From f0b207d0c2d47dbb30aac5dd7335b74f40a5a404 Mon Sep 17 00:00:00 2001
From: Andrew Nicols <andrew@nicols.co.uk>
Date: Thu, 8 Oct 2015 10:36:28 +0800
Subject: [PATCH] Supply the image alt text when converting images

In order to provide an accessible version of HTML content, where an alt tag
has been provided, this should be used.

Where text refers to an image inline (for example, an emoticon), complete
removal of that image can confuse the meaning of the text. As per WCAG 2.0
guidelines (http://www.w3.org/TR/WCAG20-TECHS/H37.html), images which
convey words should be provided with alt text which inclues those words.

This also relates to images which convey meaning in some other fashion.

We have been using this modification to Html2Text since August 2010.
---
 src/Html2Text.php  |  4 +++-
 test/ImageTest.php | 42 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 45 insertions(+), 1 deletion(-)
 create mode 100644 test/ImageTest.php

diff --git a/src/Html2Text.php b/src/Html2Text.php
index 0dfb8d4..5ead1ff 100644
--- a/src/Html2Text.php
+++ b/src/Html2Text.php
@@ -67,6 +67,7 @@ class Html2Text
         '/(<tr[^>]*>|<\/tr>)/i',                          // <tr> and </tr>
         '/<td[^>]*>(.*?)<\/td>/i',                        // <td> and </td>
         '/<span class="_html2text_ignore">.+?<\/span>/i', // <span class="_html2text_ignore">...</span>
+        '/<(img)[^>]*alt=\"([^>"]+)\"[^>]*>/i',           // <img> with alt tag
     );
 
     /**
@@ -97,7 +98,8 @@ class Html2Text
         "\n\n",                          // <table> and </table>
         "\n",                            // <tr> and </tr>
         "\t\t\\1\n",                     // <td> and </td>
-        ""                               // <span class="_html2text_ignore">...</span>
+        "",                              // <span class="_html2text_ignore">...</span>
+        '[\\2]',                         // <img> with alt tag
     );
 
     /**
diff --git a/test/ImageTest.php b/test/ImageTest.php
new file mode 100644
index 0000000..04a3aa6
--- /dev/null
+++ b/test/ImageTest.php
@@ -0,0 +1,42 @@
+<?php
+
+namespace Html2Text;
+
+class ImageTest extends \PHPUnit_Framework_TestCase
+{
+    public function testImageDataProvider() {
+        return array(
+            'Without alt tag' => array(
+                'html' => '<img src="http://example.com/example.jpg">',
+                'expected'  => '',
+            ),
+            'Without alt tag, wrapped in text' => array(
+                'html' => 'xx<img src="http://example.com/example.jpg">xx',
+                'expected'  => 'xxxx',
+            ),
+            'With alt tag' => array(
+                'html' => '<img src="http://example.com/example.jpg" alt="An example image">',
+                'expected'  => '[An example image]',
+            ),
+            'With alt, and title tags' => array(
+                'html' => '<img src="http://example.com/example.jpg" alt="An example image" title="Should be ignored">',
+                'expected'  => '[An example image]',
+            ),
+            'With alt tag, wrapped in text' => array(
+                'html' => 'xx<img src="http://example.com/example.jpg" alt="An example image">xx',
+                'expected'  => 'xx[An example image]xx',
+            ),
+        );
+    }
+
+    /**
+     * @dataProvider testImageDataProvider
+     */
+    public function testImages($html, $expected)
+    {
+        $html2text = new Html2Text($html);
+        $output = $html2text->getText();
+
+        $this->assertEquals($expected, $output);
+    }
+}