feat(search): extend hasText to match sender/receiver/tag names, add hasTagPartial

- hasText now JOINs sender (LEFT JOIN) and uses EXISTS subqueries for
  receivers and tags to avoid duplicate rows
- hasTagPartial added for live debounced tag text filter (ILIKE partial match)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/src/main/java/org/raddatz/familienarchiv/repository/DocumentSpecifications.java

@@ -8,24 +8,58 @@ import java.util.UUID;
 
 import org.raddatz.familienarchiv.model.Document;
 import org.raddatz.familienarchiv.model.DocumentStatus;
+import org.raddatz.familienarchiv.model.Person;
 import org.raddatz.familienarchiv.model.Tag;
 import org.springframework.data.jpa.domain.Specification;
 import org.springframework.util.StringUtils;
 
 public class DocumentSpecifications {
 
-    // Filtert nach Text (in Titel, Dateiname oder Transkription)
+    // Filtert nach Text (in Titel, Dateiname, Transkription, Ort, Absender- und Empfängername, Tags)
     public static Specification<Document> hasText(String text) {
         return (root, query, cb) -> {
             if (!StringUtils.hasText(text))
                 return null;
             String likePattern = "%" + text.toLowerCase() + "%";
 
+            // LEFT JOIN on sender (ManyToOne — no duplicate rows)
+            Join<Document, Person> senderJoin = root.join("sender", JoinType.LEFT);
+
+            // EXISTS subquery for receiver name — avoids duplicate rows for multi-receiver docs
+            Subquery<Long> receiverSub = query.subquery(Long.class);
+            Root<Document> receiverRoot = receiverSub.from(Document.class);
+            Join<Document, Person> receiverJoin = receiverRoot.join("receivers");
+            receiverSub.select(cb.literal(1L))
+                    .where(
+                            cb.equal(receiverRoot.get("id"), root.get("id")),
+                            cb.or(
+                                    cb.like(cb.lower(receiverJoin.get("lastName")), likePattern),
+                                    cb.like(cb.lower(receiverJoin.get("firstName")), likePattern)
+                            )
+                    );
+
+            // EXISTS subquery for tag name — avoids duplicate rows for multi-tag docs
+            Subquery<Long> tagSub = query.subquery(Long.class);
+            Root<Document> tagRoot = tagSub.from(Document.class);
+            Join<Document, Tag> tagJoin = tagRoot.join("tags");
+            tagSub.select(cb.literal(1L))
+                    .where(
+                            cb.equal(tagRoot.get("id"), root.get("id")),
+                            cb.like(cb.lower(tagJoin.get("name")), likePattern)
+                    );
+
+            query.distinct(true);
+
             return cb.or(
                     cb.like(cb.lower(root.get("title")), likePattern),
                     cb.like(cb.lower(root.get("originalFilename")), likePattern),
                     cb.like(cb.lower(root.get("transcription")), likePattern),
-                    cb.like(cb.lower(root.get("location")), likePattern));
+                    cb.like(cb.lower(root.get("location")), likePattern),
+                    cb.like(cb.lower(senderJoin.get("lastName")), likePattern),
+                    cb.like(cb.lower(senderJoin.get("firstName")), likePattern),
+                    cb.exists(receiverSub),
+                    cb.exists(tagSub)
+            );
         };
     }
 
@@ -55,13 +89,13 @@ public class DocumentSpecifications {
             return cb.lessThanOrEqualTo(root.get("documentDate"), end);
         };
     }
-    
+
     // Filtert nach Status
     public static Specification<Document> hasStatus(DocumentStatus status) {
         return (root, query, cb) -> status == null ? null : cb.equal(root.get("status"), status);
     }
 
-    // Filtert nach Schlagworten (UND-Verknüpfung)
+    // Filtert nach Schlagworten (UND-Verknüpfung, exakter Match)
     public static Specification<Document> hasTags(List<String> tags) {
         return (root, query, cb) -> {
             if (tags == null || tags.isEmpty())
@@ -72,15 +106,13 @@ public class DocumentSpecifications {
             for (String tagName : tags) {
                 if (!StringUtils.hasText(tagName)) continue;
 
-                // Subquery erstellen: "Gibt es für dieses Dokument (root.id) einen Tag mit dem Namen X?"
-                // Dies stellt sicher, dass ALLE Tags vorhanden sein müssen (AND Logik).
                 Subquery<Long> subquery = query.subquery(Long.class);
                 Root<Document> subRoot = subquery.from(Document.class);
                 Join<Document, Tag> subTags = subRoot.join("tags");
 
                 subquery.select(subRoot.get("id"))
                         .where(
-                                cb.equal(subRoot.get("id"), root.get("id")), // Korrelation zum Haupt-Query
+                                cb.equal(subRoot.get("id"), root.get("id")),
                                 cb.equal(cb.lower(subTags.get("name")), tagName.trim().toLowerCase())
                         );
 
@@ -90,5 +122,26 @@ public class DocumentSpecifications {
             return cb.and(predicates.toArray(new Predicate[0]));
         };
     }
-    
-}
+
+    // Filtert nach partiellem Tag-Namen (ILIKE) — für Live-Tag-Suche
+    public static Specification<Document> hasTagPartial(String tagQ) {
+        return (root, query, cb) -> {
+            if (!StringUtils.hasText(tagQ))
+                return null;
+            String likePattern = "%" + tagQ.toLowerCase() + "%";
+
+            Subquery<Long> subquery = query.subquery(Long.class);
+            Root<Document> subRoot = subquery.from(Document.class);
+            Join<Document, Tag> tagJoin = subRoot.join("tags");
+
+            subquery.select(cb.literal(1L))
+                    .where(
+                            cb.equal(subRoot.get("id"), root.get("id")),
+                            cb.like(cb.lower(tagJoin.get("name")), likePattern)
+                    );
+
+            return cb.exists(subquery);
+        };
+    }
+
+}

backend/src/test/java/org/raddatz/familienarchiv/repository/DocumentSpecificationsTest.java

@@ -250,6 +250,62 @@ class DocumentSpecificationsTest {
         assertThat(result).isEmpty();
     }
 
+    @Test
+    void hasText_findsByPartialSenderLastName() {
+        List<Document> result = documentRepository.findAll(Specification.where(hasText("üller")));
+        assertThat(result).extracting(Document::getTitle)
+                .containsExactlyInAnyOrder("Alter Brief", "Neuerer Brief");
+    }
+
+    @Test
+    void hasText_findsByPartialReceiverLastName() {
+        List<Document> result = documentRepository.findAll(Specification.where(hasText("schmid")));
+        assertThat(result).extracting(Document::getTitle).containsExactly("Alter Brief");
+    }
+
+    @Test
+    void hasText_findsByPartialTagName() {
+        List<Document> result = documentRepository.findAll(Specification.where(hasText("amili")));
+        assertThat(result).extracting(Document::getTitle)
+                .containsExactlyInAnyOrder("Alter Brief", "Familienfoto");
+    }
+
+    @Test
+    void hasText_doesNotProduceDuplicatesForDocumentWithMultipleReceivers() {
+        Person receiver2 = personRepository.save(Person.builder().firstName("Karl").lastName("Schmidt").build());
+        briefEarly.setReceivers(new java.util.HashSet<>(Set.of(receiver, receiver2)));
+        documentRepository.save(briefEarly);
+
+        List<Document> result = documentRepository.findAll(Specification.where(hasText("schmid")));
+        assertThat(result).hasSize(1);
+    }
+
+    // ─── hasTagPartial ────────────────────────────────────────────────────────
+
+    @Test
+    void hasTagPartial_returnsAllDocuments_whenTextIsNull() {
+        List<Document> result = documentRepository.findAll(Specification.where(hasTagPartial(null)));
+        assertThat(result).hasSize(3);
+    }
+
+    @Test
+    void hasTagPartial_findsByPartialTagName() {
+        List<Document> result = documentRepository.findAll(Specification.where(hasTagPartial("amili")));
+        assertThat(result).extracting(Document::getTitle).containsExactly("Alter Brief");
+    }
+
+    @Test
+    void hasTagPartial_isCaseInsensitive() {
+        List<Document> result = documentRepository.findAll(Specification.where(hasTagPartial("URLAUB")));
+        assertThat(result).extracting(Document::getTitle).containsExactly("Neuerer Brief");
+    }
+
+    @Test
+    void hasTagPartial_returnsEmpty_whenNoTagMatches() {
+        List<Document> result = documentRepository.findAll(Specification.where(hasTagPartial("xyz")));
+        assertThat(result).isEmpty();
+    }
+
     // ─── hasStatus ────────────────────────────────────────────────────────────
 
     @Test