GDCC/8449 use original files in archival bag

src/main/java/edu/harvard/iq/dataverse/util/bagit/OREMap.java

@@ -166,7 +166,32 @@ public JsonObjectBuilder getOREMapBuilder(boolean aggregationOnly) throws Except
                 } else {
                     addIfNotNull(aggRes, JsonLDTerm.schemaOrg("description"), df.getDescription());
                 }
-                addIfNotNull(aggRes, JsonLDTerm.schemaOrg("name"), fmd.getLabel()); // "label" is the filename
+                String fileName = fmd.getLabel();// "label" is the filename
+                long fileSize = df.getFilesize();
+                String mimeType = df.getContentType();
+                String currentIngestedName = null;
+                boolean ingested=df.getOriginalFileName()!= null || df.getOriginalFileSize()!=null || df.getOriginalFileFormat()!=null;
+                if(ingested) {
+                    if(df.getOriginalFileName()!=null) {
+                        currentIngestedName= fileName;
+                        fileName = df.getOriginalFileName();
+                    } else {
+                        logger.warning("Missing Original file name for id: " + df.getId());
+                    }
+                    if(df.getOriginalFileSize()!=null) {
+                        fileSize = df.getOriginalFileSize();
+                    } else {
+                        logger.warning("Missing Original file size for id: " + df.getId());
+                    }
+                    if(df.getOriginalFileFormat()!=null) {
+                        mimeType = df.getOriginalFileFormat();
+                    } else {
+                        logger.warning("Missing Original file format for id: " + df.getId());
+                    }
+
+
+                }
+                addIfNotNull(aggRes, JsonLDTerm.schemaOrg("name"), fileName); 
                 addIfNotNull(aggRes, JsonLDTerm.restricted, fmd.isRestricted());
                 addIfNotNull(aggRes, JsonLDTerm.directoryLabel, fmd.getDirectoryLabel());
                 addIfNotNull(aggRes, JsonLDTerm.schemaOrg("version"), fmd.getVersion());
@@ -189,21 +214,20 @@ public JsonObjectBuilder getOREMapBuilder(boolean aggregationOnly) throws Except
                 if (df.getGlobalId().asString().length() != 0) {
                     fileId = df.getGlobalId().asString();
                     fileSameAs = SystemConfig.getDataverseSiteUrlStatic()
-                            + "/api/access/datafile/:persistentId?persistentId=" + fileId;
+                            + "/api/access/datafile/:persistentId?persistentId=" + fileId + (ingested ? "&format=original":"");
                 } else {
                     fileId = SystemConfig.getDataverseSiteUrlStatic() + "/file.xhtml?fileId=" + df.getId();
-                    fileSameAs = SystemConfig.getDataverseSiteUrlStatic() + "/api/access/datafile/" + df.getId();
+                    fileSameAs = SystemConfig.getDataverseSiteUrlStatic() + "/api/access/datafile/" + df.getId() + (ingested ? "?format=original":"");
                 }
                 aggRes.add("@id", fileId);
                 aggRes.add(JsonLDTerm.schemaOrg("sameAs").getLabel(), fileSameAs);
                 fileArray.add(fileId);
 
                 aggRes.add("@type", JsonLDTerm.ore("AggregatedResource").getLabel());
-                addIfNotNull(aggRes, JsonLDTerm.schemaOrg("fileFormat"), df.getContentType());
-                addIfNotNull(aggRes, JsonLDTerm.filesize, df.getFilesize());
+                addIfNotNull(aggRes, JsonLDTerm.schemaOrg("fileFormat"), mimeType);
+                addIfNotNull(aggRes, JsonLDTerm.filesize, fileSize);
                 addIfNotNull(aggRes, JsonLDTerm.storageIdentifier, df.getStorageIdentifier());
-                addIfNotNull(aggRes, JsonLDTerm.originalFileFormat, df.getOriginalFileFormat());
-                addIfNotNull(aggRes, JsonLDTerm.originalFormatLabel, df.getOriginalFormatLabel());
+                addIfNotNull(aggRes, JsonLDTerm.currentIngestedName, currentIngestedName);
                 addIfNotNull(aggRes, JsonLDTerm.UNF, df.getUnf());
                 addIfNotNull(aggRes, JsonLDTerm.rootDataFileId, df.getRootDataFileId());
                 addIfNotNull(aggRes, JsonLDTerm.previousDataFileId, df.getPreviousDataFileId());

src/main/java/edu/harvard/iq/dataverse/util/json/JsonLDTerm.java

@@ -34,8 +34,11 @@ public class JsonLDTerm {
     public static JsonLDTerm categories = JsonLDTerm.DVCore("categories");
     public static JsonLDTerm filesize = JsonLDTerm.DVCore("filesize");
     public static JsonLDTerm storageIdentifier = JsonLDTerm.DVCore("storageIdentifier");
+    @Deprecated
     public static JsonLDTerm originalFileFormat = JsonLDTerm.DVCore("originalFileFormat");
+    @Deprecated
     public static JsonLDTerm originalFormatLabel = JsonLDTerm.DVCore("originalFormatLabel");
+    public static JsonLDTerm currentIngestedName= JsonLDTerm.DVCore("currentIngestedName");
     public static JsonLDTerm UNF = JsonLDTerm.DVCore("UNF");
     public static JsonLDTerm rootDataFileId = JsonLDTerm.DVCore("rootDataFileId");
     public static JsonLDTerm previousDataFileId = JsonLDTerm.DVCore("previousDataFileId");

src/test/java/edu/harvard/iq/dataverse/api/BagIT.java

@@ -0,0 +1,78 @@
+package edu.harvard.iq.dataverse.api;
+
+import com.jayway.restassured.RestAssured;
+import com.jayway.restassured.response.Response;
+import edu.harvard.iq.dataverse.engine.command.impl.LocalSubmitToArchiveCommand;
+import edu.harvard.iq.dataverse.settings.SettingsServiceBean;
+import static javax.ws.rs.core.Response.Status.CREATED;
+import static javax.ws.rs.core.Response.Status.OK;
+import org.junit.AfterClass;
+import org.junit.BeforeClass;
+import org.junit.Test;
+
+public class BagIT {
+
+    @BeforeClass
+    public static void setUpClass() {
+
+        RestAssured.baseURI = UtilIT.getRestAssuredBaseUri();
+
+        Response setArchiverClassName = UtilIT.setSetting(SettingsServiceBean.Key.ArchiverClassName, LocalSubmitToArchiveCommand.class.getCanonicalName());
+        setArchiverClassName.then().assertThat()
+                .statusCode(OK.getStatusCode());
+
+        Response setArchiverSettings = UtilIT.setSetting(SettingsServiceBean.Key.ArchiverSettings, ":BagItLocalPath, :BagGeneratorThreads");
+        setArchiverSettings.then().assertThat()
+                .statusCode(OK.getStatusCode());
+
+        Response setBagItLocalPath = UtilIT.setSetting(":BagItLocalPath", "/tmp");
+        setBagItLocalPath.then().assertThat()
+                .statusCode(OK.getStatusCode());
+
+    }
+
+    @Test
+    public void testBagItExport() {
+
+        Response createUser = UtilIT.createRandomUser();
+        createUser.then().assertThat().statusCode(OK.getStatusCode());
+        String username = UtilIT.getUsernameFromResponse(createUser);
+        String apiToken = UtilIT.getApiTokenFromResponse(createUser);
+        Response toggleSuperuser = UtilIT.makeSuperUser(username);
+        toggleSuperuser.then().assertThat()
+                .statusCode(OK.getStatusCode());
+
+        Response createDataverse = UtilIT.createRandomDataverse(apiToken);
+        createDataverse.then().assertThat().statusCode(CREATED.getStatusCode());
+        String dataverseAlias = UtilIT.getAliasFromResponse(createDataverse);
+        Integer dataverseId = UtilIT.getDataverseIdFromResponse(createDataverse);
+
+        Response createDataset = UtilIT.createRandomDatasetViaNativeApi(dataverseAlias, apiToken);
+        createDataset.prettyPrint();
+        createDataset.then().assertThat()
+                .statusCode(CREATED.getStatusCode());
+
+        String datasetPid = UtilIT.getDatasetPersistentIdFromResponse(createDataset);
+
+        Response publishDataverse = UtilIT.publishDataverseViaNativeApi(dataverseAlias, apiToken);
+        publishDataverse.then().assertThat().statusCode(OK.getStatusCode());
+        Response publishDataset = UtilIT.publishDatasetViaNativeApi(datasetPid, "major", apiToken);
+        publishDataset.then().assertThat().statusCode(OK.getStatusCode());
+
+        Response archiveDataset = UtilIT.archiveDataset(datasetPid, "1.0", apiToken);
+        archiveDataset.prettyPrint();
+        archiveDataset.then().assertThat().statusCode(OK.getStatusCode());
+
+    }
+
+    @AfterClass
+    public static void tearDownClass() {
+
+        // Not checking if delete happened. Hopefully, it did.
+        UtilIT.deleteSetting(SettingsServiceBean.Key.ArchiverClassName);
+        UtilIT.deleteSetting(SettingsServiceBean.Key.ArchiverSettings);
+        UtilIT.deleteSetting(":BagItLocalPath");
+
+    }
+
+}

src/test/java/edu/harvard/iq/dataverse/api/UtilIT.java

@@ -1879,6 +1879,14 @@ static Response deleteSetting(SettingsServiceBean.Key settingKey) {
         return response;
     }
 
+    /**
+     * @param settingKey Include the colon like :BagItLocalPath
+     */
+    static Response deleteSetting(String settingKey) {
+        Response response = given().when().delete("/api/admin/settings/" + settingKey);
+        return response;
+    }
+
     static Response getSetting(SettingsServiceBean.Key settingKey) {
         Response response = given().when().get("/api/admin/settings/" + settingKey);
         return response;
@@ -1889,6 +1897,14 @@ static Response setSetting(SettingsServiceBean.Key settingKey, String value) {
         return response;
     }
 
+    /**
+     * @param settingKey Include the colon like :BagItLocalPath
+     */
+    public static Response setSetting(String settingKey, String value) {
+        Response response = given().body(value).when().put("/api/admin/settings/" + settingKey);
+        return response;
+    }
+
     static Response getRoleAssignmentsOnDataverse(String dataverseAliasOrId, String apiToken) {
         String url = "/api/dataverses/" + dataverseAliasOrId + "/assignments";
         return given()
@@ -2927,6 +2943,20 @@ static Response getDatasetVersionArchivalStatus(Integer datasetId, String versio
                 .get("/api/datasets/" + datasetId + "/" + version + "/archivalStatus");
         return response;
     }
+
+    static Response archiveDataset(String idOrPersistentIdOfDataset, String version, String apiToken) {
+        String idInPath = idOrPersistentIdOfDataset;
+        String optionalQueryParam = "";
+        if (!NumberUtils.isNumber(idOrPersistentIdOfDataset)) {
+            idInPath = ":persistentId";
+            optionalQueryParam = "?persistentId=" + idOrPersistentIdOfDataset;
+        }
+        Response response = given()
+                .header(API_TOKEN_HTTP_HEADER, apiToken)
+                .post("/api/admin/submitDatasetVersionToArchive/" + idInPath + "/" + version + optionalQueryParam);
+        return response;
+    }
+
     static Response setDatasetVersionArchivalStatus(Integer datasetId, String version, String apiToken, String status, String message) {
         Response response = given()
                 .header(API_TOKEN_HTTP_HEADER, apiToken).contentType("application/json; charset=utf-8").body("{\"status\":\"" + status + "\", \"message\":\"" + message + "\"}")

tests/integration-tests.txt

@@ -1 +1 @@
-DataversesIT,DatasetsIT,SwordIT,AdminIT,BuiltinUsersIT,UsersIT,UtilIT,ConfirmEmailIT,FileMetadataIT,FilesIT,SearchIT,InReviewWorkflowIT,HarvestingServerIT,MoveIT,MakeDataCountApiIT,FileTypeDetectionIT,EditDDIIT,ExternalToolsIT,AccessIT,DuplicateFilesIT,DownloadFilesIT,LinkIT,DeleteUsersIT,DeactivateUsersIT,AuxiliaryFilesIT,InvalidCharactersIT,LicensesIT,NotificationsIT
+DataversesIT,DatasetsIT,SwordIT,AdminIT,BuiltinUsersIT,UsersIT,UtilIT,ConfirmEmailIT,FileMetadataIT,FilesIT,SearchIT,InReviewWorkflowIT,HarvestingServerIT,MoveIT,MakeDataCountApiIT,FileTypeDetectionIT,EditDDIIT,ExternalToolsIT,AccessIT,DuplicateFilesIT,DownloadFilesIT,LinkIT,DeleteUsersIT,DeactivateUsersIT,AuxiliaryFilesIT,InvalidCharactersIT,LicensesIT,NotificationsIT,BagIT