Duplicated results in JPA Repository - postgresql

Question

I have a JPA Repository for my Document entity. And I have a problem with one long method, where I want to achieve something like full text search across multiple tables. Document Entity has some OneToMany relationships.
public interface DocumentRepository extends JpaRepository<Document, Long>, JpaSpecificationExecutor<Document> {
Page<Document> findByInstanceNumberContainingIgnoreCaseOrAttachmentsNameContainingIgnoreCaseOrDocumentIdEqualsOrHeaderAuthorNameContainingIgnoreCaseOrCommentsCommentContainingIgnoreCaseOrFormFieldsSvalueContainingIgnoreCaseOrItemListsRowsCellsSvalueContainingIgnoreCaseOrderByInstanceNumber(Pageable pageable, String instanceNumber, String attachmentName, Integer documentId, String authorName, String comment, String formContent, String cellContent);
}
Unfortunately, that method returns multiplied and duplicated results. For example, currently I have 120 records in my Documents table, but I am getting 10485638 results in this query.
Here is something what was generated and used to fetch results
select count(document0_.id) as col_0_0_ from documents document0_ left outer join attachments attachment1_ on document0_.id=attachment1_.document_id left outer join headers header2_ on document0_.header_id=header2_.id left outer join authors author3_ on header2_.author_id=author3_.id left outer join comments comments4_ on document0_.id=comments4_.document_id left outer join form_fields formfields5_ on document0_.id=formfields5_.document_id left outer join item_list itemlists6_ on document0_.id=itemlists6_.document_id left outer join rows rows7_ on itemlists6_.id=rows7_.item_list_id left outer join cells cells8_ on rows7_.id=cells8_.row_id where upper(document0_.instance_number) like upper('%TEST%') escape '\' or upper(attachment1_.name) like upper('%TEST%') escape '\' or document0_.external_document_id=123 or upper(author3_.name) like upper('%TEST%') escape '\' or upper(comments4_.comment) like upper('%TEST%') escape '\' or upper(formfields5_.svalue) like upper('%TEST%') escape '\' or upper(cells8_.svalue) like upper('%TEST%') escape '\''
What can I do to fix that issue and get only unique documents? I need paging, so I can not use Set here.
Edit:
Document Entity class (without getters, setters and constructors):
#Entity
#ToString
#AllArgsConstructor
#NoArgsConstructor
#Table(name = "documents", indexes = {
#Index(columnList = "external_document_id", unique = true),
#Index(columnList = "instance_number")
})
public class Document {
#Id
#GeneratedValue(strategy = GenerationType.SEQUENCE, generator = "documents_generator")
#SequenceGenerator(name="documents_generator", sequenceName = "documents_seq", allocationSize=50)
#Column(name = "id", nullable = false)
private Long id;
#Column(name = "external_document_id", nullable = false, unique=true)
private Integer documentId;
#Column(name = "version", nullable = false)
private Integer version;
#Column(name = "instance_number", nullable = false)
private String instanceNumber;
#ManyToOne(fetch = FetchType.LAZY, cascade = CascadeType.ALL)
#JoinColumn(name = "task_id")
private Task tasks;
#ManyToOne(fetch = FetchType.LAZY, cascade = CascadeType.ALL)
#JoinColumn(name = "header_id")
private Header header;
#OneToMany(mappedBy = "document", cascade = CascadeType.PERSIST)
#ToString.Exclude
#EqualsAndHashCode.Exclude
private Set<Comment> comments = new LinkedHashSet<>();
#OneToMany(mappedBy = "document", cascade = CascadeType.PERSIST)
#ToString.Exclude
#EqualsAndHashCode.Exclude
private Set<Attachment> attachments = new LinkedHashSet<>();
#OneToMany(mappedBy = "document", cascade = CascadeType.PERSIST)
#ToString.Exclude
#EqualsAndHashCode.Exclude
private Set<FormField> formFields = new LinkedHashSet<>();
#OneToMany(mappedBy = "document", cascade = CascadeType.PERSIST)
#ToString.Exclude
#EqualsAndHashCode.Exclude
private Set<ItemList> itemLists = new LinkedHashSet<>();
}