🐛 harmonize extractors

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

@@ -19,6 +19,7 @@ public class ExtractedImage {
   private final String filename;
   private final String saveFormat;
   private final int pageId;
+  private final int elementId;
 
   /**
    * Default constructor.
@@ -27,11 +28,18 @@ 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, int pageId) {
+  public ExtractedImage(
+      BufferedImage image,
+      String filename,
+      String saveFormat,
+      int pageId,
+      int elementId
+  ) {
     this.image = image;
     this.filename = filename;
     this.saveFormat = saveFormat;
     this.pageId = pageId;
+    this.elementId = elementId;
   }
 
   /**
@@ -53,11 +61,10 @@ public void writeToFile(String outputPath) throws IOException {
    * @throws IOException Throws if the file can't be accessed.
    */
   public void writeToFile(Path outputPath) throws IOException {
-    if (Files.isDirectory(outputPath)) {
-      outputPath = outputPath.resolve(this.filename);
+    if (!Files.isDirectory(outputPath)) {
+      throw new IllegalArgumentException("Provided path is not a directory.");
     }
-    var outputfile = outputPath.toFile();
-    ImageIO.write(this.image, this.saveFormat, outputfile);
+    ImageIO.write(this.image, this.saveFormat, outputPath.resolve(this.filename).toFile());
   }
 
   /**

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

@@ -22,19 +22,17 @@ public class ImageExtractor {
   public ImageExtractor(LocalInputSource source) throws IOException {
 
     this.pageImages = new ArrayList<>();
+    this.filename = source.getFilename();
 
     if (source.isPDF()) {
       this.saveFormat = "jpg";
       var pdfPageImages = getPDFRasterizer().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();
-
       var input = new ByteArrayInputStream(source.getFile());
       this.pageImages.add(ImageIO.read(input));
     }
@@ -64,53 +62,29 @@ public int getPageCount() {
    *
    * @param <FieldT> Type of field (needs to support positioning data).
    * @param fields List of Fields to extract.
-   * @param pageIndex The page index to extract, begins at 0.
+   * @param pageId The page index to extract, begins at 0.
    * @return A list of {@link ExtractedImage}.
    */
   public <FieldT extends PositionDataField> ExtractedImages extractImagesFromPage(
       List<FieldT> fields,
-      int pageIndex
+      int pageId
   ) {
-    return extractImagesFromPage(fields, pageIndex, this.filename);
-  }
-
-  /**
-   * Extract multiple images on a given page from a list of fields having position data.
-   *
-   * @param <FieldT> Type of field (needs to support positioning data).
-   * @param fields List of Fields to extract.
-   * @param pageIndex The page index to extract, begins at 0.
-   * @param outputName The base output filename, must have an image extension.
-   * @return A list of {@link ExtractedImage}.
-   */
-  public <FieldT extends PositionDataField> ExtractedImages extractImagesFromPage(
-      List<FieldT> fields,
-      int pageIndex,
-      String outputName
-  ) {
-    String filename;
-    if (this.getPageCount() > 1) {
-      String[] splitName = InputSourceUtils.splitNameStrict(outputName);
-      filename = splitName[0] + "." + this.saveFormat;
-    } else {
-      filename = outputName;
-    }
-    return extractFromPage(fields, pageIndex, filename);
+    return extractFromPage(fields, pageId, this.filename);
   }
 
   private <FieldT extends PositionDataField> ExtractedImages extractFromPage(
       List<FieldT> fields,
-      int pageIndex,
+      int pageId,
       String outputName
   ) {
-    String[] splitName = InputSourceUtils.splitNameStrict(outputName);
-    var filename = String
-      .format("%s_page-%3s.%s", splitName[0], pageIndex + 1, splitName[1])
-      .replace(" ", "0");
-
     var extractedImages = new ExtractedImages();
-    for (int i = 0; i < fields.size(); i++) {
-      ExtractedImage extractedImage = extractImage(fields.get(i), pageIndex, i + 1, filename);
+    for (int elementId = 0; elementId < fields.size(); elementId++) {
+      ExtractedImage extractedImage = extractImage(
+        fields.get(elementId),
+        pageId,
+        elementId,
+        outputName
+      );
       if (extractedImage != null) {
         extractedImages.add(extractedImage);
       }
@@ -123,33 +97,31 @@ private <FieldT extends PositionDataField> ExtractedImages extractFromPage(
    *
    * @param <FieldT> Type of field (needs to support positioning data).
    * @param field The field to extract.
-   * @param index The index to use for naming the extracted image.
+   * @param elementId The index to use for naming the extracted image.
    * @param filename Name of the file.
-   * @param pageIndex The page index to extract, begins at 0.
+   * @param pageId The page index to extract, begins at 0.
    * @return The {@link ExtractedImage}, or <code>null</code> if the field does not have valid
    * position data.
    */
   public <FieldT extends PositionDataField> ExtractedImage extractImage(
       FieldT field,
-      int pageIndex,
-      int index,
+      int pageId,
+      int elementId,
       String filename
   ) {
     String[] splitName = InputSourceUtils.splitNameStrict(filename);
-    String saveFormat = splitName[1].toLowerCase();
     var polygon = field.getPolygon();
     if (polygon == null) {
       return null;
     }
-    String fieldFilename = splitName[0]
-      + String.format("_%3s", index).replace(" ", "0")
-      + "."
-      + saveFormat;
     return new ExtractedImage(
-      extractImage(polygon.getAsBbox(), pageIndex),
-      fieldFilename,
-      saveFormat,
-      pageIndex
+      extractImage(polygon.getAsBbox(), pageId),
+      String
+        .format("%s_page-%3s-item-%3s.%s", splitName[0], pageId + 1, elementId + 1, this.saveFormat)
+        .replace(" ", "0"),
+      this.saveFormat,
+      pageId,
+      elementId
     );
   }
 
@@ -158,17 +130,17 @@ public <FieldT extends PositionDataField> ExtractedImage extractImage(
    *
    * @param <FieldT> Type of field (needs to support positioning data).
    * @param field The field to extract.
-   * @param index The index to use for naming the extracted image.
-   * @param pageIndex The 0-based page index to extract.
+   * @param elementId The index to use for naming the extracted image.
+   * @param pageId The 0-based page index to extract.
    * @return The {@link ExtractedImage}, or <code>null</code> if the field does not have valid
    * position data.
    */
   public <FieldT extends PositionDataField> ExtractedImage extractImage(
       FieldT field,
-      int pageIndex,
-      int index
+      int pageId,
+      int elementId
   ) {
-    return extractImage(field, pageIndex, index, this.filename);
+    return extractImage(field, pageId, elementId, this.filename);
   }
 
   private BufferedImage extractImage(Bbox bbox, int pageIndex) {

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

@@ -88,9 +88,9 @@ public ExtractedPDFs extractSubDocuments(List<List<Integer>> pageIndexes) throws
   protected String makeFilename(List<Integer> pageNumbers) {
     String[] splitName = InputSourceUtils.splitNameStrict(filename);
     return splitName[0]
-      + String.format("_%3s", pageNumbers.get(0)).replace(" ", "0")
+      + String.format("_pages-%3s", pageNumbers.get(0) + 1).replace(" ", "0")
       + "-"
-      + String.format("%3s", pageNumbers.get(pageNumbers.size() - 1)).replace(" ", "0")
+      + String.format("%3s", pageNumbers.get(pageNumbers.size() - 1) + 1).replace(" ", "0")
       + "."
       + splitName[1];
   }

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

@@ -33,10 +33,11 @@ public ExtractedPDF(byte[] fileBytes, String filename) {
    * @throws IOException Throws if the file can't be accessed.
    */
   public void writeToFile(Path outputPath) throws IOException {
-    if (Files.isDirectory(outputPath)) {
-      outputPath = outputPath.resolve(this.filename);
+    if (!Files.isDirectory(outputPath)) {
+      throw new IllegalArgumentException("Provided path is not a directory.");
     }
-    Files.write(outputPath, this.fileBytes);
+
+    Files.write(outputPath.resolve(this.filename), this.fileBytes);
   }
 
   /**

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

@@ -7,6 +7,9 @@
 import com.mindee.v2.product.crop.CropItem;
 import java.io.IOException;
 import java.util.List;
+import java.util.Map;
+import java.util.stream.Collectors;
+import java.util.stream.IntStream;
 
 public class Crop {
   private final ImageExtractor imageExtractor;
@@ -15,21 +18,41 @@ public Crop(LocalInputSource inputSource) throws IOException {
     this.imageExtractor = new ImageExtractor(inputSource);
   }
 
-  public ExtractedImage extractSingleCrop(CropItem cropItem) throws IOException {
+  public ExtractedImage extractSingleCrop(CropItem cropItem) {
     return this.imageExtractor
       .extractImage(cropItem.getLocation(), cropItem.getLocation().getPage(), 0);
   }
 
   public ExtractedImages extractMultipleCrops(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)
-        );
+    if (cropItems == null || cropItems.isEmpty()) {
+      return new ExtractedImages();
     }
+
+    // Group crops by page, preserving insertion order
+    Map<Integer, List<CropItem>> cropsByPage = cropItems
+      .stream()
+      .collect(
+        Collectors
+          .groupingBy(
+            item -> item.getLocation().getPage(),
+            java.util.LinkedHashMap::new,
+            Collectors.toList()
+          )
+      );
+
+    var extractedImages = new ExtractedImages();
+    cropsByPage
+      .forEach(
+        (page, pageCrops) -> IntStream
+          .range(0, pageCrops.size())
+          .forEach(
+            elementId -> extractedImages
+              .add(
+                this.imageExtractor
+                  .extractImage(pageCrops.get(elementId).getLocation(), page, elementId)
+              )
+          )
+      );
     return extractedImages;
   }
 }

src/test/java/com/mindee/TestingUtilities.java

@@ -9,6 +9,17 @@
 import org.junit.jupiter.api.Assertions;
 
 public class TestingUtilities {
+  public static void deleteRecursively(Path path) throws IOException {
+    if (Files.exists(path)) {
+      try (var entries = Files.walk(path)) {
+        entries
+          .sorted(java.util.Comparator.reverseOrder())
+          .map(Path::toFile)
+          .forEach(java.io.File::delete);
+      }
+    }
+  }
+
   public static Path getResourcePath(String filePath) {
     return Paths.get("src/test/resources/" + filePath);
   }

src/test/java/com/mindee/image/ImageExtractorTest.java

@@ -74,7 +74,9 @@ public void givenAnImage_shouldExtractPositionFields() throws IOException {
         LocalInputSource source = extractedImage.asInputSource();
         Assertions
           .assertEquals(
-            String.format("default_sample_page-001_%3s.jpg", i + 1).replace(" ", "0"),
+            String
+              .format("default_sample_page-%3s-item-%3s.jpg", page.getPageId() + 1, i + 1)
+              .replace(" ", "0"),
             source.getFilename()
           );
       }
@@ -93,28 +95,22 @@ public void givenAnImage_shouldExtractValueFields() throws IOException {
 
     for (Page<BarcodeReaderV1Document> page : inference.getPages()) {
       List<ExtractedImage> codes1D = extractor
-        .extractImagesFromPage(
-          page.getPrediction().getCodes1D(),
-          page.getPageId(),
-          "barcodes_1D.png"
-        );
+        .extractImagesFromPage(page.getPrediction().getCodes1D(), page.getPageId());
       for (int i = 0; i < codes1D.size(); i++) {
         ExtractedImage extractedImage = codes1D.get(i);
         Assertions.assertNotNull(extractedImage.getImage());
         LocalInputSource source = extractedImage.asInputSource();
         Assertions
           .assertEquals(
-            String.format("barcodes_1D_page-001_%3s.png", i + 1).replace(" ", "0"),
+            String
+              .format("default_sample_page-%3s-item-%3s.jpg", page.getPageId() + 1, i + 1)
+              .replace(" ", "0"),
             source.getFilename()
           );
         extractedImage.writeToFile(getResourcePath("output/"));
       }
       List<ExtractedImage> codes2D = extractor
-        .extractImagesFromPage(
-          page.getPrediction().getCodes2D(),
-          page.getPageId(),
-          "barcodes_2D.png"
-        );
+        .extractImagesFromPage(page.getPrediction().getCodes2D(), page.getPageId());
       for (ExtractedImage extractedImage : codes2D) {
         Assertions.assertNotNull(extractedImage.getImage());
         extractedImage.writeToFile(getResourcePath("output/"));

src/test/java/com/mindee/v1/fileoperations/InvoiceSplitterExtractionIT.java

@@ -70,8 +70,10 @@ 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_000-000.pdf", extractedPDFsStrict.get(0).getFilename());
-    Assertions.assertEquals("default_sample_001-001.pdf", extractedPDFsStrict.get(1).getFilename());
+    Assertions
+      .assertEquals("default_sample_pages-001-001.pdf", extractedPDFsStrict.get(0).getFilename());
+    Assertions
+      .assertEquals("default_sample_pages-002-002.pdf", extractedPDFsStrict.get(1).getFilename());
 
     PredictResponse<InvoiceV4> invoice0 = getInvoicePrediction(
       extractedPDFsStrict.get(0).asInputSource()

src/test/java/com/mindee/v1/fileoperations/InvoiceSplitterExtractionTest.java

@@ -31,9 +31,12 @@ public void givenAPDF_shouldExtractInvoicesNoStrict() throws IOException {
     var extractedPDFSNoStrict = extractor
       .extractInvoices(inference.getPrediction().getInvoicePageGroups(), false);
     Assertions.assertEquals(3, extractedPDFSNoStrict.size());
-    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());
+    Assertions
+      .assertEquals("invoice_5p_pages-001-001.pdf", extractedPDFSNoStrict.get(0).getFilename());
+    Assertions
+      .assertEquals("invoice_5p_pages-002-004.pdf", extractedPDFSNoStrict.get(1).getFilename());
+    Assertions
+      .assertEquals("invoice_5p_pages-005-005.pdf", extractedPDFSNoStrict.get(2).getFilename());
   }
 
   @Test
@@ -48,7 +51,9 @@ public void givenAPDF_shouldExtractInvoicesStrict() throws IOException {
     var extractedPDFStrict = extractor
       .extractInvoices(inference.getPrediction().getInvoicePageGroups(), true);
     Assertions.assertEquals(2, extractedPDFStrict.size());
-    Assertions.assertEquals("invoice_5p_000-000.pdf", extractedPDFStrict.get(0).getFilename());
-    Assertions.assertEquals("invoice_5p_001-004.pdf", extractedPDFStrict.get(1).getFilename());
+    Assertions
+      .assertEquals("invoice_5p_pages-001-001.pdf", extractedPDFStrict.get(0).getFilename());
+    Assertions
+      .assertEquals("invoice_5p_pages-002-005.pdf", extractedPDFStrict.get(1).getFilename());
   }
 }