keeps
diff --git a/‎CLAUDE.md‎
Lines changed: 15 additions & 13 deletions b/‎CLAUDE.md‎
Lines changed: 15 additions & 13 deletions
diff --git a/‎roda-common/roda-common-data/src/main/java/org/roda/core/data/common/RodaConstants.java‎
Lines changed: 11 additions & 0 deletions b/‎roda-common/roda-common-data/src/main/java/org/roda/core/data/common/RodaConstants.java‎
Lines changed: 11 additions & 0 deletions
diff --git a/‎roda-common/roda-common-data/src/main/java/org/roda/core/data/v2/index/FindRequest.java‎
Lines changed: 14 additions & 0 deletions b/‎roda-common/roda-common-data/src/main/java/org/roda/core/data/v2/index/FindRequest.java‎
Lines changed: 14 additions & 0 deletions
diff --git a/‎roda-core/roda-core-tests/src/main/java/org/roda/core/CorporaConstants.java‎
Lines changed: 4 additions & 0 deletions b/‎roda-core/roda-core-tests/src/main/java/org/roda/core/CorporaConstants.java‎
Lines changed: 4 additions & 0 deletions
diff --git a/‎roda-core/roda-core-tests/src/main/java/org/roda/core/index/EmailArchiveCrosswalkTest.java‎
Lines changed: 239 additions & 0 deletions b/‎roda-core/roda-core-tests/src/main/java/org/roda/core/index/EmailArchiveCrosswalkTest.java‎
Lines changed: 239 additions & 0 deletions
@@ -84,7 +84,7 @@ roda/
 
 1. **Java 21** (Oracle JDK) — strictly required for compilation
 2. **Maven 3.8.6+** — build tool
-3. **Docker & Docker Compose** — for running Solr, PostgreSQL, LDAP, etc.
+3. **Docker** — required for running the application locally and for tests (via Testcontainers — no manual `docker compose` needed for tests)
 4. **GitHub account with PAT** — required for GitHub Packages dependency resolution
 
 **Configure Maven for GitHub Packages** (`~/.m2/settings.xml`):
@@ -112,6 +112,10 @@ The PAT must have `read:packages` permission. Without this, the build will fail
 
 ### Starting Development Dependencies
 
+**For running tests:** No manual setup needed — tests use **Testcontainers** (`TestContainersManager`) which automatically starts ZooKeeper, Solr, PostgreSQL, Mailpit, ClamAV, and Siegfried as ephemeral Docker containers. The `RodaContainersLifecycleListener` TestNG suite listener wires this up before any test class loads. Docker must be running on the host, but no `docker compose` command is required.
+
+**For running the application locally:**
+
 ```bash
 # Create required data directories
 mkdir -p $HOME/.roda/data/{storage,staging-storage}
@@ -120,7 +124,7 @@ mkdir -p $HOME/.roda/data/{storage,staging-storage}
 docker compose -f deploys/standalone/docker-compose-dev.yaml up -d
 ```
 
-Services and ports:
+Services and ports (for local app, not tests):
 - ZooKeeper: `2181`
 - Apache Solr: `8983`
 - PostgreSQL: `5432`
@@ -196,25 +200,23 @@ mvn -f dev/codeserver gwt:codeserver -DrodaPath=$(pwd)
 
 ### Running Tests
 
+Tests use **Testcontainers** — no environment variables or `docker compose` setup required. Docker must be available on the host.
+
 ```bash
-# All tests (requires Docker services running)
-mvn clean test
+# All tests
+mvn clean test -Pcore
 
 # CI subset only (faster)
-mvn -Dtestng.groups="travis" -Denforcer.skip=true clean org.jacoco:jacoco-maven-plugin:prepare-agent test
+mvn -Dtestng.groups="travis" -Denforcer.skip=true clean org.jacoco:jacoco-maven-plugin:prepare-agent test -Pcore
+
+# Specific test class
+mvn -pl roda-core/roda-core-tests -am test -Dtest=NestedDocumentSearchTest -Dtestng.groups=dev -Denforcer.skip=true -DfailIfNoTests=false
 
 # Skip tests
 mvn clean package -Dmaven.test.skip=true
 ```
 
-### Required Environment Variables (for tests matching CI)
-
-```
-RODA_CORE_SOLR_TYPE=CLOUD
-SIEGFRIED_MODE=standalone
-```
-
-See `.github/workflows/CI.yml` for the full CI test environment configuration.
+See `.github/workflows/CI.yml` for the full CI configuration.
 
 ### Key Test Classes
 
 
@@ -176,6 +176,15 @@ public enum DateGranularity {
 
   public static final String UI_LISTS_PAGE_SIZE_INITIAL = "pageSize.initial";
   public static final String UI_LISTS_PAGE_SIZE_INCREMENT = "pageSize.increment";
+  public static final String UI_LISTS_INCLUDE_NESTED_DOCUMENTS = "includeNestedDocuments";
+
+  public static final String UI_CATALOGUE_VIRTUAL_PROPERTY = "ui.catalogue.virtual";
+  public static final String UI_LISTS_CATALOGUE_LABEL_I18N_PROPERTY = "catalogue.label.i18n";
+  public static final String UI_LISTS_CATALOGUE_ICON_PROPERTY = "catalogue.icon";
+  public static final String UI_LISTS_CATALOGUE_FILTER = "catalogue.filter";
+  public static final String UI_LISTS_CATALOGUE_CHILDOF_FILTER = "catalogue.childOf.filter";
+  public static final String UI_LISTS_CATALOGUE_CLICK_ACTION = "catalogue.click_action";
+  public static final String UI_LISTS_CATALOGUE_CLICK_ACTION_BROWSE_PARENT = "browse_parent";
 
   public static final String UI_ICONS_CLASS = "ui.icons.class";
   public static final String UI_SERVICE_DROPFOLDER_URL = "ui.service.dropfolder.url";
@@ -1980,6 +1989,8 @@ public String toString() {
   public static final String SEARCH_FIELD_TYPE_SUGGEST_FIELD = "suggestField";
   public static final String SEARCH_FIELD_TYPE_SUGGEST_PARTIAL = "suggestPartial";
   public static final String SEARCH_FIELD_TYPE_CONTROLLED = "controlled";
+  public static final String SEARCH_FIELD_NESTED_TYPE = "nestedType";
+  public static final String SEARCH_FIELD_NESTED_PARENT_TYPE = "nestedParentType";
   public static final String SEARCH_WITH_PREFILTER_HANDLER = "$prefilter";
   public static final String SEARCH_WITH_SAVED_HANDLER = "$savedSearch";
 
 
@@ -62,6 +62,8 @@ public class FindRequest extends CountRequest {
   private Long childrenLimit;
   @JsonProperty("childrenFilter")
   private Filter childrenFilter;
+  @JsonProperty("includeNestedDocuments")
+  private boolean includeNestedDocuments;
 
   public FindRequest() {
     super();
@@ -85,6 +87,7 @@ private FindRequest(FindRequestBuilder builder) {
     this.childrenFieldsToReturn = builder.childrenFieldsToReturn;
     this.childrenLimit = builder.childrenLimit;
     this.childrenFilter = builder.childrenFilter;
+    this.includeNestedDocuments = builder.includeNestedDocuments;
   }
 
   @JsonCreator
@@ -176,6 +179,10 @@ public Filter getChildrenFilter() {
     return childrenFilter;
   }
 
+  public boolean isIncludeNestedDocuments() {
+    return includeNestedDocuments;
+  }
+
   public void setChildrenFilter(Filter childrenFilter) {
     this.childrenFilter = childrenFilter;
   }
@@ -195,6 +202,7 @@ public static class FindRequestBuilder {
     private List<String> childrenFieldsToReturn;
     private Long childrenLimit;
     private Filter childrenFilter;
+    private boolean includeNestedDocuments;
 
     public FindRequestBuilder(@JsonProperty("filter") final Filter filter,
       @JsonProperty("onlyActive") boolean onlyActive) {
@@ -209,6 +217,7 @@ public FindRequestBuilder(@JsonProperty("filter") final Filter filter,
       this.fieldsToReturn = Collections.emptyList();
       this.collapse = null;
       this.children = false;
+      this.includeNestedDocuments = false;
     }
 
     public FindRequest build() {
@@ -269,5 +278,10 @@ public FindRequestBuilder withChildrenFilter(Filter childrenFilter) {
       this.childrenFilter = childrenFilter;
       return this;
     }
+
+    public FindRequestBuilder withIncludeNestedDocuments(boolean includeNestedDocuments) {
+      this.includeNestedDocuments = includeNestedDocuments;
+      return this;
+    }
   }
 }
@@ -94,6 +94,10 @@ public final class CorporaConstants {
 
   public static final String SOURCE_DESC_METADATA_CONTAINER = "DescriptiveMetadata";
   public static final String STRANGE_DESC_METADATA_FILE = "strange.xml";
+  public static final String EMAIL_ARCHIVE_FULL_FILE = "emailarchive_full.xml";
+  public static final String EMAIL_ARCHIVE_MINIMAL_FILE = "emailarchive_minimal.xml";
+  public static final String EMAIL_ARCHIVE_NO_EMAILS_FILE = "emailarchive_no_emails.xml";
+  public static final String EMAIL_ARCHIVE_METADATA_TYPE = "emailarchive";
 
   public static final String TEXT_XML = "text/xml";
   public static final String REPRESENTATION_1_PREMIS_EVENT_ID = "urn:roda:premis:event:roda_398";
 
@@ -0,0 +1,239 @@
+/**
+ * The contents of this file are subject to the license and copyright
+ * detailed in the LICENSE file at the root of the source
+ * tree and available online at
+ *
+ * https://github.com/keeps/roda
+ */
+package org.roda.core.index;
+
+import static org.testng.AssertJUnit.assertEquals;
+import static org.testng.AssertJUnit.assertNotNull;
+import static org.testng.AssertJUnit.assertNull;
+
+import java.io.IOException;
+import java.net.URISyntaxException;
+import java.net.URL;
+import java.nio.file.Path;
+import java.nio.file.Paths;
+import java.util.Collection;
+
+import org.apache.solr.common.SolrInputDocument;
+import org.apache.solr.common.SolrInputField;
+import org.roda.core.CorporaConstants;
+import org.roda.core.RodaCoreFactory;
+import org.roda.core.data.common.RodaConstants;
+import org.roda.core.data.exceptions.GenericException;
+import org.roda.core.data.exceptions.NotFoundException;
+import org.roda.core.data.exceptions.RODAException;
+import org.roda.core.index.utils.SolrUtils;
+import org.roda.core.storage.Binary;
+import org.roda.core.storage.DefaultStoragePath;
+import org.roda.core.storage.StorageService;
+import org.roda.core.storage.fs.FileStorageService;
+import org.testng.Assert;
+import org.testng.annotations.AfterMethod;
+import org.testng.annotations.BeforeClass;
+import org.testng.annotations.BeforeMethod;
+import org.testng.annotations.Test;
+
+@Test(groups = {RodaConstants.TEST_GROUP_ALL, RodaConstants.TEST_GROUP_DEV, RodaConstants.TEST_GROUP_TRAVIS})
+public class EmailArchiveCrosswalkTest {
+
+  private static StorageService corporaService;
+
+  @BeforeClass
+  public static void setUp() throws URISyntaxException, GenericException {
+    URL corporaURL = IndexServiceTest.class.getResource("/corpora");
+    Path corporaPath = Paths.get(corporaURL.toURI());
+    corporaService = new FileStorageService(corporaPath);
+  }
+
+  @BeforeMethod
+  public void init() {
+    RodaCoreFactory.instantiateTest(false, false, false, false, false, false, false);
+  }
+
+  @AfterMethod
+  public void cleanup() throws NotFoundException, GenericException, IOException {
+    RodaCoreFactory.shutdown();
+  }
+
+  // ---------------------------------------------------------------------------
+  // Full fixture — 3 emails
+  // ---------------------------------------------------------------------------
+
+  @Test
+  public void testFullCrosswalkProducesParentFields() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_FULL_FILE);
+
+    assertNotNull(doc);
+    assertFieldValue(doc, "custodian_txt", "João Silva");
+    assertFieldValue(doc, "emailAddress_s", "joao.silva@empresa.pt");
+    assertFieldValue(doc, "totalMessages_i", "3");
+    assertFieldValue(doc, "originalFormat_s", "PST");
+    assertFieldValue(doc, "archivingMotive_txt", "Offboarding");
+    assertFieldValue(doc, "content_type", "emailarchive");
+  }
+
+  @Test
+  public void testFullCrosswalkProducesDateFields() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_FULL_FILE);
+
+    assertNotNull(doc);
+    assertFieldValue(doc, "dateStart_dt", "2020-01-01T00:00:00Z");
+    assertFieldValue(doc, "dateEnd_dt", "2023-12-31T00:00:00Z");
+  }
+
+  @Test
+  public void testFullCrosswalkProducesThreeChildDocuments() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_FULL_FILE);
+
+    assertNotNull(doc);
+    SolrInputField emailsField = doc.getField("emails");
+    assertNotNull("'emails' field must be present for nested children", emailsField);
+
+    Collection<SolrInputDocument> children = getChildDocuments(emailsField);
+    assertEquals("Expected 3 child email documents", 3, children.size());
+  }
+
+  @Test
+  public void testFullCrosswalkFirstChildFields() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_FULL_FILE);
+    SolrInputDocument first = getChildAt(doc, 0);
+
+    assertFieldValue(first, "content_type", "email");
+    assertFieldValue(first, "messageId_s", "<msg001@empresa.pt>");
+    assertFieldValue(first, "subject_txt", "Quarterly Report Q1 2021");
+    assertFieldValue(first, "sender_s", "joao.silva@empresa.pt");
+    assertFieldValue(first, "sentDate_dt", "2021-03-15T09:42:00Z");
+    assertFieldValue(first, "folderPath_s", "Inbox/Projects");
+    assertFieldValue(first, "hasAttachments_b", "true");
+    assertFieldValue(first, "filePath_s", "Inbox/Projects/msg_001.eml");
+  }
+
+  @Test
+  public void testFullCrosswalkMultipleRecipients() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_FULL_FILE);
+    SolrInputDocument first = getChildAt(doc, 0);
+
+    // First email has two recipients: ana.costa and rui.pinto
+    SolrInputField recipientsField = first.getField("recipients_txt");
+    assertNotNull("recipients_txt field must be present", recipientsField);
+    Collection<?> values = recipientsField.getValues();
+    assertNotNull(values);
+    assertEquals("Expected 2 recipient values", 2, values.size());
+  }
+
+  @Test
+  public void testFullCrosswalkThirdChildFields() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_FULL_FILE);
+    SolrInputDocument third = getChildAt(doc, 2);
+
+    assertFieldValue(third, "subject_txt", "Budget Approval Request");
+    assertFieldValue(third, "folderPath_s", "Sent");
+    assertFieldValue(third, "filePath_s", "Sent/msg_003.eml");
+  }
+
+  // ---------------------------------------------------------------------------
+  // Minimal fixture — 1 email, only required fields
+  // ---------------------------------------------------------------------------
+
+  @Test
+  public void testMinimalCrosswalkProducesRequiredFields() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_MINIMAL_FILE);
+
+    assertNotNull(doc);
+    assertFieldValue(doc, "custodian_txt", "Jane Doe");
+    assertFieldValue(doc, "emailAddress_s", "jane.doe@example.org");
+    assertFieldValue(doc, "content_type", "emailarchive");
+  }
+
+  @Test
+  public void testMinimalCrosswalkOmitsAbsentOptionalFields() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_MINIMAL_FILE);
+
+    assertNotNull(doc);
+    assertNull("dateStart_dt should be absent when not in source", doc.getField("dateStart_dt"));
+    assertNull("dateEnd_dt should be absent when not in source", doc.getField("dateEnd_dt"));
+    assertNull("totalMessages_i should be absent when not in source", doc.getField("totalMessages_i"));
+    assertNull("originalFormat_s should be absent when not in source", doc.getField("originalFormat_s"));
+    assertNull("archivingMotive_txt should be absent when not in source", doc.getField("archivingMotive_txt"));
+  }
+
+  @Test
+  public void testMinimalCrosswalkProducesOneChild() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_MINIMAL_FILE);
+    SolrInputField emailsField = doc.getField("emails");
+    assertNotNull(emailsField);
+    assertEquals(1, getChildDocuments(emailsField).size());
+  }
+
+  @Test
+  public void testMinimalCrosswalkChildHasRequiredFields() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_MINIMAL_FILE);
+    SolrInputDocument child = getChildAt(doc, 0);
+
+    assertFieldValue(child, "content_type", "email");
+    assertFieldValue(child, "messageId_s", "<only-email@example.org>");
+    assertFieldValue(child, "subject_txt", "Hello World");
+    assertFieldValue(child, "hasAttachments_b", "false");
+  }
+
+  // ---------------------------------------------------------------------------
+  // No-emails fixture — mailbox with zero email records
+  // ---------------------------------------------------------------------------
+
+  @Test
+  public void testNoEmailsCrosswalkProducesParentFieldsOnly() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_NO_EMAILS_FILE);
+
+    assertNotNull(doc);
+    assertFieldValue(doc, "custodian_txt", "Empty Mailbox User");
+    assertFieldValue(doc, "content_type", "emailarchive");
+    assertFieldValue(doc, "totalMessages_i", "0");
+  }
+
+  @Test
+  public void testNoEmailsCrosswalkProducesNoChildDocumentsField() throws RODAException {
+    SolrInputDocument doc = getCrosswalkResult(CorporaConstants.EMAIL_ARCHIVE_NO_EMAILS_FILE);
+    assertNull("'emails' field must be absent when there are no child emails", doc.getField("emails"));
+  }
+
+  // ---------------------------------------------------------------------------
+  // Helpers
+  // ---------------------------------------------------------------------------
+
+  private SolrInputDocument getCrosswalkResult(String filename) throws RODAException {
+    try {
+      DefaultStoragePath path = DefaultStoragePath.parse(
+        CorporaConstants.SOURCE_DESC_METADATA_CONTAINER, filename);
+      Binary binary = corporaService.getBinary(path);
+      return SolrUtils.getDescriptiveMetadataFields(binary, CorporaConstants.EMAIL_ARCHIVE_METADATA_TYPE, null);
+    } catch (Exception e) {
+      Assert.fail("Unexpected exception loading fixture '" + filename + "': " + e.getMessage());
+      return null;
+    }
+  }
+
+  private void assertFieldValue(SolrInputDocument doc, String fieldName, String expectedValue) {
+    SolrInputField field = doc.getField(fieldName);
+    assertNotNull("Field '" + fieldName + "' must be present", field);
+    assertEquals("Field '" + fieldName + "' value mismatch", expectedValue, field.getValue().toString());
+  }
+
+  @SuppressWarnings("unchecked")
+  private Collection<SolrInputDocument> getChildDocuments(SolrInputField emailsField) {
+    Object value = emailsField.getValue();
+    assertNotNull("'emails' field value must not be null", value);
+    return (Collection<SolrInputDocument>) value;
+  }
+
+  private SolrInputDocument getChildAt(SolrInputDocument parent, int index) {
+    SolrInputField emailsField = parent.getField("emails");
+    assertNotNull(emailsField);
+    Collection<SolrInputDocument> children = getChildDocuments(emailsField);
+    return children.stream().skip(index).findFirst()
+      .orElseThrow(() -> new AssertionError("No child document at index " + index));
+  }
+}