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

RED-3086: Fixed overlapping rule and dictionary entries

Browse Source
This commit is contained in:
deiflaender 2021-12-15 10:32:40 +01:00
parent 8935af9c76
commit a62c94b344
4 changed files with 11 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/redaction/utils/EntitySearchUtils.java
View File

@ -159,6 +159,8 @@ public class EntitySearchUtils {
// HashSet keeps old value but we want the new. // HashSet keeps old value but we want the new.
entities.removeAll(found); entities.removeAll(found);
entities.addAll(found); entities.addAll(found);
removeEntitiesContainedInLarger(entities);
} }

2
redaction-service-v1/redaction-service-server-v1/src/test/java/com/iqser/red/service/redaction/v1/server/RedactionIntegrationTest.java
View File

@ -673,7 +673,7 @@ public class RedactionIntegrationTest {
public void redactionTest() throws IOException { public void redactionTest() throws IOException {
long start = System.currentTimeMillis(); long start = System.currentTimeMillis();
ClassPathResource pdfFileResource = new ClassPathResource("files/Metolachlor/S-Metolachlor_RAR_01_Volume_1_2018-09-06.pdf"); ClassPathResource pdfFileResource = new ClassPathResource("files/new/just-a-document-20211215.pdf");
AnalyzeRequest request = prepareStorage(pdfFileResource.getInputStream()); AnalyzeRequest request = prepareStorage(pdfFileResource.getInputStream());
request.setExcludedPages(Set.of(1)); request.setExcludedPages(Set.of(1));

8
redaction-service-v1/redaction-service-server-v1/src/test/resources/dictionaries/CBI_address.txt
View File

@ -1655,3 +1655,11 @@ Zyma SA, Nyon, Switzerland
Mambo-Tox Ltd. Biomedical Sciences Building Bassett Crescent East Southampton SO16 7PX UK Mambo-Tox Ltd. Biomedical Sciences Building Bassett Crescent East Southampton SO16 7PX UK
Syngenta Environmental Sciences Jealott’s Hill International Research Centre Bracknell, Berkshire RG42 6EY UK Syngenta Environmental Sciences Jealott’s Hill International Research Centre Bracknell, Berkshire RG42 6EY UK
Test Ignored Hint CBI_ADDRESS Test Ignored Hint CBI_ADDRESS
CTL
CTL - Central Toxicology Laboratory
CTL/P/3557
CTL/P/3607
CTL/P/S 161
CTL/Pl4333
CTL/with dictionary entry

BIN
redaction-service-v1/redaction-service-server-v1/src/test/resources/files/new/just-a-document-20211215.pdf Normal file
View File

Binary file not shown.