Manage Guestbooks page performance (#12053)

* A quick pr speeding up Manage Guestbooks page.

* Removed the code that allowed to run "select count(o) from guestbookresponse", on the entire table.

* Added another useful index on guestbookresponse table (that we've been using at HDV since April)

Author: landreev

doc/release-notes/6.9-release-notes.md

@@ -95,6 +95,7 @@ See [the guides](https://guides.dataverse.org/en/6.9/developers/workflows.html#c
 - In prior versions of Dataverse, publishing a dataset via the superuser-only update-current-version option would not set the current curation status (if enabled/used) to none/empty and, in v6.7, would not maintain the curation status history. These issues are now resolved and the update-current-version option works the same as normal publication of a new version with regard to curation status. See #11783 and #11784.
 - This release fixes problems with guestbook questions being displayed at download when files are selected from the dataset files table when guestbook-at-request is enabled and not displaying when they should when access is requested from the file page. See #11800, #11808, and #11835.
 - The optional Croissant exporter has been updated to 0.1.6 to prevent variable names, variable descriptions, and variable types from being exposed for restricted files. See https://github.com/gdcc/exporter-croissant/pull/20 and #11752.
+- Manage Gustbooks page was optimized to load much faster for collections with large numbers of downloads recorded. 
 
 ## API Updates
 

src/main/java/edu/harvard/iq/dataverse/GuestbookResponse.java

@@ -32,7 +32,9 @@
         @Index(columnList = "datafile_id"),
         @Index(columnList = "datasetversion_id"),
         @Index(columnList = "authenticateduser_id"),
-        @Index(columnList = "dataset_id")
+        @Index(columnList = "dataset_id"),
+        @Index(columnList = "dataset_id, guestbook_id", name="INDEX_GUESTBOOKRESPONSE_dataset_id_guestbook_id"),
+        @Index(columnList = "dataset_id, eventtype", name="INDEX_GUESTBOOKRESPONSE_dataset_id_eventtype")
 })
 
 @NamedQueries(

src/main/java/edu/harvard/iq/dataverse/GuestbookResponseServiceBean.java

@@ -488,17 +488,25 @@ public Long findCount30Days(Long dataverseId) {
         return (Long) query.getSingleResult();
     }
 
-    public Long findCountAll() {
-        return findCountAll(null);
-    }
-
     public Long findCountAll(Long dataverseId) {
-        String queryString;
-        if (dataverseId != null) {
-            queryString = "select count(o.id) from GuestbookResponse  o,  DvObject v where o.dataset_id = v.id and v.owner_id = " + dataverseId + " ";
-        } else {
-            queryString = "select count(o.id) from GuestbookResponse  o ";
+      
+        if (dataverseId == null) {
+            return null;
         }
+        
+        // Note that this method used to support NULL dataverseId, 
+        // in which case it counted ALL the guestbookresponse rows
+        // for the entire instance: 
+        // queryString = "select count(o.id) from GuestbookResponse  o ";
+        // I removed this code (it was not being used, thankfully) since
+        // the query can be insanely expensive on a large production table. 
+        // That's why we use a stored procedure to "estimate" its size, in 
+        // the dedicated getTotalDownloadCount() method further below, for 
+        // example, when we need to show the total number of downloads on 
+        // the homepage. (L.A.)
+        
+        String queryString = "select count(o.id) from GuestbookResponse  o, DvObject v, Dataset d where o.dataset_id = v.id and v.id = d.id and v.owner_id = " + dataverseId + " ";
+            
 
         Query query = em.createNativeQuery(queryString);
         return (Long) query.getSingleResult();

src/main/resources/db/migration/V6.8.0.4.sql

@@ -0,0 +1,2 @@
+CREATE INDEX IF NOT EXISTS  INDEX_GUESTBOOKRESPONSE_dataset_id_guestbook_id ON GUESTBOOKRESPONSE (dataset_id, guestbook_id);
+CREATE INDEX IF NOT EXISTS  INDEX_GUESTBOOKRESPONSE_dataset_id_eventtype ON GUESTBOOKRESPONSE (dataset_id, eventtype);