StringBuilder(len)
+                var hasQuote = false
+                var hasComma = false
+                var hasNewline = false
+                var hasMultiByte = false
+                for (i in 0 until len) {
+                    val pickMulti = spec.utf8MultiByteRate > 0.0 && random.nextDouble() < spec.utf8MultiByteRate / 2.0
+                    val ch: Char = when {
+                        pickMulti -> {
+                            hasMultiByte = true
+                            hiragana[random.nextInt(hiragana.size)]
+                        }
+                        else -> asciiPrintable[random.nextInt(asciiPrintable.size)]
+                    }
+                    sb.append(ch)
+                }
+                if (spec.embeddedCommaRate > 0.0 && random.nextDouble() < spec.embeddedCommaRate) {
+                    sb.insert(random.nextInt(sb.length + 1), ',')
+                    hasComma = true
+                }
+                if (spec.embeddedNewlineRate > 0.0 && random.nextDouble() < spec.embeddedNewlineRate) {
+                    sb.insert(random.nextInt(sb.length + 1), '\n')
+                    hasNewline = true
+                }
+                if (spec.quoteRate > 0.0 && random.nextDouble() < spec.quoteRate) {
+                    sb.insert(random.nextInt(sb.length + 1), '"')
+                    hasQuote = true
+                }
+                val cell = sb.toString()
+                row.add(cell)
+                totalLen += cell.length
+                if (cell.length > maxLen) maxLen = cell.length
+                if (hasQuote) quoteCells++
+                if (hasComma) commaCells++
+                if (hasNewline) newlineCells++
+                if (hasMultiByte) multiByteCells++
+            }
+            rows.add(row)
+        }
+        val csvText = encode(rows)
+        val csvBytes = csvText.toByteArray(Charsets.UTF_8)
+        val totalCells = spec.rows.toLong() * spec.cols
+        val stats = DataStats(
+            dataset = spec.name,
+            rows = spec.rows,
+            cols = spec.cols,
+            seed = seed,
+            avgCellLen = if (totalCells == 0L) 0.0 else totalLen.toDouble() / totalCells,
+            maxCellLen = maxLen,
+            quoteRate = if (totalCells == 0L) 0.0 else quoteCells.toDouble() / totalCells,
+            embeddedCommaRate = if (totalCells == 0L) 0.0 else commaCells.toDouble() / totalCells,
+            embeddedNewlineRate = if (totalCells == 0L) 0.0 else newlineCells.toDouble() / totalCells,
+            utf8MultiByteCellRate = if (totalCells == 0L) 0.0 else multiByteCells.toDouble() / totalCells,
+            totalBytes = csvBytes.size.toLong(),
+        )
+        return Generated(rows, csvText, csvBytes, stats)
+    }
+
+    private fun encode(rows: List<List<String>>): String {
+        val sb = StringBuilder()
+        for (row in rows) {
+            var first = true
+            for (cell in row) {
+                if (!first) sb.append(',')
+                first = false
+                if (cell.contains('"') || cell.contains(',') || cell.contains('\n') || cell.contains('\r')) {
+                    sb.append('"').append(cell.replace("\"", "\"\"")).append('"')
+                } else {
+                    sb.append(cell)
+                }
+            }
+            sb.append("\r\n")
+        }
+        return sb.toString()
+    }
+}

benchmark/shared/src/main/kotlin/com/jsoizo/kotlincsv/bench/shared/DataStats.kt

@@ -0,0 +1,31 @@
+package com.jsoizo.kotlincsv.bench.shared
+
+data class DataStats(
+    val dataset: String,
+    val rows: Int,
+    val cols: Int,
+    val seed: Long,
+    val avgCellLen: Double,
+    val maxCellLen: Int,
+    val quoteRate: Double,
+    val embeddedCommaRate: Double,
+    val embeddedNewlineRate: Double,
+    val utf8MultiByteCellRate: Double,
+    val totalBytes: Long,
+) {
+    fun toJson(): String = buildString {
+        append("{")
+        append("\"dataset\":\"").append(dataset).append("\",")
+        append("\"rows\":").append(rows).append(",")
+        append("\"cols\":").append(cols).append(",")
+        append("\"seed\":").append(seed).append(",")
+        append("\"avgCellLen\":").append(avgCellLen).append(",")
+        append("\"maxCellLen\":").append(maxCellLen).append(",")
+        append("\"quoteRate\":").append(quoteRate).append(",")
+        append("\"embeddedCommaRate\":").append(embeddedCommaRate).append(",")
+        append("\"embeddedNewlineRate\":").append(embeddedNewlineRate).append(",")
+        append("\"utf8MultiByteCellRate\":").append(utf8MultiByteCellRate).append(",")
+        append("\"totalBytes\":").append(totalBytes)
+        append("}")
+    }
+}

benchmark/shared/src/main/kotlin/com/jsoizo/kotlincsv/bench/shared/DatasetSpec.kt

@@ -0,0 +1,15 @@
+package com.jsoizo.kotlincsv.bench.shared
+
+enum class DatasetSpec(
+    val rows: Int,
+    val cols: Int,
+    val quoteRate: Double,
+    val embeddedCommaRate: Double,
+    val embeddedNewlineRate: Double,
+    val utf8MultiByteRate: Double,
+) {
+    SMALL(rows = 1_000, cols = 10, quoteRate = 0.0, embeddedCommaRate = 0.0, embeddedNewlineRate = 0.0, utf8MultiByteRate = 0.0),
+    MEDIUM(rows = 100_000, cols = 20, quoteRate = 0.0, embeddedCommaRate = 0.0, embeddedNewlineRate = 0.0, utf8MultiByteRate = 0.0),
+    LARGE(rows = 1_000_000, cols = 10, quoteRate = 0.0, embeddedCommaRate = 0.0, embeddedNewlineRate = 0.0, utf8MultiByteRate = 0.0),
+    HARD(rows = 10_000, cols = 10, quoteRate = 0.30, embeddedCommaRate = 0.10, embeddedNewlineRate = 0.05, utf8MultiByteRate = 0.20),
+}

benchmark/shared/src/main/kotlin/com/jsoizo/kotlincsv/bench/shared/EnvProbe.kt

@@ -0,0 +1,49 @@
+package com.jsoizo.kotlincsv.bench.shared
+
+object EnvProbe {
+    data class Snapshot(
+        val jdkVersion: String,
+        val jdkVendor: String,
+        val osName: String,
+        val osVersion: String,
+        val osArch: String,
+        val availableProcessors: Int,
+        val totalMemoryMB: Long,
+        val kotlinStdlibVersion: String,
+        val benchSide: String,
+        val timestampEpochMs: Long,
+    ) {
+        fun toJson(): String = buildString {
+            append("{")
+            append("\"jdkVersion\":\"").append(escape(jdkVersion)).append("\",")
+            append("\"jdkVendor\":\"").append(escape(jdkVendor)).append("\",")
+            append("\"osName\":\"").append(escape(osName)).append("\",")
+            append("\"osVersion\":\"").append(escape(osVersion)).append("\",")
+            append("\"osArch\":\"").append(escape(osArch)).append("\",")
+            append("\"availableProcessors\":").append(availableProcessors).append(",")
+            append("\"totalMemoryMB\":").append(totalMemoryMB).append(",")
+            append("\"kotlinStdlibVersion\":\"").append(escape(kotlinStdlibVersion)).append("\",")
+            append("\"benchSide\":\"").append(escape(benchSide)).append("\",")
+            append("\"timestampEpochMs\":").append(timestampEpochMs)
+            append("}")
+        }
+
+        private fun escape(s: String): String = s.replace("\\", "\\\\").replace("\"", "\\\"")
+    }
+
+    fun snapshot(benchSide: String): Snapshot {
+        val runtime = Runtime.getRuntime()
+        return Snapshot(
+            jdkVersion = System.getProperty("java.version") ?: "",
+            jdkVendor = System.getProperty("java.vendor") ?: "",
+            osName = System.getProperty("os.name") ?: "",
+            osVersion = System.getProperty("os.version") ?: "",
+            osArch = System.getProperty("os.arch") ?: "",
+            availableProcessors = runtime.availableProcessors(),
+            totalMemoryMB = runtime.totalMemory() / (1024L * 1024L),
+            kotlinStdlibVersion = KotlinVersion.CURRENT.toString(),
+            benchSide = benchSide,
+            timestampEpochMs = System.currentTimeMillis(),
+        )
+    }
+}