Merge branch 'RED-10714-bp' into 'release/0.188.x'

RED-10714: fix IndexOutOfBoundsException

See merge request fforesight/layout-parser!261

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/SectionIdentifier.java

@@ -16,8 +16,8 @@ import lombok.experimental.FieldDefaults;
 @FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
 public class SectionIdentifier {
 
-    public static Pattern numericalIdentifierPattern = Pattern.compile("^[\\s]?(\\d+)[\\s.,;]?(\\d+)?[\\s.,;]?(\\d+)?[\\s.,;]?(\\d+)?");
-    public static Pattern alphanumericIdentifierPattern = Pattern.compile("^[\\s]?[A-Za-z][\\s.,;]?(\\d+)[\\s.,;]?(\\d+)?[\\s.,;]?(\\d+)?[\\s.,;]?(\\d+)?[\\s.,;]?");
+    public static Pattern numericalIdentifierPattern = Pattern.compile("^[\\s]?(\\d{1,2})[\\s.,;]?(\\d{1,2})?[\\s.,;]?(\\d{1,2})?[\\s.,;]?(\\d{1,2})?");
+    public static Pattern alphanumericIdentifierPattern = Pattern.compile("^[\\s]?[A-Za-z][\\s.,;]?(\\d{1,2})[\\s.,;]?(\\d{1,2})?[\\s.,;]?(\\d{1,2})?[\\s.,;]?(\\d{1,2})?[\\s.,;]?");
 
     public enum Format {
         EMPTY,

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/text/ListIdentifier.java

@@ -14,11 +14,12 @@ import lombok.experimental.FieldDefaults;
 @FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
 public class ListIdentifier {
 
-    public static final Pattern NUMBER_WITH_DOT = Pattern.compile("^\\s*([1-9]+)\\.\\s+");
-    public static final Pattern NUMBER_IN_PARENTHESES = Pattern.compile("^\\s*\\(([1-9]+)\\)\\s+");
+    public static final Pattern NUMBER_WITH_DOT = Pattern.compile("^\\s*([1-9]{1,4})\\.\\s+");
+    public static final Pattern NUMBER_IN_PARENTHESES = Pattern.compile("^\\s*\\(([1-9]{1,4})\\)\\s+");
 
     enum Format {
-        NUMBER_WITH_DOT, NUMBER_IN_PARENTHESES
+        NUMBER_WITH_DOT,
+        NUMBER_IN_PARENTHESES
     }
 
     Format format;
@@ -48,17 +49,33 @@ public class ListIdentifier {
         Matcher numberMatcher = NUMBER_WITH_DOT.matcher(text);
 
         if (numberMatcher.find()) {
-            return Optional.of(new ListIdentifier(Format.NUMBER_WITH_DOT, sequences.get(0), page, Integer.parseInt(numberMatcher.group(1))));
+            Optional<Integer> representation = parseInteger(numberMatcher.group(1));
+            if (representation.isPresent()) {
+                return Optional.of(new ListIdentifier(Format.NUMBER_WITH_DOT, sequences.get(0), page, representation.get()));
+            }
         }
 
         Matcher parenthesisMatcher = NUMBER_IN_PARENTHESES.matcher(text);
         if (parenthesisMatcher.find()) {
-            return Optional.of(new ListIdentifier(Format.NUMBER_IN_PARENTHESES, sequences.get(0), page, Integer.parseInt(parenthesisMatcher.group(1))));
+            Optional<Integer> representation = parseInteger(parenthesisMatcher.group(1));
+            if (representation.isPresent()) {
+                return Optional.of(new ListIdentifier(Format.NUMBER_IN_PARENTHESES, sequences.get(0), page, representation.get()));
+            }
         }
         return Optional.empty();
     }
 
 
+    private static Optional<Integer> parseInteger(String text) {
+
+        try {
+            return Optional.of(Integer.parseInt(text));
+        } catch (NumberFormatException e) {
+            return Optional.empty();
+        }
+    }
+
+
     public static boolean isInOrder(List<ListIdentifier> listIdentifiers) {
 
         if (listIdentifiers.size() <= 1) {

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/classification/TableOfContentsClassificationService.java

@@ -69,6 +69,9 @@ public class TableOfContentsClassificationService {
 
     private int identifyTOCItems(int start, List<TextBlockOnPage> textBlocks, ClassificationDocument document) {
 
+        if (start >= textBlocks.size()) {
+            return start;
+        }
         ClassificationPage startPage = textBlocks.get(start).page();
         List<TextBlockOnPage> initialLookAhead = textBlocks.subList(start, Math.min(start + SURROUNDING_BLOCKS_RADIUS, textBlocks.size()));
         HashMap<Word, TextBlockOnPage> lookup = new HashMap<>();
@@ -170,7 +173,7 @@ public class TableOfContentsClassificationService {
 
             Word word = sequences.get(i);
 
-            if (!NUMERIC.matcher(word).matches()) {
+            if (!NUMERIC.matcher(word).matches() || word.length() > 5) {
                 continue;
             }