From 06377b1df1f47bdefef6bc3d261466ad1ac61515 Mon Sep 17 00:00:00 2001
From: Taylor Lovett <tlovett88@gmail.com>
Date: Tue, 1 May 2018 14:42:43 +0900
Subject: [PATCH] Strip HTML before analyzing; add post_content_filtered field.
 #1070

---
 classes/class-ep-api.php  | 50 +++++++++++++++++++--------------------
 includes/mappings/5-2.php |  4 ++++
 2 files changed, 29 insertions(+), 25 deletions(-)

diff --git a/classes/class-ep-api.php b/classes/class-ep-api.php
index 633a7c025..b40e04a5d 100644
--- a/classes/class-ep-api.php
+++ b/classes/class-ep-api.php
@@ -614,31 +614,31 @@ public function prepare_post( $post_id ) {
 		remove_action( 'updated_postmeta', array( EP_Sync_Manager::factory(), 'action_queue_meta_sync' ), 10 );
 
 		$post_args = array(
-			'post_id'           => $post_id,
-			'ID'                => $post_id,
-			'post_author'       => $user_data,
-			'post_date'         => $post_date,
-			'post_date_gmt'     => $post_date_gmt,
-			'post_title'        => $this->prepare_text_content( get_the_title( $post_id ) ),
-			'post_excerpt'      => $this->prepare_text_content( $post->post_excerpt ),
-			'post_content'      => $this->prepare_text_content( apply_filters( 'the_content', $post->post_content ) ),
-			'post_status'       => $post->post_status,
-			'post_name'         => $post->post_name,
-			'post_modified'     => $post_modified,
-			'post_modified_gmt' => $post_modified_gmt,
-			'post_parent'       => $post->post_parent,
-			'post_type'         => $post->post_type,
-			'post_mime_type'    => $post->post_mime_type,
-			'permalink'         => get_permalink( $post_id ),
-			'terms'             => $this->prepare_terms( $post ),
-			'meta'              => $this->prepare_meta_types( $this->prepare_meta( $post ) ), // post_meta removed in 2.4
-			'date_terms'        => $this->prepare_date_terms( $post_date ),
-			'comment_count'     => $comment_count,
-			'comment_status'    => $comment_status,
-			'ping_status'       => $ping_status,
-			'menu_order'        => $menu_order,
-			'guid'				=> $post->guid
-			//'site_id'         => get_current_blog_id(),
+			'post_id'               => $post_id,
+			'ID'                    => $post_id,
+			'post_author'           => $user_data,
+			'post_date'             => $post_date,
+			'post_date_gmt'         => $post_date_gmt,
+			'post_title'            => $post->post_title,
+			'post_excerpt'          => $post->post_excerpt,
+			'post_content_filtered' => apply_filters( 'the_content', $post->post_content ),
+			'post_content'          => $post->post_content,
+			'post_status'           => $post->post_status,
+			'post_name'             => $post->post_name,
+			'post_modified'         => $post_modified,
+			'post_modified_gmt'     => $post_modified_gmt,
+			'post_parent'           => $post->post_parent,
+			'post_type'             => $post->post_type,
+			'post_mime_type'        => $post->post_mime_type,
+			'permalink'             => get_permalink( $post_id ),
+			'terms'                 => $this->prepare_terms( $post ),
+			'meta'                  => $this->prepare_meta_types( $this->prepare_meta( $post ) ), // post_meta removed in 2.4
+			'date_terms'            => $this->prepare_date_terms( $post_date ),
+			'comment_count'         => $comment_count,
+			'comment_status'        => $comment_status,
+			'ping_status'           => $ping_status,
+			'menu_order'            => $menu_order,
+			'guid'                  => $post->guid,
 		);
 
 		/**
diff --git a/includes/mappings/5-2.php b/includes/mappings/5-2.php
index c4dff2ce8..365c0bd0b 100644
--- a/includes/mappings/5-2.php
+++ b/includes/mappings/5-2.php
@@ -19,6 +19,7 @@
 				'default' => array(
 					'tokenizer' => 'standard',
 					'filter' => array( 'standard', 'ewp_word_delimiter', 'lowercase', 'stop', 'ewp_snowball' ),
+					'char_filter' => array( 'html_strip' ),
 					'language' => apply_filters( 'ep_analyzer_language', 'english', 'analyzer_default' ),
 				),
 				'shingle_analyzer' => array(
@@ -255,6 +256,9 @@
 				'post_content' => array(
 					'type' => 'text',
 				),
+				'post_content_filtered' => array(
+					'type' => 'text',
+				),
 				'post_status' => array(
 					'type' => 'keyword',
 				),