Merge pull request IQSS#12145 from QualitativeDataRepository/IQSS/7956-show_original_file_format

landreev · web-flow · commit 6b188a44db98 · 2026-04-21T09:47:14.000-04:00
IQSS/7956 show original file format
diff --git a/doc/release-notes/7956-show_original_file_format.md b/doc/release-notes/7956-show_original_file_format.md
@@ -0,0 +1,2 @@
+This release changes the display behavior of the file, dataset and collection ("dataverse") pages for tabular files. They will show the original file name and information, and will allow editing of the original file name. The ingested *.tab version is still available from the download menu. 
+ 
diff --git a/src/main/java/edu/harvard/iq/dataverse/DataCitation.java b/src/main/java/edu/harvard/iq/dataverse/DataCitation.java
@@ -141,7 +141,7 @@ public DataCitation(FileMetadata fm, boolean direct) {
         getCommonValuesFrom(dsv);
 
         // file Title for direct File citation
-        fileTitle = fm.getLabel();
+        fileTitle = fm.getLabelForOriginal();
         DataFile df = fm.getDataFile();
 
         // File description
diff --git a/src/main/java/edu/harvard/iq/dataverse/DataFile.java b/src/main/java/edu/harvard/iq/dataverse/DataFile.java
@@ -547,6 +547,11 @@ public Long getOriginalFileSize() {
         return null;
     }
     
+    public String getFriendlyOriginalFileSize() {
+        Long size = (getOriginalFileSize()==null) ? filesize : getOriginalFileSize();
+        return getFriendlySize(size);
+    }
+    
     public String getOriginalFileName() {
         if (isTabularData()) {
             DataTable dataTable = getDataTable();
@@ -559,7 +564,7 @@ public String getOriginalFileName() {
     }
 
     
-    private String getDerivedOriginalFileName() {
+    public String getDerivedOriginalFileName() {
         FileMetadata fm = getFileMetadata();
         String filename = fm.getLabel();
         String originalExtension = FileUtil.generateOriginalExtension(getOriginalFileFormat());
@@ -709,8 +714,12 @@ public void setFilesize(long filesize) {
      * @return 
      */
     public String getFriendlySize() {
-        if (filesize != null) {
-            return FileSizeChecker.bytesToHumanReadable(filesize);
+        return getFriendlySize(filesize);
+    }
+    
+    private String getFriendlySize(Long size) {
+        if (size != null) {
+            return FileSizeChecker.bytesToHumanReadable(size);
         } else {
             return BundleUtil.getStringFromBundle("file.sizeNotAvailable");
         }
diff --git a/src/main/java/edu/harvard/iq/dataverse/FileLabelValidator.java b/src/main/java/edu/harvard/iq/dataverse/FileLabelValidator.java
@@ -0,0 +1,34 @@
+package edu.harvard.iq.dataverse;
+
+import jakarta.validation.ConstraintValidator;
+import jakarta.validation.ConstraintValidatorContext;
+
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
+public class FileLabelValidator implements ConstraintValidator<ValidateDataFileLabel, String> {
+
+    @Override
+    public void initialize(ValidateDataFileLabel constraintAnnotation) {
+        
+    }
+
+    @Override
+    public boolean isValid(String value, ConstraintValidatorContext context) {
+        return isFileLabelValid(value, context);
+        
+    }
+    
+    public static boolean isFileLabelValid(String value, ConstraintValidatorContext context) {
+
+        if (value == null || value.isEmpty()) {
+            return true;
+        }
+        String validCharacters = "^[^:<>;#/\"\\*\\|\\?\\\\]*$";
+        Pattern p = Pattern.compile(validCharacters);
+        Matcher m = p.matcher(value);
+        return m.matches();
+
+    }
+    
+}
diff --git a/src/main/java/edu/harvard/iq/dataverse/FileMetadata.java b/src/main/java/edu/harvard/iq/dataverse/FileMetadata.java
@@ -39,6 +39,7 @@
 import jakarta.persistence.NamedNativeQuery;
 import jakarta.persistence.OneToMany;
 import jakarta.persistence.OrderBy;
+import jakarta.persistence.PostLoad;
 import jakarta.persistence.SqlResultSetMapping;
 import jakarta.persistence.Table;
 import jakarta.persistence.Transient;
@@ -153,6 +154,18 @@ public class FileMetadata implements Serializable {
     @OneToMany (mappedBy="fileMetadata", cascade={ CascadeType.REMOVE, CascadeType.MERGE,CascadeType.PERSIST})
     private Collection<VariableMetadata> variableMetadatas;
         
+    // A transient field is needed for JSF UI - validation errors on label do not get routed directly to the input for labelNoExtension, causing rollback.
+    // With a separate transient field kept in sync with label, the validation can be done on the labelNoExtension field, which avoids the issue and allows proper validation.
+    @Transient
+    @ValidateDataFileLabel(message = "{filename.illegalCharacters}")
+    String labelNoExtension;
+    
+    // Initialize the labelNoExtension from label after loading the entity
+    @PostLoad
+    public void postLoad() {
+        getLabelNoExtension();
+    }
+
     /**
      * Creates a copy of {@code this}, with identical business logic fields, making the bi-drectional connections to the specified version.
      * 
@@ -173,13 +186,46 @@ public FileMetadata createCopyInVersion(DatasetVersion dsv) {
     }
     
     public String getLabel() {
+        getLabelNoExtension();
         return label;
     }
     
     public void setLabel(String label) {
         this.label = label;
+        getLabelNoExtension();
+    }
+
+
+    public String getLabelNoExtension() {
+        int last = label.lastIndexOf(".");
+        labelNoExtension = (last == -1) ? label : label.substring(0, last);
+        return labelNoExtension;
+    }
+
+    public String getOriginalExtension() {
+        String origFilename = getLabelForOriginal();
+        int last = origFilename.lastIndexOf(".");
+        return (last == -1) ? "" : origFilename.substring(last);
     }
 
+    public void setLabelNoExtension(String name) {
+        labelNoExtension = name;
+        int last = this.label.lastIndexOf(".");
+        if (last == -1) {
+            this.label = name;
+        } else {
+            this.label = name + this.label.substring(last);
+        }
+    }
+    
+    public String getLabelForOriginal() {
+        if(dataFile.isTabularData()) {
+            return dataFile.getDerivedOriginalFileName();
+        } else {
+            return label;
+        }
+    }
+    
     public FileMetadata() {
         variableMetadatas = new ArrayList<VariableMetadata>();
         varGroups = new ArrayList<VarGroup>();
diff --git a/src/main/java/edu/harvard/iq/dataverse/ValidateDataFileLabel.java b/src/main/java/edu/harvard/iq/dataverse/ValidateDataFileLabel.java
@@ -0,0 +1,25 @@
+package edu.harvard.iq.dataverse;
+
+import jakarta.validation.Constraint;
+import jakarta.validation.Payload;
+
+import java.lang.annotation.Documented;
+import java.lang.annotation.Retention;
+import java.lang.annotation.Target;
+
+import static java.lang.annotation.ElementType.FIELD;
+import static java.lang.annotation.RetentionPolicy.RUNTIME;
+
+@Target({FIELD})
+@Retention(RUNTIME)
+@Constraint(validatedBy = {FileLabelValidator.class})
+@Documented
+public @interface ValidateDataFileLabel {
+    
+    String message() default "Failed Validation for Validate Data File Label";
+
+    Class<?>[] groups() default {};
+
+    Class<? extends Payload>[] payload() default {};
+    
+}
diff --git a/src/main/java/edu/harvard/iq/dataverse/search/IndexServiceBean.java b/src/main/java/edu/harvard/iq/dataverse/search/IndexServiceBean.java
@@ -1611,7 +1611,7 @@ public SolrInputDocuments toSolrDocs(IndexableDataset indexableDataset, Set<Long
 
                     String filenameCompleteFinal = "";
                     if (fileMetadata != null) {
-                        String filenameComplete = fileMetadata.getLabel();
+                        String filenameComplete = fileMetadata.getLabelForOriginal();
                         if (filenameComplete != null) {
                             String filenameWithoutExtension = "";
                             // String extension = "";
diff --git a/src/main/webapp/editFilesFragment.xhtml b/src/main/webapp/editFilesFragment.xhtml
@@ -424,13 +424,14 @@
                                         </div>
                                     </div>
                                     <div class="col-md-10 file-metadata-block">
-                                        <label class="control-label" for="datasetForm:filesTable:#{rowNum}:fileName" style="margin-right:1em;margin-bottom:.5em;">
+                                        <label class="control-label" jsf:for="fileName" style="margin-right:1em;margin-bottom:.5em;">
                                             #{bundle['file.fileName']}
                                         </label>
-                                        <p:inputText id="fileName" value="#{fileMetadata.label}" style="width:60%; margin-bottom:.5em;" disabled="#{fileMetadata.dataFile.filePackage}"/>
+                                        <p:inputText id="fileName" value="#{fileMetadata.labelNoExtension}" style="width:60%; margin-bottom:.5em;" disabled="#{fileMetadata.dataFile.filePackage}"/>
+                                        <h:outputText value="#{fileMetadata.originalExtension}"/>
                                         <p:message for="datasetForm:filesTable:#{rowNum}:fileName"/>
 
-                                        <label class="control-label" for="datasetForm:filesTable:#{rowNum}:fileDirectoryName" style="margin-right:1em;margin-bottom:.5em;">
+                                        <label class="control-label" jsf:for="fileDirectoryName" style="margin-right:1em;margin-bottom:.5em;">
                                             #{bundle['file.metadataTab.fileMetadata.hierarchy.label']}    
                                             <span class="glyphicon glyphicon-question-sign tooltip-icon"
                                                   data-toggle="tooltip" data-placement="auto right" data-original-title="#{bundle['file.metadataTab.fileMetadata.hierarchy.tip']}"></span>
diff --git a/src/main/webapp/file-info-fragment.xhtml b/src/main/webapp/file-info-fragment.xhtml
@@ -67,19 +67,19 @@
                 <ui:fragment rendered="#{!editDatafilesPage}">
                     <!-- conditional render in file page links, if file has global id, use that; if not, we use database id -->
                     <a href="#{widgetWrapper.wrapURL('/file.xhtml?'.concat(fileIdParam).concat(fileIdValue).concat('&amp;version=').concat(fileMetadata.datasetVersion.friendlyVersionNumber))}">
-                        #{fileMetadata.label}
+                        #{fileMetadata.labelForOriginal}
                     </a>
                 </ui:fragment>
                 <ui:fragment rendered="#{editDatafilesPage}">
-                    <span>#{fileMetadata.label}</span>
+                    <span>#{fileMetadata.labelForOriginal}</span>
                 </ui:fragment>
             </div>
             <!-- HIERARCHY + TYPE + SIZE + DATE + CHECKSUM -->
             <div class="text-muted small">
-                <!-- id:#{dataFile.id} -->
+                <!-- id:#{fileMetadata.dataFile.id} -->
                 <h:outputText id="fileHierarchy" styleClass="show" value="#{fileMetadata.directoryLabel}/" rendered="#{hasDirectoryLabel}"/>
-                <h:outputText id="fileType" value="#{dataFile.tabularData ? bundle['file.type.tabularData'] : dataFile.friendlyType}"/>
-                <h:outputText id="fileSize" value=" - #{dataFile.friendlySize}"/>
+                <h:outputText id="fileType" value="#{dataFile.friendlyType}"/>
+                <h:outputText id="fileSize" value=" - #{dataFile.friendlyOriginalFileSize}"/>
                 <div class="dateCreatePublish-block">
                     <h:outputText id="fileCreatePublishDate" value="#{isReleased ? ((dataFile.embargo!=null) ? bundle['metadatareleased'] : bundle['published']) : bundle['deposited']} #{fileMetadata.getFileDateToDisplay()}" rendered="#{hasFileId}"/>
                     <h:outputText id="embargoDate" class="embargoright" value="#{isReleased ? ((dataFileServiceBean.isActivelyEmbargoed(fileMetadata)) ? bundle['embargoed.until'] : bundle['embargoed.wasthrough']):bundle['embargoed.willbeuntil']} #{dataFile.embargo.formattedDateAvailable}" rendered="#{hasEmbargo}"/>
@@ -91,7 +91,7 @@
                     </h:outputFormat>
                 </div>
                 <!-- CHECKSUM -->
-                <div class="checksum-block" jsf:rendered="#{!dataFile.tabularData and !(empty dataFile.checksumValue)}">
+                <div class="checksum-block" jsf:rendered="#{!(empty dataFile.checksumValue)}">
                     <span class="#{!editDatafilesPage ? 'checksum-truncate checksum-tooltip' : ''}" style="margin-right:3px;"
                                   data-toggle="tooltip" data-placement="top" data-html="true"
                                   data-clipboard-action="copy"
@@ -105,6 +105,7 @@
                 </div>
                 <!-- TABULAR VARIABLES, OBSERVATIONS + UNF -->
                 <div class="unf-block" jsf:rendered="#{isTabularData}">
+                    <h:outputText value="#{bundle['file.type.tabularData']}: "/>
                     <h:outputText id="fileNumVars" value="#{dataFile.dataTable.varQuantity} #{bundle['file.metaData.dataFile.dataTab.variables']}, "/>
                     <h:outputText id="fileNumObs" value="#{dataFile.dataTable.caseQuantity} #{bundle['file.metaData.dataFile.dataTab.observations']} "/>
                     <span jsf:id="fileUNF" class="#{!editDatafilesPage ? 'checksum-truncate checksum-tooltip' : ''}" style="margin-right:3px;"
diff --git a/src/main/webapp/file.xhtml b/src/main/webapp/file.xhtml
@@ -15,7 +15,7 @@
 
     <h:body>
         <ui:composition template="/dataverse_template.xhtml">
-            <ui:param name="pageTitle" value="#{FilePage.fileMetadata.dataFile.displayName} - #{FilePage.file.owner.owner.displayName}"/>
+            <ui:param name="pageTitle" value="#{FilePage.fileMetadata.labelForOriginal} - #{FilePage.file.owner.owner.displayName}"/>
             <ui:param name="dataverse" value="#{FilePage.file.owner.owner}"/>
             <ui:param name="showDataverseHeader" value="#{!FilePage.isAnonymizedAccess()}"/>
             <ui:param name="showMessagePanel" value="#{true}"/>
@@ -45,7 +45,7 @@
                         <div class="col-xs-12">
                             <div class="row">
                                 <div id="file-title-block" class="col-xs-12 margin-bottom-half">
-                                    <span class="file-title-label">#{FilePage.fileMetadata.label}</span>
+                                    <span class="file-title-label">#{FilePage.fileMetadata.labelForOriginal}</span>
 
                                     <!-- FILE LEVEL MSGs -->
                                     <div class="bg-info text-info text-center margin-bottom-half" jsf:rendered="#{FilePage.fileMetadata.dataFile.ingestProblem and FilePage.canUpdateDataset()}">
@@ -590,11 +590,11 @@
                                                         <td>#{FilePage.file.retention.reason}
                                                         </td>
                                                     </tr>
-                                                    <tr id="fileSizeBlock" jsf:rendered="#{!(empty FilePage.file.friendlySize)}">
+                                                    <tr id="fileSizeBlock" jsf:rendered="#{!(empty FilePage.file.friendlyOriginalFileSize)}">
                                                         <th scope="row">
                                                             #{bundle['file.metadataTab.fileMetadata.size.label']}
                                                         </th>
-                                                        <td>#{FilePage.file.friendlySize}</td>
+                                                        <td>#{FilePage.file.friendlyOriginalFileSize}</td>
                                                     </tr>
                                                     <tr id="fileTypeBlock" jsf:rendered="#{!(empty FilePage.file.friendlyType)}">
                                                         <th scope="row">
diff --git a/src/test/java/edu/harvard/iq/dataverse/api/SearchIT.java b/src/test/java/edu/harvard/iq/dataverse/api/SearchIT.java
@@ -2220,7 +2220,7 @@ public void testTabularFiles() throws IOException {
         search.prettyPrint();
         search.then().assertThat()
                 .statusCode(OK.getStatusCode())
-                .body("data.items[0].name", is("data.tab"))
+                .body("data.items[0].name", is("data.csv")) // as of 6.11, we are indexing the original names of ingested tab. files
                 .body("data.items[0].variables", is(4))
                 .body("data.items[0].observations", is(3));
     }

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,2 @@`
	`1`	`+This release changes the display behavior of the file, dataset and collection ("dataverse") pages for tabular files. They will show the original file name and information, and will allow editing of the original file name. The ingested *.tab version is still available from the download menu.`
	`2`	`+`