Merge branch 'RED-5223' into 'main'

RED-5223: Use tracing-commons from fforesight

See merge request fforesight/layout-parser!90

.gitlab-ci.yml

@@ -10,7 +10,7 @@ deploy:
   script:
     - echo "Building with gradle version ${BUILDVERSION}"
     - gradle -Pversion=${BUILDVERSION} publish
-    - gradle bootBuildImage --cleanCache --publishImage -PbuildbootDockerHostNetwork=true -Pversion=${BUILDVERSION}
+    - gradle bootBuildImage --publishImage -PbuildbootDockerHostNetwork=true -Pversion=${BUILDVERSION}
     - echo "BUILDVERSION=$BUILDVERSION" >> version.env
   artifacts:
     reports:

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

@@ -244,9 +244,9 @@ public class LayoutParsingPipeline {
 
     private void increaseDocumentStatistics(ClassificationPage classificationPage, ClassificationDocument document) {
 
-        //        if (!classificationPage.isLandscape()) {
+       if (!classificationPage.isLandscape()) {
         document.getFontSizeCounter().addAll(classificationPage.getFontSizeCounter().getCountPerValue());
-//        }
+      }
         document.getFontCounter().addAll(classificationPage.getFontCounter().getCountPerValue());
         document.getTextHeightCounter().addAll(classificationPage.getTextHeightCounter().getCountPerValue());
         document.getFontStyleCounter().addAll(classificationPage.getFontStyleCounter().getCountPerValue());

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/python_api/adapter/CvTableParsingAdapter.java

@@ -8,6 +8,7 @@ import java.util.Map;
 
 import org.springframework.stereotype.Service;
 
+import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.PageInfo;
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableCells;
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableServiceResponse;
 
@@ -24,24 +25,26 @@ public class CvTableParsingAdapter {
         Map<Integer, List<TableCells>> tableCells = new HashMap<>();
         tableServiceResponse.getData()
                 .forEach(tableData -> tableCells.computeIfAbsent(tableData.getPageInfo().getNumber(), tableCell -> new ArrayList<>())
-                        .addAll(convertTableCells(tableData.getTableCells())));
+                        .addAll(convertTableCells(tableData.getTableCells(), tableData.getPageInfo())));
 
         return tableCells;
     }
 
 
-    private Collection<TableCells> convertTableCells(List<TableCells> tableCells) {
+    private Collection<TableCells> convertTableCells(List<TableCells> tableCells, PageInfo pageInfo) {
 
         List<TableCells> cvParsedTableCells = new ArrayList<>();
 
-        tableCells.forEach(t -> cvParsedTableCells.add(TableCells.builder()
-                .y0(t.getY0())
-                .x1(t.getX1())
-                .y1(t.getY1())
-                .x0(t.getX0())
-                .width(t.getWidth())
-                .height(t.getHeight())
-                .build()));
+        tableCells.stream()
+                .filter(cell -> cell.getWidth() < pageInfo.getWidth() * 0.98 && cell.getHeight() < pageInfo.getHeight() * 0.98)
+                .forEach(t -> cvParsedTableCells.add(TableCells.builder()
+                        .y0(t.getY0())
+                        .x1(t.getX1())
+                        .y1(t.getY1())
+                        .x0(t.getX0())
+                        .width(t.getWidth())
+                        .height(t.getHeight())
+                        .build()));
 
         return cvParsedTableCells;
     }

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/parsing/PDFLinesTextStripper.java

@@ -329,11 +329,6 @@ public class PDFLinesTextStripper extends PDFTextStripper {
                 .getXDirAdj() - (previous.getXDirAdj() + previous.getWidthDirAdj()) < maximumGapSize;
     }
 
-    // !(sublist.isEmpty() || sublist.size() == 1 && (sublist.get(0).getUnicode().equals(" ") || sublist.get(0).getUnicode().equals("\u00A0") || sublist.get(0)
-    //                        .getUnicode()
-    //                        .equals("\t")))
-
-
     @Override
     public String getText(PDDocument doc) throws IOException {
 

layoutparser-service/layoutparser-service-server/build.gradle.kts

@@ -3,7 +3,7 @@ import org.springframework.boot.gradle.tasks.bundling.BootBuildImage
 plugins {
     application
     id("com.knecon.fforesight.java-conventions")
-    id("org.springframework.boot") version "3.1.3"
+    id("org.springframework.boot") version "3.1.5"
     id("io.spring.dependency-management") version "1.1.3"
     id("org.sonarqube") version "4.3.0.3225"
     id("io.freefair.lombok") version "8.2.2"
@@ -12,31 +12,43 @@ plugins {
 
 description = "layoutparser-service-server"
 
+configurations {
+    all {
+        exclude(group = "commons-logging", module = "commons-logging")
+        exclude(group = "org.springframework.boot", module = "spring-boot-starter-log4j2")
+        exclude(group = "com.iqser.red.commons", module = "logging-commons")
+    }
+}
+
+
 val jacksonVersion = "2.15.2"
 val pdfBoxVersion = "3.0.0"
+val springBootStarterVersion = "3.1.5"
 
 dependencies {
     implementation(project(":layoutparser-service-processor"))
     implementation(project(":layoutparser-service-internal-api"))
 
     implementation("com.iqser.red.commons:storage-commons:2.45.0")
-    implementation("com.knecon.fforesight:tenant-commons:0.14.0")
-    implementation("org.springframework.boot:spring-boot-starter-actuator:3.1.3")
+    implementation("com.knecon.fforesight:tenant-commons:0.19.0")
+    implementation("com.knecon.fforesight:tracing-commons:0.2.0")
+    implementation("org.springframework.boot:spring-boot-starter-actuator:${springBootStarterVersion}")
+    implementation("org.springframework.boot:spring-boot-starter-amqp:${springBootStarterVersion}")
     implementation("com.amazonaws:aws-java-sdk-s3:1.12.536")
-    implementation("org.springframework.boot:spring-boot-starter-amqp:3.1.3")
     implementation("org.springframework.cloud:spring-cloud-starter-openfeign:4.0.4")
+    implementation("net.logstash.logback:logstash-logback-encoder:7.4")
 
     // for integration testing only
     testImplementation(project(":layoutparser-service-internal-api"))
 
-    testImplementation("org.springframework.boot:spring-boot-starter-amqp:3.1.3")
+    testImplementation("org.springframework.boot:spring-boot-starter-amqp:${springBootStarterVersion}")
+    testImplementation("org.springframework.boot:spring-boot-starter-test:${springBootStarterVersion}")
     testImplementation("com.iqser.red.service:persistence-service-shared-api-v1:2.144.0")
     testImplementation("com.iqser.red.commons:jackson-commons:1.3.0")
     testImplementation("com.fasterxml.jackson.module:jackson-module-afterburner:${jacksonVersion}")
     testImplementation("com.fasterxml.jackson.datatype:jackson-datatype-jsr310:${jacksonVersion}")
     testImplementation("org.apache.pdfbox:pdfbox:${pdfBoxVersion}")
     testImplementation("org.apache.pdfbox:pdfbox-tools:${pdfBoxVersion}")
-    testImplementation("org.springframework.boot:spring-boot-starter-test:3.1.3")
 }
 
 /*

layoutparser-service/layoutparser-service-server/src/main/java/com/knecon/fforesight/service/layoutparser/server/queue/MessagingConfiguration.java

@@ -24,8 +24,7 @@ public class MessagingConfiguration {
     @Bean
     public Queue layoutparsingResponseQueue() {
 
-        return QueueBuilder.durable(LayoutParsingQueueNames.LAYOUT_PARSING_FINISHED_EVENT_QUEUE)//
-                .withArgument("x-dead-letter-exchange", "").withArgument("x-dead-letter-routing-key", LayoutParsingQueueNames.LAYOUT_PARSING_DLQ).build();
+        return QueueBuilder.durable(LayoutParsingQueueNames.LAYOUT_PARSING_FINISHED_EVENT_QUEUE).build();
     }
 
 

layoutparser-service/layoutparser-service-server/src/main/resources/application.yml

@@ -5,10 +5,19 @@ tenant-user-management-service.url: "http://tenant-user-management-service:8080/
 
 fforesight.tenants.remote: true
 
+logging.pattern.level: "%5p [${spring.application.name},%X{traceId:-},%X{spanId:-}]"
+
+logging.type: ${LOGGING_TYPE:CONSOLE}
+kubernetes.namespace: ${NAMESPACE:default}
+project.version: 1.0-SNAPSHOT
+
+
 server:
   port: 8080
 
 spring:
+  application:
+    name: layoutparser-service
   main:
     allow-circular-references: true # FIXME
   rabbitmq:
@@ -32,6 +41,13 @@ management:
     prometheus.enabled: ${monitoring.enabled:false}
     health.enabled: true
   endpoints.web.exposure.include: prometheus, health
+  tracing:
+    enabled: ${TRACING_ENABLED:false}
+    sampling:
+      probability: ${TRACING_PROBABILITY:1.0}
+  otlp:
+    tracing:
+      endpoint: ${OTLP_ENDPOINT:http://otel-collector-opentelemetry-collector.otel-collector:4318/v1/traces}
 
 
 storage:

layoutparser-service/layoutparser-service-server/src/main/resources/logback-spring.xml

@@ -0,0 +1,17 @@
+<configuration>
+
+    <springProperty scope="configuration" name="logType" source="logging.type"/>
+    <springProperty scope="context" name="application.name" source="spring.application.name"/>
+    <springProperty scope="context" name="version" source="project.version"/>
+    <include resource="org/springframework/boot/logging/logback/defaults.xml"/>
+    <include resource="org/springframework/boot/logging/logback/console-appender.xml"/>
+
+    <appender name="JSON" class="ch.qos.logback.core.ConsoleAppender">
+        <encoder class="net.logstash.logback.encoder.LogstashEncoder"/>
+    </appender>
+
+    <root level="INFO">
+        <appender-ref ref="${logType}"/>
+    </root>
+
+</configuration>

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/segmentation/PdfSegmentationServiceTest.java

@@ -61,12 +61,13 @@ public class PdfSegmentationServiceTest extends AbstractTest {
     private SectionsBuilderService sectionsBuilderService;
 
 
-    public ClassificationDocument buildClassificationDocument(PDDocument originDocument) {
+    @SneakyThrows
+    public ClassificationDocument buildClassificationDocument(PDDocument originDocument, TableServiceResponse tableServiceResponse) {
 
         ClassificationDocument classificationDocument = layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
                 originDocument,
                 new ImageServiceResponse(),
-                new TableServiceResponse());
+                tableServiceResponse);
 
         redactManagerClassificationService.classifyDocument(classificationDocument);
 
@@ -76,6 +77,13 @@ public class PdfSegmentationServiceTest extends AbstractTest {
     }
 
 
+    @SneakyThrows
+    public ClassificationDocument buildClassificationDocument(PDDocument originDocument) {
+
+        return buildClassificationDocument(originDocument, new TableServiceResponse());
+    }
+
+
     @Test
     public void tablesToHtmlDebugger() throws IOException {
 
@@ -88,6 +96,39 @@ public class PdfSegmentationServiceTest extends AbstractTest {
     }
 
 
+    @Test
+    public void tablesToHtmlDebuggerWithCVResponse() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/cv_tables/ScanRotationBorder.pdf");
+        ClassPathResource cvTablesResource = new ClassPathResource("files/cv_tables/ScanRotationBorder.TABLES.json");
+        var tableServiceResponse = objectMapper.readValue(cvTablesResource.getInputStream(), TableServiceResponse.class);
+
+        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()), tableServiceResponse);
+
+        toHtml(document, "/tmp/ScanRotationBorder.html");
+    }
+
+
+    @Test
+    public void testScanRotationBorderIsIgnored() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/cv_tables/ScanRotationBorder.pdf");
+        ClassPathResource cvTablesResource = new ClassPathResource("files/cv_tables/ScanRotationBorder.TABLES.json");
+        var tableServiceResponse = objectMapper.readValue(cvTablesResource.getInputStream(), TableServiceResponse.class);
+
+        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()), tableServiceResponse);
+        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
+        var tables = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList();
+
+        // Quality of the table parsing is not good, because the file is rotated at scanning.
+        // We only asset that the table border is not the page border.
+        tables.forEach(table -> {
+            assertThat(table.getMinX()).isGreaterThan(75);
+            assertThat(table.getMaxX()).isLessThan(512);
+        });
+    }
+
+
     @Test
     @SneakyThrows
     public void testMapping() {