Nomad changes

docs/snippets/kotlin/ocr/cloud_ocr_backend.md

@@ -0,0 +1,45 @@
+```kotlin title="Kotlin"
+import dev.kreuzberg.*
+import java.nio.file.Path
+
+class CloudOcrBackend(
+    private val apiKey: String,
+    private val supportedLangs: List<String>,
+) : IOcrBackend {
+
+    override fun name(): String = "cloud-ocr"
+
+    override fun version(): String = "1.0.0"
+
+    override fun process_image(image_bytes: ByteArray, config: OcrConfig): ExtractionResult {
+        val text = callCloudApi(image_bytes, config.language())
+        return ExtractionResult.builder()
+            .withContent(text)
+            .withMimeType("text/plain")
+            .withMetadata(Metadata.builder().build())
+            .build()
+    }
+
+    override fun process_image_file(path: Path, config: OcrConfig): ExtractionResult {
+        return process_image(java.nio.file.Files.readAllBytes(path), config)
+    }
+
+    override fun supports_language(lang: String): Boolean = supportedLangs.contains(lang)
+
+    override fun backend_type(): OcrBackendType = OcrBackendType.Custom
+
+    override fun supported_languages(): List<String> = supportedLangs
+
+    override fun supports_table_detection(): Boolean = false
+
+    override fun supports_document_processing(): Boolean = false
+
+    override fun process_document(_path: Path, _config: OcrConfig): ExtractionResult {
+        throw UnsupportedOperationException("document processing not supported")
+    }
+
+    private fun callCloudApi(image: ByteArray, language: String): String {
+        return "Extracted text"
+    }
+}
+```

docs/snippets/kotlin/ocr/image_extraction.md

@@ -0,0 +1,18 @@
+```kotlin title="Kotlin"
+import dev.kreuzberg.*
+import java.nio.file.Paths
+import java.util.Optional
+
+fun main() {
+    val images = ImageExtractionConfig.builder()
+        .withExtractImages(true)
+        .build()
+
+    val config = ExtractionConfig.builder()
+        .withImages(Optional.of(images))
+        .build()
+
+    val result = Kreuzberg.extractFileSync(Paths.get("document.pdf"), null, config)
+    println("Extracted images: ${result.images()?.size ?: 0}")
+}
+```

docs/snippets/kotlin/ocr/image_preprocessing.md

@@ -0,0 +1,20 @@
+```kotlin title="Kotlin"
+import dev.kreuzberg.*
+import java.nio.file.Paths
+import java.util.Optional
+
+fun main() {
+    val images = ImageExtractionConfig.builder()
+        .withExtractImages(true)
+        .withTargetDpi(300)
+        .withMaxImageDimension(4096)
+        .build()
+
+    val config = ExtractionConfig.builder()
+        .withImages(Optional.of(images))
+        .build()
+
+    val result = Kreuzberg.extractFileSync(Paths.get("document.pdf"), null, config)
+    println("Extracted images: ${result.images()?.size ?: 0}")
+}
+```

docs/snippets/kotlin/ocr/ocr_easyocr.md

@@ -0,0 +1,19 @@
+```kotlin title="Kotlin"
+import dev.kreuzberg.*
+import java.nio.file.Paths
+import java.util.Optional
+
+fun main() {
+    val ocr = OcrConfig.builder()
+        .withBackend("easyocr")
+        .withLanguage("en")
+        .build()
+
+    val config = ExtractionConfig.builder()
+        .withOcr(Optional.of(ocr))
+        .build()
+
+    val result = Kreuzberg.extractFileSync(Paths.get("document.pdf"), null, config)
+    println("Extracted text: ${result.content()}")
+}
+```

docs/snippets/kotlin/ocr/ocr_elements.md

@@ -0,0 +1,31 @@
+```kotlin title="Kotlin"
+import dev.kreuzberg.*
+import java.nio.file.Paths
+import java.util.Optional
+
+fun main() {
+    val elementConfig = OcrElementConfig.builder()
+        .withIncludeElements(true)
+        .build()
+
+    val ocr = OcrConfig.builder()
+        .withBackend("paddleocr")
+        .withLanguage("en")
+        .withElementConfig(Optional.of(elementConfig))
+        .build()
+
+    val config = ExtractionConfig.builder()
+        .withOcr(Optional.of(ocr))
+        .build()
+
+    val result = Kreuzberg.extractFileSync(Paths.get("scanned.pdf"), null, config)
+
+    result.ocrElements()?.forEach { element ->
+        println("Text: ${element.text()}")
+        println("Confidence: ${element.confidence().recognition()}")
+        println("Geometry: ${element.geometry()}")
+        element.rotation()?.let { println("Rotation: ${it}") }
+        println()
+    }
+}
+```

docs/snippets/kotlin/ocr/ocr_extraction.md

@@ -0,0 +1,19 @@
+```kotlin title="Kotlin"
+import dev.kreuzberg.*
+import java.nio.file.Paths
+import java.util.Optional
+
+fun main() {
+    val ocr = OcrConfig.builder()
+        .withBackend("tesseract")
+        .withLanguage("eng")
+        .build()
+
+    val config = ExtractionConfig.builder()
+        .withOcr(Optional.of(ocr))
+        .build()
+
+    val result = Kreuzberg.extractFileSync(Paths.get("scanned.pdf"), null, config)
+    println(result.content())
+}
+```

docs/snippets/kotlin/ocr/ocr_force_all_pages.md

@@ -0,0 +1,20 @@
+```kotlin title="Kotlin"
+import dev.kreuzberg.*
+import java.nio.file.Paths
+import java.util.Optional
+
+fun main() {
+    val ocr = OcrConfig.builder()
+        .withBackend("tesseract")
+        .withLanguage("eng")
+        .build()
+
+    val config = ExtractionConfig.builder()
+        .withOcr(Optional.of(ocr))
+        .withForceOcr(true)
+        .build()
+
+    val result = Kreuzberg.extractFileSync(Paths.get("document.pdf"), null, config)
+    println(result.content())
+}
+```

docs/snippets/kotlin/ocr/ocr_multi_language.md

@@ -0,0 +1,19 @@
+```kotlin title="Kotlin"
+import dev.kreuzberg.*
+import java.nio.file.Paths
+import java.util.Optional
+
+fun main() {
+    val ocr = OcrConfig.builder()
+        .withBackend("tesseract")
+        .withLanguage("eng+deu")
+        .build()
+
+    val config = ExtractionConfig.builder()
+        .withOcr(Optional.of(ocr))
+        .build()
+
+    val result = Kreuzberg.extractFileSync(Paths.get("multilingual.pdf"), null, config)
+    println(result.content())
+}
+```

docs/snippets/kotlin/ocr/ocr_paddleocr.md

@@ -0,0 +1,19 @@
+```kotlin title="Kotlin"
+import dev.kreuzberg.*
+import java.nio.file.Paths
+import java.util.Optional
+
+fun main() {
+    val ocr = OcrConfig.builder()
+        .withBackend("paddleocr")
+        .withLanguage("en")
+        .build()
+
+    val config = ExtractionConfig.builder()
+        .withOcr(Optional.of(ocr))
+        .build()
+
+    val result = Kreuzberg.extractFileSync(Paths.get("document.pdf"), null, config)
+    println("Extracted text: ${result.content()}")
+}
+```