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

RED-10127: Paragraphs with multiple table, appendix, figure can't be headlines

Browse Source
This commit is contained in:
Kilian Schuettler 2024-12-06 13:41:44 +01:00
parent 2fdc53429c
commit fcbc005102
2 changed files with 11 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/classification/ClassificationPatterns.java
View File

@ -18,10 +18,14 @@ public class ClassificationPatterns {
public static final Pattern TABLE_OR_FIGURE_PATTERN = Pattern.compile( public static final Pattern TABLE_OR_FIGURE_HEADLINE_PATTERN = Pattern.compile(
"^\\s*(?:table|continued\\s+table|appendix|figure)\\s+(?:[xvi]+|[a-z0-9]{1,3}(?:\\.[0-9]{1,3})*(?:-[0-9]{1,3})?)\\b", "^\\s*(?:table|continued\\s+table|appendix|figure)\\s+(?:[xvi]+|[a-z0-9]{1,3}(?:\\.[0-9]{1,3})*(?:-[0-9]{1,3})?)\\b",
Pattern.CASE_INSENSITIVE); Pattern.CASE_INSENSITIVE);
public static final Pattern TABLE_MID_SENTENCE_PATTERN = Pattern.compile(
"(?:table|continued\\s+table|appendix|figure)\\s+(?:[xvi]+|[a-z0-9]{1,3}(?:\\.[0-9]{1,3})*(?:-[0-9]{1,3})?)\\b",
Pattern.CASE_INSENSITIVE);
public static final Pattern ALPHANUMERIC = Pattern.compile("[a-zA-Z0-9]"); public static final Pattern ALPHANUMERIC = Pattern.compile("[a-zA-Z0-9]");
public static final Pattern NUMERIC = Pattern.compile("[0-9]+"); public static final Pattern NUMERIC = Pattern.compile("[0-9]+");

8
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/classification/DocuMineClassificationService.java
View File

@ -6,7 +6,8 @@ import static com.knecon.fforesight.service.layoutparser.processor.services.clas
import static com.knecon.fforesight.service.layoutparser.processor.services.classification.ClassificationPatterns.HEADLINE_PATTERN_WITH_SLASHES; import static com.knecon.fforesight.service.layoutparser.processor.services.classification.ClassificationPatterns.HEADLINE_PATTERN_WITH_SLASHES;
import static com.knecon.fforesight.service.layoutparser.processor.services.classification.ClassificationPatterns.HEADLINE_WITH_2_IDENTIFER_PATTERN; import static com.knecon.fforesight.service.layoutparser.processor.services.classification.ClassificationPatterns.HEADLINE_WITH_2_IDENTIFER_PATTERN;
import static com.knecon.fforesight.service.layoutparser.processor.services.classification.ClassificationPatterns.HEADLINE_WITH_SINGLE_IDENTIFER_PATTERN; import static com.knecon.fforesight.service.layoutparser.processor.services.classification.ClassificationPatterns.HEADLINE_WITH_SINGLE_IDENTIFER_PATTERN;
import static com.knecon.fforesight.service.layoutparser.processor.services.classification.ClassificationPatterns.TABLE_OR_FIGURE_PATTERN; import static com.knecon.fforesight.service.layoutparser.processor.services.classification.ClassificationPatterns.TABLE_MID_SENTENCE_PATTERN;
import static com.knecon.fforesight.service.layoutparser.processor.services.classification.ClassificationPatterns.TABLE_OR_FIGURE_HEADLINE_PATTERN;
import java.util.ArrayList; import java.util.ArrayList;
import java.util.Comparator; import java.util.Comparator;
@ -83,7 +84,8 @@ public class DocuMineClassificationService {
Matcher atLeast3Matcher = AT_LEAST_3_CHARS_PATTERN.matcher(textBlock.toString()); Matcher atLeast3Matcher = AT_LEAST_3_CHARS_PATTERN.matcher(textBlock.toString());
Matcher headlineWithSlashesMatcher = HEADLINE_PATTERN_WITH_SLASHES.matcher(textBlock.toString()); Matcher headlineWithSlashesMatcher = HEADLINE_PATTERN_WITH_SLASHES.matcher(textBlock.toString());
Matcher amountMatcher = AMOUNT_PATTERN.matcher(textBlock.toString()); Matcher amountMatcher = AMOUNT_PATTERN.matcher(textBlock.toString());
Matcher tableOrFigureMatcher = TABLE_OR_FIGURE_PATTERN.matcher(textBlock.toString()); Matcher tableOrFigureMatcher = TABLE_OR_FIGURE_HEADLINE_PATTERN.matcher(textBlock.toString());
Matcher tableMidSentenceMatcher = TABLE_MID_SENTENCE_PATTERN.matcher(textBlock.toString());
Matcher headlineWithSingleIdentifierMatcher = HEADLINE_WITH_SINGLE_IDENTIFER_PATTERN.matcher(textBlock.toString()); Matcher headlineWithSingleIdentifierMatcher = HEADLINE_WITH_SINGLE_IDENTIFER_PATTERN.matcher(textBlock.toString());
boolean isAtLeast3Characters = atLeast3Matcher.reset().find(); boolean isAtLeast3Characters = atLeast3Matcher.reset().find();
boolean headlineWithSlashesMatches = headlineWithSlashesMatcher.reset().matches(); boolean headlineWithSlashesMatches = headlineWithSlashesMatcher.reset().matches();
@ -148,6 +150,8 @@ public class DocuMineClassificationService {
&& greaterOrEqualFontThanPageAverage(textBlock, page)// && greaterOrEqualFontThanPageAverage(textBlock, page)//
&& PositionUtils.getApproxLineCount(textBlock) < 2.9// && PositionUtils.getApproxLineCount(textBlock) < 2.9//
&& (tableOrFigureMatcher.reset().find() || (headlineWithSingleIdentifierMatcher.reset().find() && listIdentifiers.isEmpty())) // && (tableOrFigureMatcher.reset().find() || (headlineWithSingleIdentifierMatcher.reset().find() && listIdentifiers.isEmpty())) //
&& tableMidSentenceMatcher.reset().results()
.count() <= 1 //
&& !isAmount// && !isAmount//
&& !headlineWithSlashesMatches) { && !headlineWithSlashesMatches) {