Marmelator/layout-parser
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 289 Wiki Activity

Merge branch 'RED-10270-bp' into 'release/0.159.x'

Browse Source 
RED-10270: fix NumberFormatException

See merge request fforesight/layout-parser!250
This commit is contained in:
Kilian Schüttler 2024-10-24 17:14:51 +02:00
commit e01c0a8d3b
3 changed files with 25 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/SectionIdentifier.java
View File

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

27
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/text/ListIdentifier.java
View File

@ -14,11 +14,12 @@ import lombok.experimental.FieldDefaults;
@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE) @FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
public class ListIdentifier { public class ListIdentifier {
public static final Pattern NUMBER_WITH_DOT = 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]+)\\)\\s+"); public static final Pattern NUMBER_IN_PARENTHESES = Pattern.compile("^\\s*\\(([1-9]{1,4})\\)\\s+");
enum Format { enum Format {
NUMBER_WITH_DOT, NUMBER_IN_PARENTHESES NUMBER_WITH_DOT,
NUMBER_IN_PARENTHESES
} }
Format format; Format format;
@ -48,17 +49,33 @@ public class ListIdentifier {
Matcher numberMatcher = NUMBER_WITH_DOT.matcher(text); Matcher numberMatcher = NUMBER_WITH_DOT.matcher(text);
if (numberMatcher.find()) { 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); Matcher parenthesisMatcher = NUMBER_IN_PARENTHESES.matcher(text);
if (parenthesisMatcher.find()) { 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(); 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) { public static boolean isInOrder(List<ListIdentifier> listIdentifiers) {
if (listIdentifiers.size() <= 1) { if (listIdentifiers.size() <= 1) {

2
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/classification/TableOfContentsClassificationService.java
View File

@ -170,7 +170,7 @@ public class TableOfContentsClassificationService {
Word word = sequences.get(i); Word word = sequences.get(i);
if (!NUMERIC.matcher(word).matches()) { if (!NUMERIC.matcher(word).matches() || word.length() > 5) {
continue; continue;
} }