RED-4510: Generated Redaction Log and Generating will not run with profile

redaction-service-v1/redaction-service-server-v1/src/test/java/com/iqser/red/service/redaction/v1/server/RulesTest.java

@@ -28,6 +28,7 @@ import java.util.Set;
 import java.util.stream.Collectors;
 
 import org.apache.commons.lang3.StringUtils;
+import org.assertj.core.data.Offset;
 import org.junit.After;
 import org.junit.Before;
 import org.junit.Ignore;
@@ -49,6 +50,7 @@ import org.springframework.context.annotation.Configuration;
 import org.springframework.context.annotation.Import;
 import org.springframework.context.annotation.Primary;
 import org.springframework.core.io.ClassPathResource;
+import org.springframework.test.annotation.IfProfileValue;
 import org.springframework.test.context.junit4.SpringRunner;
 
 import com.amazonaws.services.s3.AmazonS3;
@@ -118,6 +120,10 @@ public class RulesTest {
     private final static String TEST_DOSSIER_TEMPLATE_ID = "123";
     private final static String TEST_DOSSIER_ID = "123";
     private String TEST_FILE_ID = "123";
+
+    private int counter = 0;
+
+    private int fileSize = 0;
     private final Map<String, List<String>> dictionary = new HashMap<>();
     private final Map<String, List<String>> dossierDictionary = new HashMap<>();
     private final Map<String, List<String>> falsePositive = new HashMap<>();
@@ -206,6 +212,8 @@ public class RulesTest {
     public void generateRedactionLogForOneFile() {
 
         String fileName = "files/Compounds/31 A14111B - EU AIR3 - MCP Section 1 - Identity of the plant protection product.pdf";
+        fileName = "scanned/VV-377031.pdf";//TODO
+        fileSize = 1;
         generateAndSaveRedactionLog(fileName, true);
     }
 
@@ -216,10 +224,12 @@ public class RulesTest {
      * Please commit generated files, if not the following tests will fail
      */
     @Test
+    @Ignore//TODO
     public void generateRedactionLogForAllFiles() {
 
         Set<String> files = getFileNames(new HashSet<>(), FileSystems.getDefault().getPath(RESOURCES_PATH));
         System.out.println("Will generate RedactionLog for " + files.size() + " files.");
+        fileSize = files.size();
         files.forEach(this::generateAndSaveRedactionLog);
     }
 
@@ -229,11 +239,12 @@ public class RulesTest {
      * If RedactionLog Json does not exist, test will fail.
      */
     @Test
+//    @Ignore//TODO
     public void analyseFileAndCompareRedactionLog() {
 
-        String fileName = "files/RulesTest/SYNGENTA_EFSA_sanitisation_GFL_v1_withHighlights.pdf"; //TODO PSC
-//        String fileName = "scanned/VV-377031.pdf";
-        fileName = "files/Cyprodinil/49 Cyprodinil - EU AIR3 - MCA Section 8 Supplement - Ecotoxicological studies on the active substance.pdf";
+//        String fileName = "files/RulesTest/SYNGENTA_EFSA_sanitisation_GFL_v1_withHighlights.pdf"; //TODO PSC
+        String fileName = "scanned/VV-377031.pdf";
+        fileSize = 1;
         analyseFileAndCompareRedactionLog(fileName);
     }
 
@@ -242,12 +253,14 @@ public class RulesTest {
      * Analyses all files and compares its RedactionLog with saved one from here: REDACTION_LOG_PATH.
      * If RedactionLog Json for one file does not exist, whole test will fail.
      */
-//    @IfProfileValue(name = "test-groups", value = "rules-test") //TODO PSC
+    @IfProfileValue(name = "test-groups", value = "rules-test")
     @Test
+    @Ignore//TODO PSC
     public void analyseAllFilesAndCompareRedactionLogs() {
 
         Set<String> files = getFileNames(new HashSet<>(), FileSystems.getDefault().getPath(RESOURCES_PATH));
         System.out.println("Will analyse " + files.size() + " files and compare its RedactionLogs.");
+        fileSize = files.size();
         files.forEach(this::analyseFileAndCompareRedactionLog);
     }
 
@@ -261,13 +274,14 @@ public class RulesTest {
     @SneakyThrows
     private void generateAndSaveRedactionLog(String fileName, boolean force) {
 
+        counter++;
         if (!force && doesRedactionLogExist(fileName)) {
-            System.out.println("Skip generating RedactionLog as Json for " + fileName + " because it already exists.");
+            System.out.println("(" + counter + "/" + fileSize + ") Skip generating RedactionLog as Json for " + fileName + " because it already exists.");
 
         } else {
             generateTestFileId(fileName);
 
-            System.out.println("Generate RedactionLog as Json for " + fileName + " with fileId " + TEST_FILE_ID);
+            System.out.println("(" + counter + "/" + fileSize + ") Generate RedactionLog as Json for " + fileName + " with fileId " + TEST_FILE_ID);
 
             loadNerForTest();
 
@@ -286,8 +300,9 @@ public class RulesTest {
     @SneakyThrows
     public void analyseFileAndCompareRedactionLog(String fileName) {
 
+        counter++;
         generateTestFileId(fileName);
-        System.out.println("Analyse " + fileName + " with fileId " + TEST_FILE_ID + " and compare it with its saved RedactionLog");
+        System.out.println("(" + counter + "/" + fileSize + ") Analyse " + fileName + " with fileId " + TEST_FILE_ID + " and compare it with its saved RedactionLog");
 
         RedactionLog savedRedactionLog = loadSavedRedactionLog(fileName);
 
@@ -340,8 +355,17 @@ public class RulesTest {
             assertThat(savedRedactionLogEntry.get().getSectionNumber()).isEqualTo(redactionLogEntry.getSectionNumber());
             assertThat(savedRedactionLogEntry.get().getTextBefore()).isEqualTo(redactionLogEntry.getTextBefore());
             assertThat(savedRedactionLogEntry.get().getTextAfter()).isEqualTo(redactionLogEntry.getTextAfter());
+
+            System.out.println("OFF " + savedRedactionLogEntry.get().getId());
+            System.out.println("OFF " + savedRedactionLogEntry.get().getStartOffset() + " " + redactionLogEntry.getStartOffset());
+            System.out.println("OFF " + savedRedactionLogEntry.get().getEndOffset() + " " + redactionLogEntry.getEndOffset());
+
+            assertThat(savedRedactionLogEntry.get().getStartOffset()).isCloseTo(redactionLogEntry.getStartOffset(), Offset.offset(4));
+            assertThat(savedRedactionLogEntry.get().getEndOffset()).isCloseTo(redactionLogEntry.getEndOffset(), Offset.offset(4));
+
             assertThat(savedRedactionLogEntry.get().getStartOffset()).isEqualTo(redactionLogEntry.getStartOffset());
             assertThat(savedRedactionLogEntry.get().getEndOffset()).isEqualTo(redactionLogEntry.getEndOffset());
+
             assertThat(savedRedactionLogEntry.get().isImage()).isEqualTo(redactionLogEntry.isImage());
             assertThat(savedRedactionLogEntry.get().isImageHasTransparency()).isEqualTo(redactionLogEntry.isImageHasTransparency());
             assertThat(savedRedactionLogEntry.get().isDictionaryEntry()).isEqualTo(redactionLogEntry.isDictionaryEntry());
@@ -470,6 +494,25 @@ public class RulesTest {
     }
 
 
+    @SneakyThrows
+    private Set<String> getFileNames(Set<String> fileNames, Path dir) {
+
+        try (DirectoryStream<Path> stream = Files.newDirectoryStream(dir)) {
+            for (Path path : stream) {
+                if (path.toFile().isDirectory()) {
+                    getFileNames(fileNames, path);
+                } else if (StringUtils.endsWith(path.toAbsolutePath().toString(), ".pdf")) {
+                    String absolutePath = cleanPath(path.toAbsolutePath().toString());
+                    int pos = StringUtils.indexOf(absolutePath, RESOURCES_PATH) + RESOURCES_PATH.length();
+
+                    fileNames.add(StringUtils.substring(absolutePath, pos));
+                }
+            }
+        }
+        return fileNames;
+    }
+
+
     private String cleanPath(String path) {
 
         return StringUtils.replace(path, "\\", "/");
@@ -614,30 +657,6 @@ public class RulesTest {
     }
 
 
-    @SneakyThrows
-    private Set<String> getFileNames(Set<String> fileNames, Path dir) {
-
-        try (DirectoryStream<Path> stream = Files.newDirectoryStream(dir)) {
-            for (Path path : stream) {
-                if (path.toFile().isDirectory()) {
-                    getFileNames(fileNames, path);
-                } else if (StringUtils.endsWith(path.toAbsolutePath().toString(), ".pdf")) {
-                    String absolutePath = cleanPath(path.toAbsolutePath().toString());
-                    int pos = StringUtils.indexOf(absolutePath, RESOURCES_PATH) + RESOURCES_PATH.length();
-
-                    if (StringUtils.contains(absolutePath, "scanned/VV-377031.pdf")) {
-                        System.out.println("SKIPPED! " + absolutePath);// TODO PSC
-                    } else {
-                        fileNames.add(StringUtils.substring(absolutePath, pos));
-                    }
-
-                }
-            }
-        }
-        return fileNames;
-    }
-
-
     private Type getDictionaryResponse(String type, boolean isDossierDictionary) {
 
         return Type.builder()