From 1ef0fdac936f90b3f57cc6c77f197ba61eb32fd4 Mon Sep 17 00:00:00 2001
From: Jun Kim <jun.kim@fostermade.co>
Date: Tue, 26 Feb 2019 13:40:21 -0500
Subject: [PATCH 1/2] deleteByDistinctId added

---
 classes/AlgoliaEngine.inc.php | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/classes/AlgoliaEngine.inc.php b/classes/AlgoliaEngine.inc.php
index 8392afb..d8ddc81 100644
--- a/classes/AlgoliaEngine.inc.php
+++ b/classes/AlgoliaEngine.inc.php
@@ -126,4 +126,24 @@ public function clear_index()
             return $e->getMessage();
         }
     }
+
+    /**
+     * Deletes an object by distictId
+     * @param  $distinctId
+     * @return bool
+     */
+    public function deleteByDistinctId($distinctId)
+    {
+        $index = $this->client->initIndex($this->index);
+
+        try {
+            $index->deleteBy([
+                'filters' => 'distinctId:' . $distinctId,
+            ]);
+        } catch (AlgoliaException $e) {
+            return $e->getMessage();
+        }
+
+        return true;
+    }
 }
\ No newline at end of file

From 56297ecbb0bd1aaa78b4893d55aa4bc7900ae1af Mon Sep 17 00:00:00 2001
From: Jun Kim <jun.kim@fostermade.co>
Date: Tue, 26 Feb 2019 13:40:45 -0500
Subject: [PATCH 2/2] clean up, bugfixes

---
 classes/AlgoliaService.inc.php | 57 ++++++++++++++++------------------
 1 file changed, 26 insertions(+), 31 deletions(-)

diff --git a/classes/AlgoliaService.inc.php b/classes/AlgoliaService.inc.php
index 6af5f9c..af4645c 100644
--- a/classes/AlgoliaService.inc.php
+++ b/classes/AlgoliaService.inc.php
@@ -21,12 +21,16 @@
 define('ALGOLIA_INDEXINGSTATE_DIRTY', true);
 define('ALGOLIA_INDEXINGSTATE_CLEAN', false);
 
-// // The max. number of articles that can
-// // be indexed in a single batch.
+// The max. number of articles that can
+// be indexed in a single batch.
 define('ALGOLIA_INDEXING_MAX_BATCHSIZE', 2000);
 
+// Number of words to split
+define('ALGOLIA_WORDCOUNT_SPLIT', 250);
+
 import('classes.search.ArticleSearch');
 import('plugins.generic.algolia.classes.AlgoliaEngine');
+import('lib.pkp.classes.config.Config');
 
 class AlgoliaService {
     var $indexer = null;
@@ -176,7 +180,9 @@ function pushChangedArticles($batchSize = ALGOLIA_INDEXING_MAX_BATCHSIZE, $journ
             unset($toDelete);
             $this->indexer->clear_index();
         }else{
-            $this->indexer->delete($toDelete);
+            foreach($toDelete as $delete){
+                $this->indexer->deleteByDistinctId($delete['distinctId']);
+            }
         }
 
         foreach($toAdd as $add){
@@ -199,7 +205,9 @@ function deleteArticleFromIndex($articleId) {
 
         $toDelete = array();
         $toDelete[] = $this->buildAlgoliaObjectDelete($articleId);
-        $this->indexer->delete($toDelete);
+        foreach($toDelete as $delete){
+            $this->indexer->deleteByDistinctId($delete['distinctId']);
+        }
     }
 
     /**
@@ -328,13 +336,13 @@ function buildAlgoliaObjectDelete($articleOrArticleId){
         if(!is_numeric($articleOrArticleId)) {
             return array(
                 "objectAction" => "deleteObject",
-                "objectID" => $articleOrArticleId->getId(),
+                "distinctId" => $articleOrArticleId->getId(),
             );
         }
 
         return array(
             "objectAction" => "deleteObject",
-            "objectID" => $articleOrArticleId,
+            "distinctId" => $articleOrArticleId,
         );
     }
 
@@ -485,27 +493,7 @@ function getGalleyHTML($article){
         foreach($galleys as $galley){
             if($galley->getFileType() == "text/html"){
                 $submissionFile = $galley->getFile();
-                $contents = file_get_contents($submissionFile->getFilePath());
-
-                $contents = preg_replace(
-                    '/([Ss][Rr][Cc]|[Hh][Rr][Ee][Ff]|[Dd][Aa][Tt][Aa])\s*=\s*"([^"]*' . $pattern . ')"/',
-                    '\1="' . $fileUrl . '"',
-                    $contents
-                );
-
-                // Replacement for Flowplayer
-                $contents = preg_replace(
-                    '/[Uu][Rr][Ll]\s*\:\s*\'(' . $pattern . ')\'/',
-                    'url:\'' . $fileUrl . '\'',
-                    $contents
-                );
-
-                // Replacement for other players (ested with odeo; yahoo and google player won't work w/ OJS URLs, might work for others)
-                $contents = preg_replace(
-                    '/[Uu][Rr][Ll]=([^"]*' . $pattern . ')/',
-                    'url=' . $fileUrl ,
-                    $contents
-                );
+                $contents .= file_get_contents($submissionFile->getFilePath());
             }
         }
 
@@ -515,12 +503,19 @@ function getGalleyHTML($article){
     function chunkContent($content){
         $data = array();
         $updated_content = html_entity_decode($content);
-        $chunked_content = explode("</p>", wordwrap($updated_content, ALGOLIA_WORDCOUNT_SPLIT));
 
-        foreach($chunked_content as $chunked){
-            if($chunked){
-                $data[] = strip_tags($chunked);
+        if($updated_content){
+            $temp_content = str_replace("</p>", "", $updated_content);
+            $chunked_content = preg_split("/<p[^>]*?(\/?)>/i", $temp_content);
+
+            foreach($chunked_content as $chunked){
+                if($chunked){
+                    $tagless_content = strip_tags($chunked);
+                    $data[] = trim(wordwrap($tagless_content, ALGOLIA_WORDCOUNT_SPLIT));
+                }
             }
+        }else{
+            $data[] = trim(strip_tags($updated_content));
         }
 
         return $data;