UFAL/Matomo statistics with dimension (#813)

* Updated the version of matomo dependency and tried to change request from Custom Variables to Dimension * Added a custom dimension with item's handle URL * Send custom dimension also in oai tracker * Use only IPv4 address, the Matomo tracker has a problem with IPv6 * Do not change custom dimension when the Item is null * First custom dimension should have ID '1'. * Use a valid URL for Matomo tracker in the IT * Configure handle custom dimension ID in the clarin-dspace.cfg * Refactored ipv4 method to be more readable - return null --------- Co-authored-by: Juraj Roka <[email protected]> Co-authored-by: milanmajchrak <[email protected]> Co-authored-by: milanmajchrak <[email protected]>
dataquest-dev · Nov 29, 2024 · 3486d2d · 3486d2d
1 parent 9758648
commit 3486d2d
Show file tree

Hide file tree

Showing 5 changed files with 44 additions and 24 deletions.
diff --git a/dspace-api/pom.xml b/dspace-api/pom.xml
@@ -339,8 +339,8 @@
     <dependencies>
         <dependency>
             <groupId>org.piwik.java.tracking</groupId>
-            <artifactId>matomo-java-tracker</artifactId>
-            <version>2.0</version>
+            <artifactId>matomo-java-tracker-java11</artifactId>
+            <version>3.4.0</version>
         </dependency>
         <dependency>
             <groupId>org.apache.logging.log4j</groupId>

diff --git a/dspace-api/src/main/java/org/dspace/app/statistics/clarin/ClarinMatomoBitstreamTracker.java b/dspace-api/src/main/java/org/dspace/app/statistics/clarin/ClarinMatomoBitstreamTracker.java
@@ -9,6 +9,7 @@
 
 import java.sql.SQLException;
 import java.text.MessageFormat;
+import java.util.LinkedHashMap;
 import java.util.List;
 import java.util.Objects;
 import javax.servlet.http.HttpServletRequest;
@@ -69,7 +70,6 @@ public ClarinMatomoBitstreamTracker() {
     @Override
     protected void preTrack(Context context, MatomoRequest matomoRequest, Item item, HttpServletRequest request) {
         super.preTrack(context, matomoRequest, item, request);
-
         matomoRequest.setSiteId(siteId);
         log.debug("Logging to site " + matomoRequest.getSiteId());
         String itemIdentifier = getItemIdentifier(item);
@@ -82,6 +82,11 @@ protected void preTrack(Context context, MatomoRequest matomoRequest, Item item,
         }
         try {
             matomoRequest.setPageCustomVariable(new CustomVariable("source", "bitstream"), 1);
+            // Add the Item handle into the request as a custom dimension
+            LinkedHashMap<Long, Object> handleDimension = new LinkedHashMap<>();
+            handleDimension.put(configurationService.getLongProperty("matomo.custom.dimension.handle.id",
+                    1L), item.getHandle());
+            matomoRequest.setDimensions(handleDimension);
         } catch (MatomoException e) {
             log.error(e);
         }

diff --git a/dspace-api/src/main/java/org/dspace/app/statistics/clarin/ClarinMatomoTracker.java b/dspace-api/src/main/java/org/dspace/app/statistics/clarin/ClarinMatomoTracker.java
@@ -7,14 +7,14 @@
  */
 package org.dspace.app.statistics.clarin;
 
+import java.net.InetAddress;
+import java.net.UnknownHostException;
 import java.util.Calendar;
 import java.util.Objects;
-import java.util.concurrent.ExecutionException;
-import java.util.concurrent.Future;
+import java.util.concurrent.CompletableFuture;
 import javax.servlet.http.HttpServletRequest;
 
 import org.apache.commons.lang3.StringUtils;
-import org.apache.http.HttpResponse;
 import org.apache.logging.log4j.Logger;
 import org.dspace.content.Item;
 import org.dspace.content.factory.ClarinServiceFactory;
@@ -23,6 +23,7 @@
 import org.dspace.services.factory.DSpaceServicesFactory;
 import org.matomo.java.tracking.MatomoException;
 import org.matomo.java.tracking.MatomoRequest;
+import org.matomo.java.tracking.parameters.AcceptLanguage;
 
 /**
  * The statistics Tracker for Matomo. This class prepare and send the track GET request to the `/matomo.php`
@@ -99,13 +100,13 @@ protected MatomoRequest createMatomoRequest(HttpServletRequest request, String p
      */
     protected void preTrack(Context context, MatomoRequest matomoRequest, Item item, HttpServletRequest request) {
         if (StringUtils.isNotBlank(request.getHeader("referer"))) {
-            matomoRequest.setHeaderUserAgent(request.getHeader("referer"));
+            matomoRequest.setReferrerUrl(request.getHeader("referer"));
         }
         if (StringUtils.isNotBlank(request.getHeader("user-agent"))) {
             matomoRequest.setHeaderUserAgent(request.getHeader("user-agent"));
         }
         if (StringUtils.isNotBlank(request.getHeader("accept-language"))) {
-            matomoRequest.setHeaderUserAgent(request.getHeader("accept-language"));
+            matomoRequest.setHeaderAcceptLanguage(AcceptLanguage.fromHeader(request.getHeader("accept-language")));
         }
 
         // Creating a calendar using getInstance method
@@ -134,18 +135,13 @@ protected void preTrack(Context context, MatomoRequest matomoRequest, Item item,
      * @param matomoRequest prepared MatomoRequest for sending
      */
     public void sendTrackingRequest(MatomoRequest matomoRequest) {
-        try {
-            Future<HttpResponse> response = tracker.sendRequestAsync(matomoRequest);
-            // usually not needed:
-            HttpResponse httpResponse = response.get();
-            int statusCode = httpResponse.getStatusLine().getStatusCode();
-            if (statusCode > 399) {
-                // problem
-                log.error("Matomo tracker error the response has status code: " + statusCode);
+        CompletableFuture<MatomoRequest> completableFuture = tracker.sendRequestAsync(matomoRequest);
+
+        completableFuture.whenComplete((result, exception) -> {
+            if (exception != null) {
+                log.error("Matomo tracker error - the response exception message: {}", exception.getMessage());
             }
-        } catch (ExecutionException | InterruptedException e) {
-            e.printStackTrace();
-        }
+        });
     }
 
     protected String getFullURL(HttpServletRequest request) {
@@ -164,21 +160,35 @@ protected String getFullURL(HttpServletRequest request) {
     }
 
     /**
-     * Get IpAddress of the current user which throws this statistic event
+     * Get IpAddress of the current user which throws this statistic event. Return only the first valid IPv4 address
+     * because the Matomo tracker has a problem with IPv6 addresses.
      *
      * @param request current request
-     * @return
+     * @return only the first valid IPv4 address
      */
     protected String getIpAddress(HttpServletRequest request) {
-        String ip = "";
         String header = request.getHeader("X-Forwarded-For");
         if (header == null) {
             header = request.getRemoteAddr();
         }
         if (header != null) {
             String[] ips = header.split(", ");
-            ip = ips.length > 0 ? ips[0] : "";
+            for (String candidateIp : ips) {
+                // Validate if it's an IPv4 address
+                if (isIPv4Address(candidateIp)) {
+                    return candidateIp;
+                }
+            }
+        }
+        return null;
+    }
+
+    private boolean isIPv4Address(String ip) {
+        try {
+            InetAddress inetAddress = InetAddress.getByName(ip);
+            return inetAddress.getHostAddress().equals(ip) && inetAddress instanceof java.net.Inet4Address;
+        } catch (UnknownHostException e) {
+            return false; // Not a valid IP address
         }
-        return ip;
     }
 }
diff --git a/dspace-api/src/test/data/dspaceFolder/config/local.cfg b/dspace-api/src/test/data/dspaceFolder/config/local.cfg
@@ -312,6 +312,10 @@ webui.supported.locales = en
 # When title is something like "Type-bind test" the type-bind field will popped up
 submit.type-bind.field = dc.type,dc.identifier.citation=>dc.title
 
+# The configuration for the Matomo tracker must have a valid URL, as it will throw an exception if it does not.
+matomo.tracker.host.url = http://localhost:8135/matomo.php
+
 autocomplete.custom.separator.solr-subject_ac = \\|\\|\\|
 autocomplete.custom.separator.solr-title_ac = \\|\\|\\|
 autocomplete.custom.allowed = solr-author_ac,solr-publisher_ac,solr-dataProvider_ac,solr-dctype_ac,solr-subject_ac,solr-handle_title_ac,json_static-iso_langs.json,solr-title_ac
+
diff --git a/dspace/config/clarin-dspace.cfg b/dspace/config/clarin-dspace.cfg
@@ -155,6 +155,7 @@ matomo.site.id = 1
 matomo.tracker.bitstream.site_id = 1
 matomo.tracker.oai.site_id = 1
 matomo.tracker.host.url = http://url:port/matomo.php
+matomo.custom.dimension.handle.id = 1
 statistics.cache-server.uri = http://cache-server.none
 
 #### Statistic usage reports ####