Merge pull request #6774 from IQSS/2734-preserve-orig-filename

2734 preserve orig filename
IQSS · Apr 2, 2020 · d631b05 · d631b05
2 parents 3aede7e + 0856d47
commit d631b05
Show file tree

Hide file tree

Showing 8 changed files with 62 additions and 20 deletions.
diff --git a/src/main/java/edu/harvard/iq/dataverse/DataFile.java b/src/main/java/edu/harvard/iq/dataverse/DataFile.java
@@ -18,6 +18,7 @@
 import edu.harvard.iq.dataverse.util.BundleUtil;
 import edu.harvard.iq.dataverse.util.FileUtil;
 import edu.harvard.iq.dataverse.util.ShapefileHandler;
+import edu.harvard.iq.dataverse.util.StringUtil;
 import java.io.IOException;
 import java.util.List;
 import java.util.ArrayList;
@@ -398,6 +399,30 @@ public Long getOriginalFileSize() {
         }
         return null;
     }
+
+    public String getOriginalFileName() {
+        if (isTabularData()) {
+            DataTable dataTable = getDataTable();
+            if (dataTable != null) {
+                return dataTable.getOriginalFileName() != null ? dataTable.getOriginalFileName()
+                        : getDerivedOriginalFileName();
+            }
+        }
+        return null;
+    }
+
+
+    private String getDerivedOriginalFileName() {
+        FileMetadata fm = getFileMetadata();
+        String filename = fm.getLabel();
+        String originalExtension = FileUtil.generateOriginalExtension(getOriginalFileFormat());
+        String extensionToRemove = StringUtil.substringIncludingLast(filename, ".");
+        if (StringUtil.nonEmpty(extensionToRemove)) {
+            return filename.replaceAll(extensionToRemove + "$", originalExtension);
+        } else{
+            return filename + originalExtension ;
+        }        
+    }
 
     @Override
     public boolean isAncestorOf( DvObject other ) {

diff --git a/src/main/java/edu/harvard/iq/dataverse/DataTable.java b/src/main/java/edu/harvard/iq/dataverse/DataTable.java
@@ -109,6 +109,12 @@ public DataTable() {
 
     private Long originalFileSize; 
 
+    /**
+     * originalFileName: the file name upon upload/ingest
+     */
+    @Column( nullable = true )
+    private String originalFileName;
+
     /*
      * Getter and Setter methods:
      */
@@ -194,6 +200,14 @@ public String getOriginalFormatVersion() {
     public void setOriginalFormatVersion(String originalFormatVersion) {
         this.originalFormatVersion = originalFormatVersion;
     }
+
+    public String getOriginalFileName() {
+        return originalFileName;
+    }
+
+    public void setOriginalFileName(String originalFileName) {
+        this.originalFileName = originalFileName;
+    }
 
     /* 
      * Custom overrides for hashCode(), equals() and toString() methods:

diff --git a/src/main/java/edu/harvard/iq/dataverse/dataaccess/StoredOriginalFile.java b/src/main/java/edu/harvard/iq/dataverse/dataaccess/StoredOriginalFile.java
@@ -81,27 +81,14 @@ public static StorageIO<DataFile> retreive(StorageIO<DataFile> storageIO) {
             inputStreamIO.setMimeType("application/x-unknown");
         }
 
-        String fileName = storageIO.getFileName();
-        if (fileName != null) {
-            if (originalMimeType != null) {
-                String origFileExtension = generateOriginalExtension(originalMimeType);
-                inputStreamIO.setFileName(fileName.replaceAll(".tab$", origFileExtension));
-            } else {
-                inputStreamIO.setFileName(fileName.replaceAll(".tab$", ""));
-            }
-        }
+        inputStreamIO.setFileName(dataFile.getOriginalFileName());
 
         return inputStreamIO;
 
     }
 
-    // TODO: 
-    // do what the comment below says - move this code into the file util, 
-    // or something like that!
-    // -- L.A. 4.0 beta15
-    // Shouldn't be here; should be part of the DataFileFormatType, or 
-    // something like that... 
-
+    /* 
+            DataFile.getOriginalFileName() method replaces this code
     private static String generateOriginalExtension(String fileType) {
 
         if (fileType.equalsIgnoreCase("application/x-spss-sav")) {
@@ -125,5 +112,5 @@ private static String generateOriginalExtension(String fileType) {
         }
         logger.severe(fileType + " does not have an associated file extension");
         return "";
-    }
+    } */
 }
diff --git a/src/main/java/edu/harvard/iq/dataverse/engine/command/impl/UningestFileCommand.java b/src/main/java/edu/harvard/iq/dataverse/engine/command/impl/UningestFileCommand.java
@@ -138,16 +138,23 @@ protected void executeImpl(CommandContext ctxt) throws CommandException {
         // Modify the file name - which is stored in FileMetadata, and there
         // could be more than one: 
 
-        String originalExtension = FileUtil.generateOriginalExtension(originalFileFormat);
-
+       // String originalExtension = FileUtil.generateOriginalExtension(originalFileFormat);
+        String originalFileName = uningest.getOriginalFileName();
         for (FileMetadata fm : uningest.getFileMetadatas()) {
+
+            fm.setLabel(originalFileName);
+            ctxt.em().merge(fm);
+
+            /* 
+            getOriginalFileName method replaces this code
             String filename = fm.getLabel();
             String extensionToRemove = StringUtil.substringIncludingLast(filename, ".");
             if (StringUtil.nonEmpty(extensionToRemove)) {
                 String newFileName = filename.replace(extensionToRemove, originalExtension);
                 fm.setLabel(newFileName);
                 ctxt.em().merge(fm);
             }
+             */
 
             DatasetVersion dv = fm.getDatasetVersion();
 

diff --git a/src/main/java/edu/harvard/iq/dataverse/ingest/IngestServiceBean.java b/src/main/java/edu/harvard/iq/dataverse/ingest/IngestServiceBean.java
@@ -930,7 +930,8 @@ public boolean ingestAsTabular(Long datafile_id) {
 
                 dataFile.setDataTable(tabDataIngest.getDataTable());
                 tabDataIngest.getDataTable().setDataFile(dataFile);
-
+                tabDataIngest.getDataTable().setOriginalFileName(originalFileName);
+
                 try {
                     produceSummaryStatistics(dataFile, tabFile);
                     produceFrequencyStatistics(dataFile, tabFile);

diff --git a/src/main/java/edu/harvard/iq/dataverse/util/json/JsonPrinter.java b/src/main/java/edu/harvard/iq/dataverse/util/json/JsonPrinter.java
@@ -594,6 +594,7 @@ public static JsonObjectBuilder json(DataFile df, FileMetadata fileMetadata) {
                 .add("originalFileFormat", df.getOriginalFileFormat())
                 .add("originalFormatLabel", df.getOriginalFormatLabel())
                 .add ("originalFileSize", df.getOriginalFileSize())
+                .add("originalFileName", df.getOriginalFileName())
                 .add("UNF", df.getUnf())
                 //---------------------------------------------
                 // For file replace: rootDataFileId, previousDataFileId

diff --git a/src/main/resources/db/migration/V4.20.0.1__2734-alter-data-table-add-orig-file-name.sql b/src/main/resources/db/migration/V4.20.0.1__2734-alter-data-table-add-orig-file-name.sql
@@ -0,0 +1,2 @@
+
+ALTER TABLE datatable ADD COLUMN  IF NOT EXISTS originalfilename character varying(255);
diff --git a/src/test/java/edu/harvard/iq/dataverse/util/json/JsonPrinterTest.java b/src/test/java/edu/harvard/iq/dataverse/util/json/JsonPrinterTest.java
@@ -4,6 +4,7 @@
 import edu.harvard.iq.dataverse.DataFile;
 import edu.harvard.iq.dataverse.DataFileCategory;
 import edu.harvard.iq.dataverse.DataFileTag;
+import edu.harvard.iq.dataverse.DataTable;
 import edu.harvard.iq.dataverse.Dataset;
 import edu.harvard.iq.dataverse.DatasetField;
 import edu.harvard.iq.dataverse.DatasetFieldCompoundValue;
@@ -154,6 +155,9 @@ public void testGetFileCategories() {
         tag.setTypeByLabel("Survey");
         dataFileTags.add(tag);
         dataFile.setTags(dataFileTags);
+        DataTable dt = new DataTable();
+        dataFile.setDataTable(dt);
+        dataFile.getDataTable().setOriginalFileName("50by1000.dta");
         fmd.setDatasetVersion(dsVersion);
         fmd.setDataFile(dataFile);
         List<DataFileCategory> fileCategories = new ArrayList<>();
@@ -171,6 +175,7 @@ public void testGetFileCategories() {
         assertEquals("", jsonObject.getJsonObject("dataFile").getString("filename"));
         assertEquals(-1, jsonObject.getJsonObject("dataFile").getInt("filesize"));
         assertEquals(-1, jsonObject.getJsonObject("dataFile").getInt("rootDataFileId"));
+        assertEquals("50by1000.dta", jsonObject.getJsonObject("dataFile").getString("originalFileName"));
         assertEquals("Survey", jsonObject.getJsonObject("dataFile").getJsonArray("tabularTags").getString(0));
     }
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,2 @@

		ALTER TABLE datatable ADD COLUMN IF NOT EXISTS originalfilename character varying(255);