perf: faster zip entry extraction

gotson · Aug 26, 2024 · eeb5898 · eeb5898
1 parent 6794e24
commit eeb5898
Show file tree

Hide file tree

Showing 5 changed files with 45 additions and 15 deletions.
diff --git a/komga/build.gradle.kts b/komga/build.gradle.kts
@@ -84,7 +84,7 @@ dependencies {
   implementation("com.appmattus.crypto:cryptohash:0.10.1")
 
   implementation("org.apache.tika:tika-core:2.9.1")
-  implementation("org.apache.commons:commons-compress:1.25.0")
+  implementation("org.apache.commons:commons-compress:1.27.1")
   implementation("com.github.junrar:junrar:7.5.5")
   implementation("com.github.gotson.nightcompress:nightcompress:0.2.0")
   implementation("org.apache.pdfbox:pdfbox:3.0.1")

diff --git a/komga/src/main/kotlin/org/gotson/komga/infrastructure/mediacontainer/divina/ZipExtractor.kt b/komga/src/main/kotlin/org/gotson/komga/infrastructure/mediacontainer/divina/ZipExtractor.kt
@@ -8,6 +8,8 @@ import org.gotson.komga.domain.model.MediaContainerEntry
 import org.gotson.komga.domain.model.MediaType
 import org.gotson.komga.infrastructure.image.ImageAnalyzer
 import org.gotson.komga.infrastructure.mediacontainer.ContentDetector
+import org.gotson.komga.infrastructure.util.getZipEntryBytes
+import org.gotson.komga.infrastructure.util.use
 import org.springframework.stereotype.Service
 import java.nio.file.Path
 
@@ -26,7 +28,7 @@ class ZipExtractor(
     path: Path,
     analyzeDimensions: Boolean,
   ): List<MediaContainerEntry> =
-    ZipFile(path.toFile()).use { zip ->
+    ZipFile.builder().setPath(path).use { zip ->
       zip.entries.toList()
         .filter { !it.isDirectory }
         .map { entry ->
@@ -52,8 +54,5 @@ class ZipExtractor(
   override fun getEntryStream(
     path: Path,
     entryName: String,
-  ): ByteArray =
-    ZipFile(path.toFile()).use { zip ->
-      zip.getInputStream(zip.getEntry(entryName)).use { it.readBytes() }
-    }
+  ): ByteArray = getZipEntryBytes(path, entryName)
 }
diff --git a/komga/src/main/kotlin/org/gotson/komga/infrastructure/mediacontainer/epub/Epub.kt b/komga/src/main/kotlin/org/gotson/komga/infrastructure/mediacontainer/epub/Epub.kt
@@ -2,6 +2,7 @@ package org.gotson.komga.infrastructure.mediacontainer.epub
 
 import org.apache.commons.compress.archivers.zip.ZipFile
 import org.gotson.komga.domain.model.MediaUnsupportedException
+import org.gotson.komga.infrastructure.util.use
 import org.jsoup.Jsoup
 import org.jsoup.nodes.Document
 import org.jsoup.parser.Parser
@@ -16,7 +17,7 @@ data class EpubPackage(
 )
 
 inline fun <R> Path.epub(block: (EpubPackage) -> R): R =
-  ZipFile(this.toFile()).use { zip ->
+  ZipFile.builder().setPath(this).use { zip ->
     val opfFile = zip.getPackagePath()
     val opfDoc = zip.getInputStream(zip.getEntry(opfFile)).use { Jsoup.parse(it, null, "", Parser.xmlParser()) }
     val opfDir = Paths.get(opfFile).parent
@@ -30,7 +31,7 @@ fun ZipFile.getPackagePath(): String =
   }
 
 fun getPackageFile(path: Path): String? =
-  ZipFile(path.toFile()).use { zip ->
+  ZipFile.builder().setPath(path).use { zip ->
     try {
       zip.getInputStream(zip.getEntry(zip.getPackagePath())).reader().use { it.readText() }
     } catch (e: Exception) {

diff --git a/komga/src/main/kotlin/org/gotson/komga/infrastructure/mediacontainer/epub/EpubExtractor.kt b/komga/src/main/kotlin/org/gotson/komga/infrastructure/mediacontainer/epub/EpubExtractor.kt
@@ -2,15 +2,14 @@ package org.gotson.komga.infrastructure.mediacontainer.epub
 
 import io.github.oshai.kotlinlogging.KotlinLogging
 import org.apache.commons.compress.archivers.ArchiveEntry
-import org.apache.commons.compress.archivers.zip.ZipFile
 import org.gotson.komga.domain.model.BookPage
-import org.gotson.komga.domain.model.EntryNotFoundException
 import org.gotson.komga.domain.model.EpubTocEntry
 import org.gotson.komga.domain.model.MediaFile
 import org.gotson.komga.domain.model.R2Locator
 import org.gotson.komga.domain.model.TypedBytes
 import org.gotson.komga.infrastructure.image.ImageAnalyzer
 import org.gotson.komga.infrastructure.mediacontainer.ContentDetector
+import org.gotson.komga.infrastructure.util.getZipEntryBytes
 import org.jsoup.Jsoup
 import org.springframework.beans.factory.annotation.Value
 import org.springframework.stereotype.Service
@@ -35,11 +34,7 @@ class EpubExtractor(
   fun getEntryStream(
     path: Path,
     entryName: String,
-  ): ByteArray =
-    ZipFile(path.toFile()).use { zip ->
-      zip.getEntry(entryName)?.let { entry -> zip.getInputStream(entry).use { it.readBytes() } }
-        ?: throw EntryNotFoundException("Entry does not exist: $entryName")
-    }
+  ): ByteArray = getZipEntryBytes(path, entryName)
 
   fun isEpub(path: Path): Boolean =
     try {

diff --git a/komga/src/main/kotlin/org/gotson/komga/infrastructure/util/ZipFileUtils.kt b/komga/src/main/kotlin/org/gotson/komga/infrastructure/util/ZipFileUtils.kt
@@ -0,0 +1,35 @@
+package org.gotson.komga.infrastructure.util
+
+import org.apache.commons.compress.archivers.zip.ZipFile
+import org.gotson.komga.domain.model.EntryNotFoundException
+import java.nio.file.Path
+
+inline fun <R> ZipFile.Builder.use(block: (ZipFile) -> R) = this.get().use(block)
+
+fun getZipEntryBytes(
+  path: Path,
+  entryName: String,
+): ByteArray {
+  // fast path. Only read central directory record and try to find entry in it
+  val zipBuilder =
+    ZipFile.builder()
+      .setPath(path)
+      .setUseUnicodeExtraFields(true)
+      .setIgnoreLocalFileHeader(true)
+  val bytes = zipBuilder.use { it.getEntryBytes(entryName) }
+  if (bytes != null) return bytes
+
+  // slow path. Entry with that name wasn't in central directory record
+  // Iterate each entry and, if present, set name from Unicode extra field in local file header
+  return zipBuilder.setIgnoreLocalFileHeader(false).use {
+    it.getEntryBytes(entryName)
+      ?: throw EntryNotFoundException("Entry does not exist: $entryName")
+  }
+}
+
+private fun ZipFile.getEntryBytes(entryName: String) =
+  this.use { zip ->
+    zip.getEntry(entryName)?.let { entry ->
+      zip.getInputStream(entry).use { it.readBytes() }
+    }
+  }