Added DataFileCategoryService to manage data file categories configuration

doc/release-notes/8461-filecategories-config.md

@@ -0,0 +1,3 @@
+### The default file categories are now configurable 
+
+The default list of the pre-defined file categories - "Documentation", "Data" and "Code" - can now be redefined via a database setting `:FileCategories`. Consult the [Database Settings](https://guides.dataverse.org/en/latest/installation/config.html) section of the Guides for more information.

doc/sphinx-guides/source/installation/config.rst

@@ -2576,3 +2576,18 @@ For example:
 ++++++++++++++++++++++++++++++++
 
 When set to ``true``, this setting allows a superuser to publish and/or update Dataverse collections and datasets bypassing the external validation checks (specified by the settings above). In an event where an external script is reporting validation failures that appear to be in error, this option gives an admin with superuser privileges a quick way to publish the dataset or update a collection for the user. 
+
+:FileCategories
++++++++++++++++
+
+Overrides the default list of file categories that is used in the UI when adding tags to files. The default list is Documentation, Data, and Code.
+
+This setting is a comma-separated list of the new tags.
+
+To override the default list with Docs, Data, Code, and Workflow:
+
+``curl -X PUT -d 'Docs,Data,Code,Workflow' http://localhost:8080/api/admin/settings/:FileCategories``
+
+To remove the override and go back to the default list:
+
+``curl -X PUT -d '' http://localhost:8080/api/admin/settings/:FileCategories``

src/main/java/edu/harvard/iq/dataverse/DataFileCategoryServiceBean.java

@@ -0,0 +1,87 @@
+package edu.harvard.iq.dataverse;
+
+import edu.harvard.iq.dataverse.settings.SettingsServiceBean;
+import edu.harvard.iq.dataverse.util.BundleUtil;
+
+import javax.ejb.EJB;
+import javax.ejb.Stateless;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.List;
+import java.util.Optional;
+import java.util.stream.Collectors;
+
+/**
+ * Service to manage the default values for Dataset file categories and allow
+ * to be overridden with a :FileCategories Settings configuration.
+ *
+ * @author adaybujeda
+ */
+@Stateless
+public class DataFileCategoryServiceBean {
+
+    public static final String FILE_CATEGORIES_KEY = ":FileCategories";
+
+    @EJB
+    private SettingsServiceBean settingsService;
+
+    public List<String> mergeDatasetFileCategories(List<DataFileCategory> datasetFileCategories) {
+        List<DataFileCategory> fileCategories = Optional.ofNullable(datasetFileCategories).orElse(Collections.emptyList());
+        List<String> defaultFileCategories = getFileCategories();
+
+        //avoid resizing
+        List<String> mergedFileCategories = new ArrayList<>(defaultFileCategories.size() + fileCategories.size());
+
+        for(DataFileCategory category: fileCategories) {
+            mergedFileCategories.add(category.getName());
+        }
+
+        for(String defaultCategory: defaultFileCategories) {
+            if (!mergedFileCategories.contains(defaultCategory)) {
+                mergedFileCategories.add(defaultCategory);
+            }
+        }
+
+        return mergedFileCategories;
+    }
+
+    public List<String> getFileCategories() {
+        List<String> fileCategoriesOverride = getFileCategoriesOverride();
+        return fileCategoriesOverride.isEmpty() ? getFileCategoriesDefault() : fileCategoriesOverride;
+    }
+
+    private List<String> getFileCategoriesDefault() {
+        // "Documentation", "Data" and "Code" are the 3 default categories that we
+        // present by default
+        return Arrays.asList(
+                BundleUtil.getStringFromBundle("dataset.category.documentation"),
+                BundleUtil.getStringFromBundle("dataset.category.data"),
+                BundleUtil.getStringFromBundle("dataset.category.code")
+        );
+    }
+
+    private List<String> getFileCategoriesOverride() {
+        String applicationLanguage = BundleUtil.getCurrentLocale().getLanguage();
+        Optional<String> fileCategoriesOverride = Optional.ofNullable(settingsService.get(FILE_CATEGORIES_KEY));
+
+        if (fileCategoriesOverride.isPresent()) {
+            // There is an override, check if there is language specific value
+            String overrideValue = settingsService.get(FILE_CATEGORIES_KEY, applicationLanguage, fileCategoriesOverride.get());
+
+            return parseCategoriesString(overrideValue);
+        }
+
+        return Collections.emptyList();
+    }
+
+    private List<String> parseCategoriesString(String categoriesString) {
+        if (categoriesString == null) {
+            return Collections.emptyList();
+        }
+
+        String[] categories = categoriesString.split(",");
+        return Arrays.stream(categories).map(item -> item.trim()).filter(item -> !item.isBlank()).collect(Collectors.toUnmodifiableList());
+    }
+
+}

src/main/java/edu/harvard/iq/dataverse/Dataset.java

@@ -448,24 +448,6 @@ public void addFileCategory(DataFileCategory category) {
         dataFileCategories.add(category);
     }
 
-    public Collection<String> getCategoriesByName() {
-        Collection<String> ret = getCategoryNames();
-
-        // "Documentation", "Data" and "Code" are the 3 default categories that we 
-        // present by default:
-        if (!ret.contains(BundleUtil.getStringFromBundle("dataset.category.documentation"))) {
-            ret.add(BundleUtil.getStringFromBundle("dataset.category.documentation"));
-        }
-        if (!ret.contains(BundleUtil.getStringFromBundle("dataset.category.data"))) {
-            ret.add(BundleUtil.getStringFromBundle("dataset.category.data"));
-        }
-        if (!ret.contains(BundleUtil.getStringFromBundle("dataset.category.code"))) {
-            ret.add(BundleUtil.getStringFromBundle("dataset.category.code"));
-        }
-
-        return ret;
-    }
-
     public void setCategoriesByName(List<String> newCategoryNames) {
         if (newCategoryNames != null) {
             Collection<String> oldCategoryNames = getCategoryNames();

src/main/java/edu/harvard/iq/dataverse/DatasetPage.java

@@ -64,9 +64,6 @@
 import java.io.InputStream;
 import java.sql.Timestamp;
 import java.text.SimpleDateFormat;
-import java.time.Instant;
-import java.time.LocalDate;
-import java.time.format.DateTimeFormatter;
 import java.util.ArrayList;
 import java.util.Date;
 import java.util.HashMap;
@@ -78,7 +75,6 @@
 import java.util.Set;
 import java.util.Collection;
 import java.util.logging.Logger;
-import java.util.stream.Collectors;
 import javax.ejb.EJB;
 import javax.ejb.EJBException;
 import javax.faces.application.FacesMessage;
@@ -88,7 +84,6 @@
 import javax.faces.view.ViewScoped;
 import javax.inject.Inject;
 import javax.inject.Named;
-import javax.json.JsonObject;
 
 import org.apache.commons.lang3.StringUtils;
 import org.primefaces.event.FileUploadEvent;
@@ -128,7 +123,6 @@
 import javax.servlet.http.HttpServletResponse;
 
 import org.apache.commons.text.StringEscapeUtils;
-import org.apache.commons.validator.routines.EmailValidator;
 import org.apache.commons.lang3.mutable.MutableBoolean;
 import org.apache.commons.io.IOUtils;
 import org.primefaces.component.selectonemenu.SelectOneMenu;
@@ -251,6 +245,8 @@ public enum DisplayMode {
     EmbargoServiceBean embargoService;
     @Inject
     LicenseServiceBean licenseServiceBean;
+    @Inject
+    DataFileCategoryServiceBean dataFileCategoryService;
 
     private Dataset dataset = new Dataset();
 
@@ -4529,7 +4525,8 @@ public void setTabFileTagsByName(List<String> tabFileTagsByName) {
 
     private void refreshCategoriesByName(){
         categoriesByName= new ArrayList<>();
-        for (String category: dataset.getCategoriesByName() ){
+        List<String> datasetFileCategories = dataFileCategoryService.mergeDatasetFileCategories(dataset.getCategories());
+        for (String category: datasetFileCategories ){
             categoriesByName.add(category);
         }
         refreshSelectedTags();

src/main/java/edu/harvard/iq/dataverse/EditDatafilesPage.java

@@ -138,6 +138,8 @@ public enum Referrer {
     SettingsWrapper settingsWrapper;
     @Inject
     LicenseServiceBean licenseServiceBean;
+    @Inject
+    DataFileCategoryServiceBean dataFileCategoryService;
 
     private Dataset dataset = new Dataset();
 
@@ -2737,7 +2739,8 @@ private void refreshSelectedTabFileTags() {
 
     private void refreshCategoriesByName(){
         categoriesByName= new ArrayList<>();
-        for (String category: dataset.getCategoriesByName() ){
+        List<String> datasetFileCategories = dataFileCategoryService.mergeDatasetFileCategories(dataset.getCategories());
+        for (String category: datasetFileCategories ){
             categoriesByName.add(category);
         }
         refreshSelectedTags();

src/test/java/edu/harvard/iq/dataverse/DataFileCategoryServiceBeanTest.java

@@ -0,0 +1,155 @@
+package edu.harvard.iq.dataverse;
+
+import edu.harvard.iq.dataverse.settings.SettingsServiceBean;
+import edu.harvard.iq.dataverse.util.BundleUtil;
+import org.hamcrest.MatcherAssert;
+import org.hamcrest.Matchers;
+import org.junit.Test;
+import org.junit.runner.RunWith;
+import org.mockito.InjectMocks;
+import org.mockito.Mock;
+import org.mockito.Mockito;
+import org.mockito.junit.MockitoJUnitRunner;
+
+import java.util.Arrays;
+import java.util.List;
+import java.util.stream.Collectors;
+
+/**
+ * 
+ * @author adaybujeda
+ */
+@RunWith(MockitoJUnitRunner.class)
+public class DataFileCategoryServiceBeanTest {
+
+    @Mock
+    private SettingsServiceBean settingsServiceBean;
+    @InjectMocks
+    private DataFileCategoryServiceBean target;
+
+    @Test
+    public void getFileCategories_should_return_default_file_categories_in_expected_order_when_no_override_configured() {
+        Mockito.when(settingsServiceBean.get(DataFileCategoryServiceBean.FILE_CATEGORIES_KEY)).thenReturn(null);
+
+        List<String> result = target.getFileCategories();
+
+        MatcherAssert.assertThat(result.size(), Matchers.is(3));
+        MatcherAssert.assertThat(result.get(0), Matchers.is("Documentation"));
+        MatcherAssert.assertThat(result.get(1), Matchers.is("Data"));
+        MatcherAssert.assertThat(result.get(2), Matchers.is("Code"));
+    }
+
+    @Test
+    public void getFileCategories_should_return_default_file_categories_in_expected_order_when_empty_override_is_configured() {
+        Mockito.when(settingsServiceBean.get(DataFileCategoryServiceBean.FILE_CATEGORIES_KEY)).thenReturn("  ");
+
+        List<String> result = target.getFileCategories();
+
+        MatcherAssert.assertThat(result.size(), Matchers.is(3));
+        MatcherAssert.assertThat(result.get(0), Matchers.is("Documentation"));
+        MatcherAssert.assertThat(result.get(1), Matchers.is("Data"));
+        MatcherAssert.assertThat(result.get(2), Matchers.is("Code"));
+    }
+
+    @Test
+    public void getFileCategories_should_return_override_file_categories_from_settings_service() {
+        setup_override("Override01, Override02");
+
+        List<String> result = target.getFileCategories();
+
+        MatcherAssert.assertThat(result.size(), Matchers.is(2));
+        MatcherAssert.assertThat(result.get(0), Matchers.is("Override01"));
+        MatcherAssert.assertThat(result.get(1), Matchers.is("Override02"));
+    }
+
+    @Test
+    public void getFileCategories_should_trim_override_values() {
+        setup_override(" Test Value 01  , Test Value 02  ");
+
+        List<String> result = target.getFileCategories();
+
+        MatcherAssert.assertThat(result.size(), Matchers.is(2));
+        MatcherAssert.assertThat(result.get(0), Matchers.is("Test Value 01"));
+        MatcherAssert.assertThat(result.get(1), Matchers.is("Test Value 02"));
+    }
+
+    @Test
+    public void getFileCategories_should_ignore_empty_override_values() {
+        setup_override(",value01,,value02,,value03,,");
+
+        List<String> result = target.getFileCategories();
+
+        MatcherAssert.assertThat(result.size(), Matchers.is(3));
+        MatcherAssert.assertThat(result.get(0), Matchers.is("value01"));
+        MatcherAssert.assertThat(result.get(1), Matchers.is("value02"));
+        MatcherAssert.assertThat(result.get(2), Matchers.is("value03"));
+    }
+
+    @Test
+    public void mergeDatasetFileCategories_should_handle_null_datafile_categories() {
+        Mockito.when(settingsServiceBean.get(DataFileCategoryServiceBean.FILE_CATEGORIES_KEY)).thenReturn(null);
+
+        List<String> result = target.mergeDatasetFileCategories(null);
+
+        MatcherAssert.assertThat(result.size(), Matchers.is(3));
+        MatcherAssert.assertThat(result.get(0), Matchers.is("Documentation"));
+        MatcherAssert.assertThat(result.get(1), Matchers.is("Data"));
+        MatcherAssert.assertThat(result.get(2), Matchers.is("Code"));
+    }
+
+    @Test
+    public void mergeDatasetFileCategories_should_add_dataset_values_first_then_default_categories() {
+        Mockito.when(settingsServiceBean.get(DataFileCategoryServiceBean.FILE_CATEGORIES_KEY)).thenReturn(null);
+
+        List<String> result = target.mergeDatasetFileCategories(setup_data_file_categories("dataset01", "dataset02"));
+
+        MatcherAssert.assertThat(result.size(), Matchers.is(5));
+        MatcherAssert.assertThat(result.get(0), Matchers.is("dataset01"));
+        MatcherAssert.assertThat(result.get(1), Matchers.is("dataset02"));
+        MatcherAssert.assertThat(result.get(2), Matchers.is("Documentation"));
+        MatcherAssert.assertThat(result.get(3), Matchers.is("Data"));
+        MatcherAssert.assertThat(result.get(4), Matchers.is("Code"));
+    }
+
+    @Test
+    public void mergeDatasetFileCategories_should_add_dataset_values_first_then_override_categories() {
+        setup_override("override01, override02");
+
+        List<String> result = target.mergeDatasetFileCategories(setup_data_file_categories("dataset01", "dataset02"));
+
+        MatcherAssert.assertThat(result.size(), Matchers.is(4));
+        MatcherAssert.assertThat(result.get(0), Matchers.is("dataset01"));
+        MatcherAssert.assertThat(result.get(1), Matchers.is("dataset02"));
+        MatcherAssert.assertThat(result.get(2), Matchers.is("override01"));
+        MatcherAssert.assertThat(result.get(3), Matchers.is("override02"));
+    }
+
+    @Test
+    public void mergeDatasetFileCategories_should_ignore_duplicates() {
+        Mockito.when(settingsServiceBean.get(DataFileCategoryServiceBean.FILE_CATEGORIES_KEY)).thenReturn(null);
+
+        List<String> result = target.mergeDatasetFileCategories(setup_data_file_categories("Code", "Data", "Custom"));
+
+        MatcherAssert.assertThat(result.size(), Matchers.is(4));
+        MatcherAssert.assertThat(result.get(0), Matchers.is("Code"));
+        MatcherAssert.assertThat(result.get(1), Matchers.is("Data"));
+        MatcherAssert.assertThat(result.get(2), Matchers.is("Custom"));
+        MatcherAssert.assertThat(result.get(3), Matchers.is("Documentation"));
+    }
+
+    private void setup_override(String overrideValue) {
+        String currentLang = BundleUtil.getCurrentLocale().getLanguage();
+
+        Mockito.when(settingsServiceBean.get(DataFileCategoryServiceBean.FILE_CATEGORIES_KEY)).thenReturn(overrideValue);
+        Mockito.when(settingsServiceBean.get(DataFileCategoryServiceBean.FILE_CATEGORIES_KEY, currentLang, overrideValue)).thenReturn(overrideValue);
+    }
+
+    private List<DataFileCategory> setup_data_file_categories(String... names) {
+        return Arrays.stream(names).map(name -> {
+            DataFileCategory dataFileCategory = new DataFileCategory();
+            dataFileCategory.setName(name);
+            return dataFileCategory;
+        }).collect(Collectors.toList());
+    }
+
+}