Improve performance

README.md

@@ -122,6 +122,8 @@ In this example, `Ksoup.parseGetRequest` fetches and parses HTML content from Wi
 #### For further documentation, please check here: [Jsoup](https://jsoup.org/)
 
 ### Ksoup vs. Jsoup Performance: Parsing & Selecting 448KB HTML File [test.tx](https://github.com/fleeksoft/ksoup/blob/develop/ksoup-test/testResources/test.txt)
+![Ksoup vs Jsoup](performance1.png)
+
 ![Ksoup vs Jsoup](performance.png)
 
 ## Open source

gradle/libs.versions.toml

@@ -8,9 +8,10 @@ ktor = "3.0.0-rc-1"
 ktor2 = "2.3.12"
 coroutines = "1.8.1"
 kotlinxDatetime = "0.6.1"
-kotlinx-io = "0.5.3"
+kotlinx-io = "0.5.4"
 okio = "3.9.0"
 dokka = "1.9.20"
+kotlinx-benchmark = "0.4.12"
 
 #korlibs = "999.0.0.999" # 999.0.0.999 is local version
 korlibs = "6.0.1"
@@ -48,10 +49,13 @@ stately-concurrency = { module = "co.touchlab:stately-concurrency", version.ref
 jsoup = { module = "org.jsoup:jsoup", version.ref = "jsoup" }
 okio = { module = "com.squareup.okio:okio", version.ref = "okio" }
 okio-nodefilesystem = { module = "com.squareup.okio:okio-nodefilesystem", version.ref = "okio" }
+kotlinx-benchmark-runtime = { module = "org.jetbrains.kotlinx:kotlinx-benchmark-runtime", version.ref = "kotlinx-benchmark" }
 
 [plugins]
 androidLibrary = { id = "com.android.library", version.ref = "agp" }
 kmp = { id = "org.jetbrains.kotlin.multiplatform", version.ref = "kotlin" }
 dokka = { id = "org.jetbrains.dokka", version.ref = "dokka" }
 power-assert = { id = "org.jetbrains.kotlin.plugin.power-assert", version.ref = "kotlin" }
 mavenPublish = { id = "com.vanniktech.maven.publish", version.ref = "mavenPublish" }
+kotlinx-benchmark = { id = "org.jetbrains.kotlinx.benchmark", version.ref = "kotlinx-benchmark" }
+allopen = { id = "org.jetbrains.kotlin.plugin.allopen", version.ref = "kotlin" }

ksoup-benchmark/build.gradle.kts

@@ -0,0 +1,23 @@
+plugins {
+    alias(libs.plugins.kotlinx.benchmark)
+    alias(libs.plugins.allopen)
+}
+
+allOpen {
+    annotation("org.openjdk.jmh.annotations.State")
+}
+
+
+benchmark {
+    targets {
+        register("jvm")
+    }
+
+    configurations {
+        named("main") {
+//            exclude("org.jsoup.parser.JsoupBenchmark")
+//            exclude("com.fleeksoft.ksoup.benchmark.KsoupBenchmark")
+        }
+    }
+
+}

ksoup-benchmark/module.yaml

@@ -0,0 +1,24 @@
+product:
+  type: lib
+  platforms: [ jvm, js, android, linuxX64, linuxArm64, tvosArm64, tvosX64, tvosSimulatorArm64, macosX64, macosArm64, iosArm64, iosSimulatorArm64, iosX64, mingwX64 ]
+
+apply: [ ../common.module-template.yaml ]
+
+aliases:
+  - jvmAndAndroid: [ jvm, android ]
+
+repositories:
+  - mavenLocal
+
+dependencies:
+  - $libs.kotlinx.io
+  - $libs.kotlinx.benchmark.runtime
+#  - com.fleeksoft.ksoup:ksoup-lite:0.1.8
+  - ../ksoup
+
+dependencies@jvm:
+  - $libs.jsoup
+
+settings:
+  kotlin:
+    optIns: [ kotlinx.cinterop.BetaInteropApi, kotlinx.cinterop.UnsafeNumber, kotlinx.cinterop.ExperimentalForeignApi, kotlin.experimental.ExperimentalNativeApi ]

ksoup-benchmark/src@jvm/com/fleeksoft/ksoup/benchmark/KsoupBenchmark.kt

@@ -0,0 +1,49 @@
+package com.fleeksoft.ksoup.benchmark
+
+import com.fleeksoft.ksoup.Ksoup
+import com.fleeksoft.ksoup.nodes.Document
+import com.fleeksoft.ksoup.nodes.Element
+import com.fleeksoft.ksoup.select.Elements
+import com.fleeksoft.ksoup.select.Evaluator
+import kotlinx.benchmark.*
+import kotlinx.io.buffered
+import kotlinx.io.files.Path
+import kotlinx.io.files.SystemFileSystem
+import kotlinx.io.readString
+
+
+@State(Scope.Benchmark)
+@Warmup(iterations = 5)
+@Measurement(iterations = 5, time = 1, timeUnit = BenchmarkTimeUnit.SECONDS)
+class KsoupBenchmark {
+    private lateinit var fileData: String
+    private lateinit var doc1: Document
+
+    @Setup
+    fun setUp() {
+        fileData =
+            SystemFileSystem.source(Path("/Users/sabeeh/IdeaProjects/ksoup-benchmark/ksoup-test/testResources/test.txt")).buffered().readString()
+        doc1 = parseHtml()
+    }
+
+    @Benchmark
+    fun parse() {
+        val doc = parseHtml()
+    }
+
+    @Benchmark
+    fun select() {
+        val doc = parseHtml()
+        doc.getElementsByClass("an-info").mapNotNull { anInfo ->
+            anInfo.parent()?.let { a ->
+                val attr = a.attr("href")
+                if (attr.isEmpty()) return@let null
+
+                attr.substringAfter("/Home/Bangumi/", "")
+                    .takeIf { it.isNotBlank() }
+            }
+        }
+    }
+
+    private fun parseHtml() = Ksoup.parse(fileData)
+}

ksoup-benchmark/src@jvm/org/jsoup/parser/JsoupBenchmark.kt

@@ -0,0 +1,45 @@
+package org.jsoup.parser
+
+import kotlinx.benchmark.*
+import kotlinx.io.buffered
+import kotlinx.io.files.Path
+import kotlinx.io.files.SystemFileSystem
+import kotlinx.io.readString
+import org.jsoup.Jsoup
+import org.jsoup.nodes.Document
+
+@State(Scope.Benchmark)
+@Warmup(iterations = 5)
+@Measurement(iterations = 5, time = 1, timeUnit = BenchmarkTimeUnit.SECONDS)
+class JsoupBenchmark {
+    private lateinit var fileData: String
+    private lateinit var doc1: Document
+
+    @Setup
+    fun setUp() {
+        fileData =
+            SystemFileSystem.source(Path("/Users/sabeeh/IdeaProjects/ksoup-benchmark/ksoup-test/testResources/test.txt")).buffered().readString()
+        doc1 = parseHtml()
+    }
+
+    @Benchmark
+    fun parse() {
+        val doc = parseHtml()
+    }
+
+    @Benchmark
+    fun select() {
+        val doc = parseHtml()
+        doc.getElementsByClass("an-info").mapNotNull { anInfo ->
+            anInfo.parent()?.let { a ->
+                val attr = a.attr("href")
+                if (attr.isEmpty()) return@let null
+
+                attr.substringAfter("/Home/Bangumi/", "")
+                    .takeIf { it.isNotBlank() }
+            }
+        }
+    }
+
+    private fun parseHtml() = Jsoup.parse(fileData)
+}

ksoup-engine-common/src/com/fleeksoft/ksoup/io/Charset.kt

@@ -20,7 +20,7 @@ interface Charset {
             }.isSuccess*/
     }
 
-    fun decode(stringBuilder: StringBuilder, byteArray: ByteArray, start: Int, end: Int): Int
+    fun decode(stringBuilder: StringBuilder, byteArray: ByteArray, start: Int, end: Int = byteArray.size): Int
     fun toByteArray(value: String): ByteArray
 
     fun onlyUtf8(): Boolean = false

ksoup-test/test@jvmAndAndroid/com/fleeksoft/ksoup/PerformanceComparisonTest.kt

@@ -44,6 +44,17 @@ class PerformanceComparisonTest {
         val jsoupParseTimes = mutableListOf<Long>()
         val jsoupSelectTimes = mutableListOf<Long>()
 
+//        warmup
+        repeat(10) {
+            ksoupTest(ksoupParseTimes, ksoupSelectTimes)
+            jsoupTest(jsoupParseTimes, jsoupSelectTimes)
+        }
+
+        ksoupParseTimes.clear()
+        ksoupSelectTimes.clear()
+        jsoupParseTimes.clear()
+        jsoupSelectTimes.clear()
+
         // Perform multiple tests
         repeat(30) {
             ksoupTest(ksoupParseTimes, ksoupSelectTimes)

ksoup/src/com/fleeksoft/ksoup/internal/StringUtil.kt

@@ -199,7 +199,7 @@ public object StringUtil {
     }
 
     public fun inSorted(needle: String, haystack: Array<out String>): Boolean {
-        return haystack.toList().binarySearch(needle) >= 0
+        return haystack.binarySearch(needle) >= 0
     }
 
     /**

ksoup/src/com/fleeksoft/ksoup/nodes/Attribute.kt

@@ -4,6 +4,7 @@ import com.fleeksoft.ksoup.helper.Validate
 import com.fleeksoft.ksoup.internal.StringUtil
 import com.fleeksoft.ksoup.nodes.Document.OutputSettings.Syntax
 import com.fleeksoft.ksoup.ported.KCloneable
+import com.fleeksoft.ksoup.ported.binarySearchBy
 import com.fleeksoft.ksoup.ported.exception.IOException
 import com.fleeksoft.ksoup.ported.exception.SerializationException
 
@@ -351,7 +352,7 @@ public open class Attribute : Map.Entry<String, String?>, KCloneable<Attribute>
          * Checks if this attribute name is defined as a boolean attribute in HTML5
          */
         public fun isBooleanAttribute(key: String): Boolean {
-            return booleanAttributes.toList().binarySearch { it.compareTo(key.lowercase()) } >= 0
+            return booleanAttributes.binarySearchBy { it.compareTo(key.lowercase()) } >= 0
         }
     }
 }

ksoup/src/com/fleeksoft/ksoup/nodes/Element.kt

@@ -17,7 +17,6 @@ import com.fleeksoft.ksoup.select.*
 import kotlin.js.JsName
 import kotlin.jvm.JvmOverloads
 import kotlin.reflect.KClass
-import kotlin.reflect.cast
 
 /**
  * An HTML Element consists of a tag name, attributes, and child nodes (including text nodes and other elements).
@@ -398,8 +397,6 @@ public open class Element : Node {
 
     private inline fun <reified T : Any> filterNodes(clazz: KClass<T>): List<T> {
         return _childNodes.filterIsInstance<T>()
-            .map { clazz.cast(it) }
-            .toList()
     }
 
     /**

ksoup/src/com/fleeksoft/ksoup/nodes/Entities.kt

@@ -69,7 +69,7 @@ public object Entities {
         if (value != null) return value
         val codepoint = extended.codepointForName(name)
         return if (codepoint != empty) {
-            charArrayOf(codepoint.toChar()).concatToString()
+            codepoint.toChar().toString()
         } else {
             emptyName
         }
@@ -420,12 +420,12 @@ public object Entities {
         }
 
         public fun codepointForName(name: String): Int {
-            val index: Int = nameKeys.toList().binarySearch(name)
+            val index: Int = nameKeys.binarySearch(name)
             return if (index >= 0) codeVals[index] else empty
         }
 
         public fun nameForCodepoint(codepoint: Int): String {
-            val index: Int = codeKeys.toList().binarySearch(codepoint)
+            val index: Int = codeKeys.binarySearch(codepoint)
             return if (index >= 0) {
                 // the results are ordered so lower case versions of same codepoint come after uppercase, and we prefer to emit lower
                 // (and binary search for same item with multi results is undefined

ksoup/src/com/fleeksoft/ksoup/nodes/FormElement.kt

@@ -19,7 +19,7 @@ public class FormElement(tag: Tag, baseUri: String?, attributes: Attributes?) :
     private val linkedEls: Elements = Elements()
 
     // contains form submittable elements that were linked during the parse (and due to parse rules, may no longer be a child of this form)
-    private val submittable = QueryParser.parse(StringUtil.join(SharedConstants.FormSubmitTags.toList(), ", "))
+    private val submittable = QueryParser.parse(SharedConstants.FormSubmitTags.joinToString(", "))
 
     /**
      * Get the list of form control elements associated with this form.