✨ add local Crop and Split for v2

src/main/java/com/mindee/image/ExtractedImage.java

@@ -17,6 +17,7 @@ public class ExtractedImage {
   private final BufferedImage image;
   private final String filename;
   private final String saveFormat;
+  private final int pageId;
 
   /**
    * Default constructor.
@@ -25,10 +26,11 @@ public class ExtractedImage {
    * @param filename Name of the extracted image.
    * @param saveFormat Format to save the image as, defaults to PNG.
    */
-  public ExtractedImage(BufferedImage image, String filename, String saveFormat) {
+  public ExtractedImage(BufferedImage image, String filename, String saveFormat, int pageId) {
     this.image = image;
     this.filename = filename;
     this.saveFormat = saveFormat;
+    this.pageId = pageId;
   }
 
   /**

src/main/java/com/mindee/image/ExtractedImages.java

@@ -0,0 +1,17 @@
+package com.mindee.image;
+
+import java.io.IOException;
+import java.nio.file.Path;
+import java.util.ArrayList;
+
+public class ExtractedImages extends ArrayList<ExtractedImage> {
+  public void saveAllToDisk(String outputPath) throws IOException {
+    saveAllToDisk(Path.of(outputPath));
+  }
+
+  public void saveAllToDisk(Path outputPath) throws IOException {
+    for (ExtractedImage image : this) {
+      image.writeToFile(outputPath);
+    }
+  }
+}

src/main/java/com/mindee/image/ImageExtractor.java

@@ -25,16 +25,18 @@ public class ImageExtractor {
   private final String saveFormat;
 
   public ImageExtractor(LocalInputSource source) throws IOException {
-    this.filename = source.getFilename();
+
     this.pageImages = new ArrayList<>();
 
     if (source.isPDF()) {
       this.saveFormat = "jpg";
-      var pdfPageImages = pdfToImages(source.getFile(), this.filename);
+      var pdfPageImages = pdfToImages(source.getFile(), source.getFilename());
       for (PDFPageImage pdfPageImage : pdfPageImages) {
         this.pageImages.add(pdfPageImage.getImage());
       }
+      this.filename = source.getFilename() + "." + this.saveFormat;
     } else {
+      this.filename = source.getFilename();
       String[] splitName = InputSourceUtils.splitNameStrict(this.filename);
       this.saveFormat = splitName[1].toLowerCase();
 
@@ -43,7 +45,7 @@ public ImageExtractor(LocalInputSource source) throws IOException {
     }
   }
 
-  public List<PDFPageImage> pdfToImages(byte[] fileBytes, String filename) throws IOException {
+  private List<PDFPageImage> pdfToImages(byte[] fileBytes, String filename) throws IOException {
     PDDocument document = Loader.loadPDF(fileBytes);
     var pdfRenderer = new PDFRenderer(document);
     List<PDFPageImage> pdfPageImages = new ArrayList<>();
@@ -90,7 +92,7 @@ public int getPageCount() {
    * @param pageIndex The page index to extract, begins at 0.
    * @return A list of {@link ExtractedImage}.
    */
-  public <FieldT extends PositionDataField> List<ExtractedImage> extractImagesFromPage(
+  public <FieldT extends PositionDataField> ExtractedImages extractImagesFromPage(
       List<FieldT> fields,
       int pageIndex
   ) {
@@ -106,7 +108,7 @@ public <FieldT extends PositionDataField> List<ExtractedImage> extractImagesFrom
    * @param outputName The base output filename, must have an image extension.
    * @return A list of {@link ExtractedImage}.
    */
-  public <FieldT extends PositionDataField> List<ExtractedImage> extractImagesFromPage(
+  public <FieldT extends PositionDataField> ExtractedImages extractImagesFromPage(
       List<FieldT> fields,
       int pageIndex,
       String outputName
@@ -121,7 +123,7 @@ public <FieldT extends PositionDataField> List<ExtractedImage> extractImagesFrom
     return extractFromPage(fields, pageIndex, filename);
   }
 
-  private <FieldT extends PositionDataField> List<ExtractedImage> extractFromPage(
+  private <FieldT extends PositionDataField> ExtractedImages extractFromPage(
       List<FieldT> fields,
       int pageIndex,
       String outputName
@@ -131,7 +133,7 @@ private <FieldT extends PositionDataField> List<ExtractedImage> extractFromPage(
       .format("%s_page-%3s.%s", splitName[0], pageIndex + 1, splitName[1])
       .replace(" ", "0");
 
-    var extractedImages = new ArrayList<ExtractedImage>();
+    var extractedImages = new ExtractedImages();
     for (int i = 0; i < fields.size(); i++) {
       ExtractedImage extractedImage = extractImage(fields.get(i), pageIndex, i + 1, filename);
       if (extractedImage != null) {
@@ -171,7 +173,8 @@ public <FieldT extends PositionDataField> ExtractedImage extractImage(
     return new ExtractedImage(
       extractImage(polygon.getAsBbox(), pageIndex),
       fieldFilename,
-      saveFormat
+      saveFormat,
+      pageIndex
     );
   }
 

src/main/java/com/mindee/pdf/BasePDFExtractor.java

@@ -7,7 +7,6 @@
 import java.io.ByteArrayInputStream;
 import java.io.ByteArrayOutputStream;
 import java.io.IOException;
-import java.util.ArrayList;
 import java.util.List;
 import javax.imageio.ImageIO;
 import org.apache.pdfbox.Loader;
@@ -56,17 +55,15 @@ public BasePDFExtractor(LocalInputSource source) throws IOException {
     }
   }
 
-  /**
-   * Converts an array to a buffered image.
-   *
-   * @param byteArray Raw byte array.
-   * @return a valid ImageIO buffer.
-   * @throws IOException Throws if the file can't be accessed.
-   */
-  public static BufferedImage byteArrayToBufferedImage(byte[] byteArray) throws IOException {
-    try (ByteArrayInputStream stream = new ByteArrayInputStream(byteArray)) {
-      return ImageIO.read(stream);
+  public ExtractedPDF extractSinglePage(
+      List<Integer> pageNumbers,
+      boolean closeOriginal
+  ) throws IOException {
+    if (pageNumbers.isEmpty()) {
+      throw new MindeeException("Empty indexes not allowed for extraction.");
     }
+    var pdfBytes = createPdfFromExistingPdf(this.sourcePdf, pageNumbers, closeOriginal);
+    return new ExtractedPDF(pdfBytes, makeFilename(pageNumbers));
   }
 
   /**
@@ -76,32 +73,41 @@ public static BufferedImage byteArrayToBufferedImage(byte[] byteArray) throws IO
    * @return A list of extracted files.
    * @throws IOException Throws if the file can't be accessed.
    */
-  public List<ExtractedPDF> extractSubDocuments(
-      List<List<Integer>> pageIndexes
-  ) throws IOException {
-    var extractedPDFs = new ArrayList<ExtractedPDF>();
+  public ExtractedPDFs extractSubDocuments(List<List<Integer>> pageIndexes) throws IOException {
+    var extractedPDFs = new ExtractedPDFs();
 
     for (List<Integer> pageIndexElement : pageIndexes) {
-      if (pageIndexElement.isEmpty()) {
-        throw new MindeeException("Empty indexes not allowed for extraction.");
-      }
-      String[] splitName = InputSourceUtils.splitNameStrict(filename);
-      String fieldFilename = splitName[0]
-        + String.format("_%3s", pageIndexElement.get(0) + 1).replace(" ", "0")
-        + "-"
-        + String
-          .format("%3s", pageIndexElement.get(pageIndexElement.size() - 1) + 1)
-          .replace(" ", "0")
-        + "."
-        + splitName[1];
-      extractedPDFs
-        .add(
-          new ExtractedPDF(mergePdfPages(this.sourcePdf, pageIndexElement, false), fieldFilename)
-        );
+      extractedPDFs.add(extractSinglePage(pageIndexElement, false));
     }
     return extractedPDFs;
   }
 
+  /**
+   * Converts an array to a buffered image.
+   *
+   * @param byteArray Raw byte array.
+   * @return a valid ImageIO buffer.
+   * @throws IOException Throws if the file can't be accessed.
+   */
+  private static BufferedImage byteArrayToBufferedImage(byte[] byteArray) throws IOException {
+    try (ByteArrayInputStream stream = new ByteArrayInputStream(byteArray)) {
+      return ImageIO.read(stream);
+    }
+  }
+
+  /**
+   * Make a nice filename for the split.
+   */
+  private String makeFilename(List<Integer> pageNumbers) {
+    String[] splitName = InputSourceUtils.splitNameStrict(filename);
+    return splitName[0]
+      + String.format("_%3s", pageNumbers.get(0)).replace(" ", "0")
+      + "-"
+      + String.format("%3s", pageNumbers.get(pageNumbers.size() - 1)).replace(" ", "0")
+      + "."
+      + splitName[1];
+  }
+
   private static PDPage clonePage(PDPage page) {
 
     COSDictionary pageDict = page.getCOSObject();
@@ -135,12 +141,4 @@ private static byte[] createPdfFromExistingPdf(
     outputStream.close();
     return output;
   }
-
-  public byte[] mergePdfPages(
-      PDDocument document,
-      List<Integer> pageNumbers,
-      boolean closeOriginal
-  ) throws IOException {
-    return createPdfFromExistingPdf(document, pageNumbers, closeOriginal);
-  }
 }

src/main/java/com/mindee/pdf/ExtractedPDFs.java

@@ -0,0 +1,6 @@
+package com.mindee.pdf;
+
+import java.util.ArrayList;
+
+public class ExtractedPDFs extends ArrayList<ExtractedPDF> {
+}

src/main/java/com/mindee/v2/fileOperations/Crop.java

@@ -0,0 +1,35 @@
+package com.mindee.v2.fileOperations;
+
+import com.mindee.image.ExtractedImage;
+import com.mindee.image.ExtractedImages;
+import com.mindee.image.ImageExtractor;
+import com.mindee.input.LocalInputSource;
+import com.mindee.v2.product.crop.CropItem;
+import java.io.IOException;
+import java.util.List;
+
+public class Crop {
+  private final ImageExtractor imageExtractor;
+
+  public Crop(LocalInputSource inputSource) throws IOException {
+    this.imageExtractor = new ImageExtractor(inputSource);
+  }
+
+  public ExtractedImage extractSingle(CropItem cropItem) throws IOException {
+    return this.imageExtractor
+      .extractImage(cropItem.getLocation(), cropItem.getLocation().getPage(), 0);
+  }
+
+  public ExtractedImages extractMultiple(List<CropItem> cropItems) {
+    var extractedImages = new ExtractedImages();
+    for (int i = 0; i < cropItems.size(); i++) {
+      var cropItem = cropItems.get(i);
+      extractedImages
+        .add(
+          this.imageExtractor
+            .extractImage(cropItem.getLocation(), cropItem.getLocation().getPage(), i + 1)
+        );
+    }
+    return extractedImages;
+  }
+}

src/main/java/com/mindee/v2/fileOperations/Split.java

@@ -0,0 +1,29 @@
+package com.mindee.v2.fileOperations;
+
+import com.mindee.input.LocalInputSource;
+import com.mindee.pdf.BasePDFExtractor;
+import com.mindee.pdf.ExtractedPDF;
+import com.mindee.pdf.ExtractedPDFs;
+import com.mindee.v2.product.split.SplitRange;
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.stream.Collectors;
+
+public class Split {
+  BasePDFExtractor pdfSplitter;
+
+  public Split(LocalInputSource inputSource) throws IOException {
+    this.pdfSplitter = new BasePDFExtractor(inputSource);
+  }
+
+  public ExtractedPDF extractSingle(SplitRange splitRange) throws IOException {
+    return this.pdfSplitter.extractSinglePage(splitRange.getPageRangeDistinct(), true);
+  }
+
+  public ExtractedPDFs extractMultiple(ArrayList<SplitRange> splitRanges) throws IOException {
+    return this.pdfSplitter
+      .extractSubDocuments(
+        splitRanges.stream().map(SplitRange::getPageRangeDistinct).collect(Collectors.toList())
+      );
+  }
+}

src/main/java/com/mindee/v2/product/split/SplitRange.java

@@ -3,6 +3,8 @@
 import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
 import com.fasterxml.jackson.annotation.JsonProperty;
 import java.util.ArrayList;
+import java.util.LinkedHashSet;
+import java.util.List;
 import lombok.AllArgsConstructor;
 import lombok.EqualsAndHashCode;
 import lombok.Getter;
@@ -22,11 +24,18 @@ public class SplitRange {
    * indicates the end page.
    */
   @JsonProperty("page_range")
-  public ArrayList<Integer> pageRange;
+  public List<Integer> pageRange;
 
   /**
    * The document type, as identified on given classification values.
    */
   @JsonProperty("document_type")
   public String documentType;
+
+  /**
+   * Returns a list of distinct page ranges. Useful for extracting pages from a document.
+   */
+  public List<Integer> getPageRangeDistinct() {
+    return new ArrayList<>(new LinkedHashSet<>(this.pageRange));
+  }
 }

src/test/java/com/mindee/v1/fileOperation/InvoiceSplitterAutoExtractionIT.java → src/test/java/com/mindee/v1/fileOperations/InvoiceSplitterAutoExtractionIT.java

@@ -1,4 +1,4 @@
-package com.mindee.v1.fileOperation;
+package com.mindee.v1.fileOperations;
 
 import static com.mindee.TestingUtilities.getV1ResourcePath;
 import static com.mindee.TestingUtilities.levenshteinRatio;
@@ -71,8 +71,8 @@ public void givenAPDF_shouldExtractInvoices() throws IOException, InterruptedExc
     List<ExtractedPDF> extractedPDFsStrict = extractor
       .extractInvoices(inference.getPrediction().getInvoicePageGroups(), false);
     Assertions.assertEquals(2, extractedPDFsStrict.size());
-    Assertions.assertEquals("default_sample_001-001.pdf", extractedPDFsStrict.get(0).getFilename());
-    Assertions.assertEquals("default_sample_002-002.pdf", extractedPDFsStrict.get(1).getFilename());
+    Assertions.assertEquals("default_sample_000-000.pdf", extractedPDFsStrict.get(0).getFilename());
+    Assertions.assertEquals("default_sample_001-001.pdf", extractedPDFsStrict.get(1).getFilename());
 
     PredictResponse<InvoiceV4> invoice0 = getInvoicePrediction(
       extractedPDFsStrict.get(0).asInputSource()

src/test/java/com/mindee/v1/pdf/PDFExtractorTest.java

@@ -31,9 +31,9 @@ public void givenAPDF_shouldExtractInvoicesNoStrict() throws IOException {
     var extractedPDFSNoStrict = extractor
       .extractInvoices(inference.getPrediction().getInvoicePageGroups(), false);
     Assertions.assertEquals(3, extractedPDFSNoStrict.size());
-    Assertions.assertEquals("invoice_5p_001-001.pdf", extractedPDFSNoStrict.get(0).getFilename());
-    Assertions.assertEquals("invoice_5p_002-004.pdf", extractedPDFSNoStrict.get(1).getFilename());
-    Assertions.assertEquals("invoice_5p_005-005.pdf", extractedPDFSNoStrict.get(2).getFilename());
+    Assertions.assertEquals("invoice_5p_000-000.pdf", extractedPDFSNoStrict.get(0).getFilename());
+    Assertions.assertEquals("invoice_5p_001-003.pdf", extractedPDFSNoStrict.get(1).getFilename());
+    Assertions.assertEquals("invoice_5p_004-004.pdf", extractedPDFSNoStrict.get(2).getFilename());
   }
 
   @Test
@@ -48,7 +48,7 @@ public void givenAPDF_shouldExtractInvoicesStrict() throws IOException {
     var extractedPDFStrict = extractor
       .extractInvoices(inference.getPrediction().getInvoicePageGroups(), true);
     Assertions.assertEquals(2, extractedPDFStrict.size());
-    Assertions.assertEquals("invoice_5p_001-001.pdf", extractedPDFStrict.get(0).getFilename());
-    Assertions.assertEquals("invoice_5p_002-005.pdf", extractedPDFStrict.get(1).getFilename());
+    Assertions.assertEquals("invoice_5p_000-000.pdf", extractedPDFStrict.get(0).getFilename());
+    Assertions.assertEquals("invoice_5p_001-004.pdf", extractedPDFStrict.get(1).getFilename());
   }
 }