Merge pull request #8732 from tcoupin/tc-citationdate-harvested-dataset

Use production date in citation for harvested dataset
IQSS · Oct 12, 2022 · f6c677b · f6c677b
2 parents 9686ab8 + 9d0ea68
commit f6c677b
Show file tree

Hide file tree

Showing 3 changed files with 60 additions and 15 deletions.
diff --git a/doc/release-notes/8732-date-in-citation-harvested-datasets.md b/doc/release-notes/8732-date-in-citation-harvested-datasets.md
@@ -0,0 +1,7 @@
+Fix the year displayed in citation for harvested dataset, specialy for oai_dc format.
+
+For normal datasets, the date used is the "citation date" which is by default the publication date (the first release date) (https://guides.dataverse.org/en/latest/api/native-api.html?highlight=citationdate#set-citation-date-field-type-for-a-dataset).
+
+But for a harvested dataset, the distribution date is used instead and this date is not always present in the harvested metadata. With oai_dc format the date tag if used as production date.
+
+Now, the production date is used for harvested dataset in addition to distribution date.
diff --git a/src/main/java/edu/harvard/iq/dataverse/DataCitation.java b/src/main/java/edu/harvard/iq/dataverse/DataCitation.java
@@ -33,6 +33,7 @@
 import javax.xml.stream.XMLStreamWriter;
 
 import edu.harvard.iq.dataverse.util.BundleUtil;
+import edu.harvard.iq.dataverse.util.DateUtil;
 import org.apache.commons.text.StringEscapeUtils;
 import org.apache.commons.lang3.StringUtils;
 
@@ -713,25 +714,24 @@ private String flattenHtml(String html) {
 
     private Date getDateFrom(DatasetVersion dsv) {
         Date citationDate = null;
-        SimpleDateFormat sdf = new SimpleDateFormat("yyyy");
-        if (!dsv.getDataset().isHarvested()) {
-            citationDate = dsv.getCitationDate();
+
+        if (dsv.getDataset().isHarvested()) {
+            citationDate = DateUtil.parseDate(dsv.getProductionDate());
             if (citationDate == null) {
-                if (dsv.getDataset().getCitationDate() != null) {
-                    citationDate = dsv.getDataset().getCitationDate();
-                } else { // for drafts
-                    citationDate = dsv.getLastUpdateTime();
-                }
+                citationDate = DateUtil.parseDate(dsv.getDistributionDate());
             }
-        } else {
-            try {
-                citationDate= sdf.parse(dsv.getDistributionDate());
-            } catch (ParseException ex) {
-                // ignore
-            } catch (Exception ex) {
-                // ignore
+        }
+
+        if (citationDate == null) {
+            if (dsv.getCitationDate() != null) {
+                citationDate = dsv.getCitationDate();
+            } else if (dsv.getDataset().getCitationDate() != null) {
+                citationDate = dsv.getDataset().getCitationDate();
+            } else { // for drafts
+                citationDate = dsv.getLastUpdateTime();
             }
         }
+
         if (citationDate == null) {
             //As a last resort, pick the current date
             logger.warning("Unable to find citation date for datasetversion: " + dsv.getId());

diff --git a/src/main/java/edu/harvard/iq/dataverse/util/DateUtil.java b/src/main/java/edu/harvard/iq/dataverse/util/DateUtil.java
@@ -16,7 +16,9 @@
  */
 public class DateUtil {
 
+    public static String YEAR_PATTERN = "yyyy";
     public static String YEAR_DASH_MONTH_PATTERN = "yyyy-MM";
+    public static String YEAR_DASH_MONTH_DASH_DAY_PATTERN = "yyyy-MM-dd";
 
     public static String formatDate(Date dateToformat) {
         String formattedDate;
@@ -63,4 +65,40 @@ public static String formatDate(Timestamp datetimeToformat) {
          }
     }
 
+    public static Date parseDate(String dateString) {
+        SimpleDateFormat sdf;
+        Date date;
+
+        // YYYY-MM-DD
+        date = parseDate(dateString, YEAR_DASH_MONTH_DASH_DAY_PATTERN);
+        if (date != null) {
+            return date;
+        }
+
+        // YYYY-MM
+        date = parseDate(dateString, YEAR_DASH_MONTH_PATTERN);
+        if (date != null) {
+            return date;
+        }
+
+        // YYYT
+        date = parseDate(dateString, YEAR_PATTERN);
+        return date;
+
+    }
+
+    public static Date parseDate(String dateString, String format) {
+
+        try {
+            SimpleDateFormat sdf = new SimpleDateFormat(format);
+            Date date = sdf.parse(dateString);
+            return date;
+        } catch (ParseException ex) {
+            // ignore
+        } catch (Exception ex) {
+            // ignore
+        }
+        return null;
+    }
+
 }