Add html:eml conversion

Author: stackpr

README.md

@@ -12,25 +12,26 @@ This table shows the number of converters configured by default between file ext
 
     fileconverter list extension-table
 
-source | dbk | docx | epub | epub3 | fb2 | html | jpg | man | md | odt | opml | pdf | ps | rtf | txt
---- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | ---
-bib |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |
-dbk | 1 | 1 | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  | 1 | 1
-doc |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  | 1
-html | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 9 |  | 1 | 1
-jpg |  |  |  |  |  |  | 1 |  |  |  |  |  |  |  |
-ltx |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |
-md | 1 | 1 | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  | 1 | 1
-opml | 1 | 1 | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  | 1 | 1
-pdb |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |
-pdf |  |  |  |  |  |  | 1 |  |  |  |  | 1 |  |  |
-ps |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |
-psw |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |
-rst | 1 | 1 | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  | 1 | 1
-rtf |  |  |  |  |  |  |  |  |  |  |  | 4 | 2 |  |
-sdw |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |
-sxw |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |
-tex | 1 | 1 | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  | 1 | 1
-textile | 1 | 1 | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  | 1 | 1
-txt | 1 | 1 | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 2 |  | 1 | 1
-vor |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |
+source | dbk | docx | eml | epub | epub3 | fb2 | html | jpg | man | md | odt | opml | pdf | pdf/grayscale | ps | rtf | txt
+--- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | ---
+bib |  |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |  | 
+dbk | 1 | 1 |  | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  |  | 1 | 1
+doc |  |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |  | 1
+html | 1 | 1 | 1 | 1 | 1 | 1 | 3 | 1 | 1 | 1 | 1 | 1 | 9 |  |  | 1 | 1
+jpg |  |  |  |  |  |  |  | 1 |  |  |  |  |  |  |  |  | 
+ltx |  |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |  | 
+md | 1 | 1 |  | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  |  | 1 | 1
+opml | 1 | 1 |  | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  |  | 1 | 1
+pdb |  |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |  | 
+pdf |  |  |  |  |  |  |  | 1 |  |  |  |  | 1 | 1 |  |  | 
+ps |  |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |  | 
+psw |  |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |  | 
+rst | 1 | 1 |  | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  |  | 1 | 1
+rtf |  |  |  |  |  |  |  |  |  |  |  |  | 4 |  | 2 |  | 
+sdw |  |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |  | 
+sxw |  |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |  | 
+tex | 1 | 1 |  | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  |  | 1 | 1
+textile | 1 | 1 |  | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 1 |  |  | 1 | 1
+txt | 1 | 1 |  | 1 | 1 | 1 | 1 |  | 1 | 1 | 1 | 1 | 2 |  |  | 1 | 1
+vor |  |  |  |  |  |  |  |  |  |  |  |  | 1 |  |  |  | 
+wiki | 1 | 1 |  | 1 | 1 | 1 | 2 |  | 1 | 2 | 1 | 1 | 1 |  |  | 1 | 1

composer.json

@@ -20,7 +20,8 @@
         "php":">=5.4.5",
         "symfony/console":"2.4.*",
         "wittiws/htmlawed":"dev-master",
-        "wittiws/phpquery":"dev-master"
+        "wittiws/phpquery":"dev-master",
+        "zendframework/zend-mail":"2.7.*"
     },
     "autoload":{
         "psr-4":{

src/Configuration/ConfigurationDefaults.php

@@ -61,6 +61,11 @@ public function __construct(&$settings) {
           '#engine' => 'Convert\\Catdoc',
         ),
       ),
+      'html->eml' => array(
+        'nativearchive:default' => array(
+          '#engine' => 'Convert\\NativeArchive',
+        ),
+      ),
       'html->pdf' => array(
         'htmldoc:default' => array(
           '#engine' => 'Convert\\Htmldoc',

src/Engine/Convert/NativeArchive.php

@@ -0,0 +1,92 @@
+<?php
+/*
+ * This file is part of the FileConverter package.
+ *
+ * (c) Greg Payne
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace FileConverter\Engine\Convert;
+use FileConverter\Engine\EngineBase;
+use Zend\Mail\Message;
+use Zend\Mime\Message as MimeMessage;
+use Zend\Mime\Mime;
+use Zend\Mime\Part as MimePart;
+
+class NativeArchive extends EngineBase {
+  /**
+   * @todo use Message::fromString($raw) to convert from eml to other formats
+   */
+  public function convertFile($source, $destination) {
+    if ($this->conversion[0] === 'html' && $this->conversion[1] === 'eml') {
+      $parts = array(
+        'html' => NULL,
+      );
+
+      // Load the HTML body.
+      $html = file_get_contents($source);
+      $html = preg_replace_callback('@(?<a>\s+src=")(?<src>[^"]+)(?<b>")@si', function ($matches) use (&$parts) {
+        // Mangle the src attribute.
+        $src = $matches['src'];
+        if (strpos($src, '://') === FALSE && !preg_match('@^/|\.\.@s', $src)) {
+          if (is_file($src)) {
+            $cid = 'imageid' . sizeof($parts);
+            $ext = strtolower(pathinfo($src, PATHINFO_EXTENSION));
+            if ($ext === 'jpg') {
+              $ext = 'jpeg';
+            }
+            $parts[$cid] = new MimePart(fopen($src, 'r'));
+            $parts[$cid]->type = "image/$ext; name=$cid.$ext";
+            $parts[$cid]->encoding = Mime::ENCODING_BASE64;
+            $parts[$cid]->id = $cid;
+            $src = "cid:$cid";
+          }
+        }
+
+        // Rebuild the src attribute.
+        $ret = $matches['a'] . $src . $matches['b'];
+        return $ret;
+      }, $html);
+
+      // Build the message.
+      $parts['html'] = new MimePart($html);
+      $parts['html']->type = Mime::TYPE_HTML;
+      $parts['html']->charset = 'utf-8';
+      $parts['html']->encoding = Mime::ENCODING_QUOTEDPRINTABLE;
+      $body = new MimeMessage();
+      $body->setParts(array_values($parts));
+
+      // Build the full output.
+      $message = new Message();
+      $message->setBody($body);
+      $output = $message->toString();
+      file_put_contents($destination, $output);
+      return $this;
+    }
+
+    throw new \InvalidArgumentException("Unsupported conversion type requested");
+  }
+
+  protected function getHelpInstallation($os, $os_version) {
+    $help = array(
+      'title' => 'Native Archives',
+    );
+    switch ($os) {
+      case 'Ubuntu':
+        $help['os'] = 'confirmed on Ubuntu 16.04';
+        $help['notes'] = array(
+          'composer update',
+        );
+        return $help;
+    }
+
+    return parent::getHelpInstallation($os, $os_version);
+  }
+
+  public function isAvailable() {
+    return (class_exists('\Zend\Mail\Message'));
+  }
+
+}