Add simple pdf page preview function

modules/extract/src/main/scala/docspell/extract/pdfbox/PdfLoader.scala

@@ -0,0 +1,24 @@
+package docspell.extract.pdfbox
+
+import cats.effect._
+import cats.implicits._
+import fs2.Stream
+
+import org.apache.pdfbox.pdmodel.PDDocument
+
+object PdfLoader {
+
+  private def readBytes1[F[_]: Sync](bytes: Array[Byte]): F[PDDocument] =
+    Sync[F].delay(PDDocument.load(bytes))
+
+  private def closePDDocument[F[_]: Sync](pd: PDDocument): F[Unit] =
+    Sync[F].delay(pd.close())
+
+  def withDocumentBytes[F[_]: Sync, A](pdf: Array[Byte])(f: PDDocument => F[A]): F[A] =
+    Sync[F].bracket(readBytes1(pdf))(f)(pd => closePDDocument(pd))
+
+  def withDocumentStream[F[_]: Sync, A](pdf: Stream[F, Byte])(
+      f: PDDocument => F[A]
+  ): F[A] =
+    pdf.compile.to(Array).flatMap(bytes => withDocumentBytes(bytes)(f))
+}

modules/extract/src/main/scala/docspell/extract/pdfbox/PdfboxPreview.scala

@@ -0,0 +1,54 @@
+package docspell.extract.pdfbox
+
+import java.awt.image.BufferedImage
+import java.awt.image.RenderedImage
+import javax.imageio.ImageIO
+
+import cats.effect._
+import cats.implicits._
+import fs2.Chunk
+import fs2.Stream
+
+import org.apache.commons.io.output.ByteArrayOutputStream
+import org.apache.pdfbox.pdmodel.PDDocument
+import org.apache.pdfbox.rendering.PDFRenderer
+
+trait PdfboxPreview[F[_]] {
+
+  def previewPNG(pdf: Stream[F, Byte]): F[Option[Stream[F, Byte]]]
+
+}
+
+object PdfboxPreview {
+
+  def apply[F[_]: Sync](dpi: Float): F[PdfboxPreview[F]] =
+    Sync[F].pure(new PdfboxPreview[F] {
+
+      def previewImage(pdf: Stream[F, Byte]): F[Option[BufferedImage]] =
+        PdfLoader.withDocumentStream(pdf)(doc => Sync[F].delay(getPageImage(doc, 0, dpi)))
+
+      def previewPNG(pdf: Stream[F, Byte]): F[Option[Stream[F, Byte]]] =
+        previewImage(pdf).map(_.map(pngStream[F]))
+
+    })
+
+  private def getPageImage(
+      pdoc: PDDocument,
+      page: Int,
+      dpi: Float
+  ): Option[BufferedImage] = {
+    val count = pdoc.getNumberOfPages
+    if (count <= 0 || page < 0 || count <= page) None
+    else {
+      val renderer = new PDFRenderer(pdoc)
+      Option(renderer.renderImageWithDPI(page, dpi))
+    }
+  }
+
+  private def pngStream[F[_]](img: RenderedImage): Stream[F, Byte] = {
+    val out = new ByteArrayOutputStream()
+    ImageIO.write(img, "PNG", out)
+    Stream.chunk(Chunk.bytes(out.toByteArray()))
+  }
+
+}