From 195bc894c1a3a892ff1d006301d75fcae0e04e3c Mon Sep 17 00:00:00 2001
From: seb303 <seb@burnit.co.uk>
Date: Tue, 20 Feb 2024 10:46:13 +0000
Subject: [PATCH 1/5] Added option 'plain_text'

If true then disables various pseudo formatting:
No converting bold, th or headings to upper case.
No character conversion to simulate strike through for <del>
No adding _ around italic text (<i> <em> and <ins> tags)
<hr> is replaced by "\n\n" rather than "\n-------------------------\n"
---
 src/Html2Text.php | 50 ++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 49 insertions(+), 1 deletion(-)
diff --git a/src/Html2Text.php b/src/Html2Text.php
index 16a6572..7297013 100644
--- a/src/Html2Text.php
+++ b/src/Html2Text.php
@@ -102,6 +102,38 @@ class Html2Text
         '[\\2]',                         // <img> with alt tag
     );
 
+    /**
+     * List of pattern replacements corresponding to patterns searched.
+     * Used with 'plain_text' option.
+     *
+     * @var array $replace
+     * @see $search
+     */
+    protected $replace_plain = array(
+        '',                              // Non-legal carriage return
+        ' ',                             // Newlines and tabs
+        '',                              // <head>
+        '',                              // <script>s -- which strip_tags supposedly has problems with
+        '',                              // <style>s -- which strip_tags supposedly has problems with
+        '\\1',                           // <i>
+        '\\1',                           // <em>
+        '\\1',                           // <ins>
+        "\n\n",                          // <ul> and </ul>
+        "\n\n",                          // <ol> and </ol>
+        "\n\n",                          // <dl> and </dl>
+        "\t* \\1\n",                     // <li> and </li>
+        " \\1\n",                        // <dd> and </dd>
+        "\t* \\1",                       // <dt> and </dt>
+        "\n\t* ",                        // <li>
+        "\n\n",                          // <hr>
+        "<div>\n",                       // <div>
+        "\n\n",                          // <table> and </table>
+        "\n",                            // <tr> and </tr>
+        "\t\t\\1\n",                     // <td> and </td>
+        "",                              // <span class="_html2text_ignore">...</span>
+        '[\\2]',                         // <img> with alt tag
+    );
+
     /**
      * List of preg* regular expression patterns to search for,
      * used in conjunction with $entReplace.
@@ -222,6 +254,12 @@ class Html2Text
         'width' => 70,          //  Maximum width of the formatted text, in columns.
                                 //  Set this value to 0 (or less) to ignore word wrapping
                                 //  and not constrain text to a fixed-width column.
+        
+        'plain_text' => false,  //  If true then disables various pseudo formatting:
+                                //  No converting bold, th or headings to upper case.
+                                //  No character conversion to simulate strike through for <del>
+                                //  No adding _ around italic text (<i> <em> and <ins> tags)
+                                //  <hr> is replaced by "\n\n" rather than "\n-------------------------\n"
     );
 
     private function legacyConstruct($html = '', $fromFile = false, array $options = array())
@@ -371,7 +409,11 @@ protected function converter(&$text)
     {
         $this->convertBlockquotes($text);
         $this->convertPre($text);
-        $text = preg_replace($this->search, $this->replace, $text);
+        if ($this->options['plain_text']) {
+            $text = preg_replace($this->search, $this->replace_plain, $text);
+        } else {
+            $text = preg_replace($this->search, $this->replace, $text);
+        }
         $text = preg_replace_callback($this->callbackSearch, array($this, 'pregCallback'), $text);
         $text = strip_tags($text);
         $text = preg_replace($this->entSearch, $this->entReplace, $text);
@@ -611,6 +653,9 @@ protected function pregPreCallback(/** @noinspection PhpUnusedParameterInspectio
      */
     protected function toupper($str)
     {
+        if ($this->options['plain_text']) {
+            return $str;
+        }
         // string can contain HTML tags
         $chunks = preg_split('/(<[^>]*>)/', $str, -1, PREG_SPLIT_NO_EMPTY | PREG_SPLIT_DELIM_CAPTURE);
 
@@ -647,6 +692,9 @@ protected function strtoupper($str)
      */
     protected function tostrike($str)
     {
+        if ($this->options['plain_text']) {
+            return $str;
+        }
         $rtn = '';
         for ($i = 0; $i < mb_strlen($str); $i++) {
             $chr = mb_substr($str, $i, 1);

From dcfe82a383c36057690b0b69635d9d785e337a6e Mon Sep 17 00:00:00 2001
From: seb303 <seb@burnit.co.uk>
Date: Tue, 20 Feb 2024 10:49:32 +0000
Subject: [PATCH 2/5] Added documentation for options parameter

---
 README.md | 32 ++++++++++++++++++++++++++++++--
 1 file changed, 30 insertions(+), 2 deletions(-)

diff --git a/README.md b/README.md
index e66da7b..80ba8c2 100644
--- a/README.md
+++ b/README.md
@@ -12,9 +12,37 @@ composer require html2text/html2text
 
 ## Basic Usage
 ```php
-$html = new \Html2Text\Html2Text('Hello, &quot;<b>world</b>&quot;');
+$html2text = new \Html2Text\Html2Text('Hello, &quot;<b>world</b>&quot;');
 
-echo $html->getText();  // Hello, "WORLD"
+echo $html2text->getText();  // Hello, "WORLD"
+```
+
+## Options
+
+The constructor takes an optional 2nd parameter which is an array of options.
+```php
+$html2text = new \Html2Text\Html2Text(string $html, array $options = []);
+```
+
+Options and defaults are:
+```php
+[
+        'do_links' => 'inline', // 'none'
+                                // 'inline' (show links inline)
+                                // 'nextline' (show links on the next line)
+                                // 'table' (if a table of link URLs should be listed after the text.
+                                // 'bbcode' (show links as bbcode)
+
+        'width' => 70,          //  Maximum width of the formatted text, in columns.
+                                //  Set this value to 0 (or less) to ignore word wrapping
+                                //  and not constrain text to a fixed-width column.
+        
+        'plain_text' => false,  //  If true then disables various pseudo formatting:
+                                //  No converting bold, th or headings to upper case.
+                                //  No character conversion to simulate strike through for <del>
+                                //  No adding _ around italic text (<i> <em> and <ins> tags)
+                                //  <hr> is replaced by "\n\n" rather than "\n-------------------------\n"
+];
 ```
 
 ## History

From 221614011cafa62ddfafc553d7f8727c6fca0b5b Mon Sep 17 00:00:00 2001
From: seb303 <seb@burnit.co.uk>
Date: Tue, 20 Feb 2024 11:47:20 +0000
Subject: [PATCH 3/5] Improved collapsing of excess whitespace

Now UTF-8 non-breaking space is supported when normalising empty lines.
Also trims any trailing whitespace.
---
 src/Html2Text.php | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/src/Html2Text.php b/src/Html2Text.php
index 7297013..04a775b 100644
--- a/src/Html2Text.php
+++ b/src/Html2Text.php
@@ -427,12 +427,15 @@ protected function converter(&$text)
         $text = str_replace('|+|amp|+|', '&', $text);
 
         // Normalise empty lines
-        $text = preg_replace("/\n\s+\n/", "\n\n", $text);
-        $text = preg_replace("/[\n]{3,}/", "\n\n", $text);
+        $text = preg_replace("/\n[(\xC2\xA0|\s)]+\n/", "\n\n", $text);
+        $text = preg_replace("/\n{3,}/", "\n\n", $text);
 
         // remove leading empty lines (can be produced by eg. P tag on the beginning)
         $text = ltrim($text, "\n");
 
+        // remove trailing white space
+        $text = preg_replace("/[(\xC2\xA0|\s)]+$/", "", $text);
+
         if ($this->options['width'] > 0) {
             $text = wordwrap($text, $this->options['width']);
         }

From 8d85323c6a30c013cce32faa9a6c2ebadb26b6f1 Mon Sep 17 00:00:00 2001
From: seb303 <seb@burnit.co.uk>
Date: Wed, 21 Feb 2024 00:40:50 +0000
Subject: [PATCH 4/5] Cleaned up regexp

No change to functionality
---
 src/Html2Text.php | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/Html2Text.php b/src/Html2Text.php
index 04a775b..7b40633 100644
--- a/src/Html2Text.php
+++ b/src/Html2Text.php
@@ -427,7 +427,7 @@ protected function converter(&$text)
         $text = str_replace('|+|amp|+|', '&', $text);
 
         // Normalise empty lines
-        $text = preg_replace("/\n[(\xC2\xA0|\s)]+\n/", "\n\n", $text);
+        $text = preg_replace("/\n(\xC2\xA0|\s)+\n/", "\n\n", $text);
         $text = preg_replace("/\n{3,}/", "\n\n", $text);
 
         // remove leading empty lines (can be produced by eg. P tag on the beginning)

From 12d07f53d746bb383fb2556f5b8b64472c5d53c3 Mon Sep 17 00:00:00 2001
From: seb303 <seb@burnit.co.uk>
Date: Tue, 26 Mar 2024 10:19:31 +0000
Subject: [PATCH 5/5] plain_text option now also suppresses * for <li> or <dt>

---
 src/Html2Text.php | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/src/Html2Text.php b/src/Html2Text.php
index 7b40633..bfd9919 100644
--- a/src/Html2Text.php
+++ b/src/Html2Text.php
@@ -121,9 +121,9 @@ class Html2Text
         "\n\n",                          // <ul> and </ul>
         "\n\n",                          // <ol> and </ol>
         "\n\n",                          // <dl> and </dl>
-        "\t* \\1\n",                     // <li> and </li>
+        "\t\\1\n",                       // <li> and </li>
         " \\1\n",                        // <dd> and </dd>
-        "\t* \\1",                       // <dt> and </dt>
+        "\t\\1",                         // <dt> and </dt>
         "\n\t* ",                        // <li>
         "\n\n",                          // <hr>
         "<div>\n",                       // <div>
@@ -259,6 +259,7 @@ class Html2Text
                                 //  No converting bold, th or headings to upper case.
                                 //  No character conversion to simulate strike through for <del>
                                 //  No adding _ around italic text (<i> <em> and <ins> tags)
+                                //  No adding * for <li> or <dt>
                                 //  <hr> is replaced by "\n\n" rather than "\n-------------------------\n"
     );
 
