Marmelator/redaction-service
1
0
Fork 0
Code Issues Pull Requests 14 Actions Packages Projects Releases 622 Wiki Activity

RED-7040: Syngenta rule 29 does not work

Browse Source
This commit is contained in:
Yannik Hampe 2023-10-16 12:36:18 +02:00
parent e395620e29
commit 9fcf02df73
2 changed files with 24 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/service/document/EntityCreationService.java
View File

@ -183,6 +183,14 @@ public class EntityCreationService {
return betweenTextRanges(startTextRanges, stopTextRanges, type, entityType, node); return betweenTextRanges(startTextRanges, stopTextRanges, type, entityType, node);
} }
public Stream<TextEntity> shortestBetweenAnyStringIgnoreCase(List<String> starts, List<String> stops, String type, EntityType entityType, SemanticNode node, int limit) {
List<TextRange> startTextRanges = RedactionSearchUtility.findTextRangesByListIgnoreCase(starts, node.getTextBlock());
List<TextRange> stopTextRanges = RedactionSearchUtility.findTextRangesByListIgnoreCase(stops, node.getTextBlock());
return betweenTextRanges(startTextRanges, stopTextRanges, type, entityType, node, limit);
}
public Stream<TextEntity> betweenRegexes(String regexStart, String regexStop, String type, EntityType entityType, SemanticNode node) { public Stream<TextEntity> betweenRegexes(String regexStart, String regexStop, String type, EntityType entityType, SemanticNode node) {
@ -203,14 +211,20 @@ public class EntityCreationService {
return betweenTextRanges(startBoundaries, stopBoundaries, type, entityType, node); return betweenTextRanges(startBoundaries, stopBoundaries, type, entityType, node);
} }
public Stream<TextEntity> betweenTextRanges(List<TextRange> startBoundaries, List<TextRange> stopBoundaries, String type, EntityType entityType, SemanticNode node) { public Stream<TextEntity> betweenTextRanges(List<TextRange> startBoundaries, List<TextRange> stopBoundaries, String type, EntityType entityType, SemanticNode node) {
return betweenTextRanges(startBoundaries, stopBoundaries, type, entityType, node,0);
}
public Stream<TextEntity> betweenTextRanges(List<TextRange> startBoundaries, List<TextRange> stopBoundaries, String type, EntityType entityType, SemanticNode node, int limit) {
if (startBoundaries.isEmpty() || stopBoundaries.isEmpty()) { if (startBoundaries.isEmpty() || stopBoundaries.isEmpty()) {
return Stream.empty(); return Stream.empty();
} }
List<TextRange> entityBoundaries = findNonOverlappingBoundariesBetweenBoundariesWithMinimalDistances(startBoundaries, stopBoundaries); List<TextRange> entityBoundaries = findNonOverlappingBoundariesBetweenBoundariesWithMinimalDistances(startBoundaries, stopBoundaries);
return entityBoundaries.stream() return entityBoundaries.stream()
.filter(range -> (limit == 0 || range.length() <= limit))
.map(textRange -> textRange.trim(node.getTextBlock())) .map(textRange -> textRange.trim(node.getTextBlock()))
.filter(textRange -> isValidEntityTextRange(node.getTextBlock(), textRange)) .filter(textRange -> isValidEntityTextRange(node.getTextBlock(), textRange))
.map(textRange -> byTextRange(textRange, type, entityType, node)) .map(textRange -> byTextRange(textRange, type, entityType, node))

9
redaction-service-v1/redaction-service-server-v1/src/test/resources/drools/all_rules.drl
View File

@ -974,6 +974,15 @@ rule "PII.9.3: Redact between \"AUTHOR(S)\" and \"STUDY COMPLETION DATE\" (verte
.forEach(authorEntity -> authorEntity.apply("PII.9.3", "AUTHOR(S) was found", "Article 39(e)(2) of Regulation (EC) No 178/2002")); .forEach(authorEntity -> authorEntity.apply("PII.9.3", "AUTHOR(S) was found", "Article 39(e)(2) of Regulation (EC) No 178/2002"));
end end
rule "PII.9.4: Redact between \"AUTHOR(S)\" and \"STUDY COMPLETION DATE\" (vertebrate study)"
when
FileAttribute(label == "Vertebrate Study", value.toLowerCase() == "yes")
$document: Document(containsStringIgnoreCase("AUTHOR(S)"), containsStringIgnoreCase("STUDY COMPLETION DATE"))
then
entityCreationService.shortestBetweenAnyStringIgnoreCase(List.of("AUTHOR(S):","AUTHOR(S)"), List.of("STUDY COMPLETION DATE:","STUDY COMPLETION DATE"), "PII", EntityType.ENTITY, $document)
.forEach(authorEntity -> authorEntity.apply("PII.9.4", "AUTHOR(S) was found", "Article 39(e)(2) of Regulation (EC) No 178/2002"));
end
// Rule unit: PII.10 // Rule unit: PII.10
rule "PII.10.0: Redact study director abbreviation" rule "PII.10.0: Redact study director abbreviation"