Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Formatting input and output completely refactored.
- Loading branch information
Showing
5 changed files
with
482 additions
and
44 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,275 @@ | ||
| <?php | ||
| /** | ||
| * | ||
| * Telegram Bridge. An extension for the phpBB Forum Software package. | ||
| * | ||
| * @copyright (c) 2022, Edgar Bolender | ||
| * @license GNU General Public License, version 2 (GPL-2.0) | ||
| * | ||
| */ | ||
|
|
||
| namespace eb\telegram\core; | ||
|
|
||
| /** Create the text and buttons response depending on the state and the users | ||
| * input. | ||
| * If an action like saving a post or saving the state is involved, this is | ||
| * also done here.*/ | ||
| class formatters | ||
| { | ||
|
|
||
| /* | ||
| public function __construct(\phpbb\config\config $config, | ||
| \phpbb\language\language $language, | ||
| \eb\telegram\core\forum_api $forum_api | ||
| ) | ||
| { | ||
| $this->config = $config; | ||
| $this->language = $language; | ||
| } | ||
| */ | ||
|
|
||
| /** Format a phpbb-post, such that the formatting information | ||
| * is transformed into valid telegram formatting. | ||
| */ | ||
| public function format_post_for_telegram($text) { | ||
| $ent = ENT_SUBSTITUTE | ENT_HTML401; //Don't substitute quotes | ||
| //The telegram bot only allows a predefined set of HTLM-Tags. | ||
| //The forum posts however surround each opening BBCode with an <s>-Tag (and each closing with an <e>) | ||
|
|
||
| //At first we remove all these tags and their BBCodes, we do not want to show | ||
| //up in telegram. | ||
| $bbcode_pattern = '~<s>\[((i|b|u|url|email)(?:(?:=|\s).*)?)\]</s>(.*?)<e>\[/\\2]</e>~'; | ||
| do | ||
| { | ||
| $text = preg_replace($bbcode_pattern, '$3', $text, 1, $count); | ||
| } while ($count > 0); | ||
|
|
||
| $allowed_tags = ['url', 'img', 'b', 'strong', 'i', 'em', 'u', 'ins', 'strike', 'del', 'a', 'code', 'pre']; | ||
|
|
||
| //Now remove all tags, which we do not allow. (Exception list contains all allowed tags.) | ||
| //This would not be necessary for "normal" text. But when the forum sends notifications, | ||
| //a lot of additional tags are surrounding the text, BB-Codes, links etc. | ||
| //$text = strip_tags($text, $allowed_tags); //Needs php version >= 7.4 | ||
| $text = strip_tags($text, '<' . implode('><',$allowed_tags) . '>'); | ||
|
|
||
| //Special handling of the url tag: Replace <URL url=...> by <a href=...> | ||
| $url_pattern = '~<URL url="([^"]*)">(.*?)</URL>~is'; | ||
| $text = preg_replace($url_pattern, '<a href="$1">$2</a>', $text); | ||
|
|
||
| //Special handling of the img tag: | ||
| //Replace the content of the anchor tag with a <IMAGE>-Placeholder | ||
| $img_pattern = '~<IMG src="([^"]*)">\[img](.*?)\[/img]</IMG>~is'; | ||
| $text = preg_replace($img_pattern, '<a href="$1"><<IMAGE>></a>', $text); | ||
|
|
||
| //Add server address to relative links (no http-protokoll) starting with a slash (/) | ||
| $rel_url_pattern = '~<a href="((?!https?://)/[^"]*)">(.*?)</a>~is'; | ||
| $replace = '<a href="' . generate_board_url(true) . '$1">$2</a>'; | ||
| $text = preg_replace($rel_url_pattern, $replace, $text); | ||
|
|
||
| //Add server address to relative links (no http-protokoll) NOT starting with a slash (/) | ||
| $rel_url_pattern = '~<a href="((?!https?://)[^"]*)">(.*?)</a>~is'; | ||
| $replace = '<a href="' . generate_board_url(false) . '/$1">$2</a>'; | ||
| $text = preg_replace($rel_url_pattern, $replace, $text); | ||
|
|
||
| //Add a non printable space (ZWSP) to all forward slashes. | ||
| //By that, telegram does not treat the forward slash as the beginning of a command. | ||
| //Exclude double // and slashes belonging to html-tags | ||
| $text = preg_replace('~([^<]/)([^/])~', "$1\u{200B}$2", $text); | ||
|
|
||
| //Revert this, for all hrefs in anchors. | ||
| //By that, telegram does not treat the forward slash as the beginning of a command. | ||
| do | ||
| { | ||
| $text = preg_replace("~(<a href=\"[^\"]*/)\u{200B}([^\"]*\">)~", '$1$2', $text, 1, $count); | ||
| } while ($count > 0); | ||
| return $text; | ||
| } | ||
|
|
||
| /** Implements the substring function (without length-param) | ||
| * such that html-tags are still correctly opened and closed. | ||
| * Cut the beginning of a text at the given offset. | ||
| * If the offset happens to lay inside a tagged area, | ||
| * the possibly cut off start tags are added again to the beginning of the | ||
| * text, such that the tags are still opened and closed correctly. | ||
| * If the offset would cut a start tag (before the ending >) or | ||
| * an end tag into pieces, the $offset is moved behind the closing >. | ||
| * By adding the start-tags before the offset, the text-length is increased. | ||
| * To ensure, the text is not longer, than it would be expected by the offset, | ||
| * the offset is increased, until the total lenght is less or equal | ||
| * than mb_strlen($text) - $offset. | ||
| */ | ||
| public function tag_aware_substr($tagged_text, $offset) { | ||
| $tags = $this->parse_tags($tagged_text); | ||
| $prefix = ''; | ||
| $cut_point = $offset - 1; | ||
| do | ||
| { | ||
| $cut_point++; | ||
| $prefix = $this->adapt_cut_point($tags, $cut_point); | ||
|
|
||
| } while ($cut_point - mb_strlen($prefix) < $offset); | ||
| return $prefix . mb_substr($tagged_text, $cut_point); | ||
| } | ||
|
|
||
| /** For a text, which is to be cut, adapt the cut-point, such | ||
| * that it does not cut a start or end tag of the text, and | ||
| * return the start-tags, which must be added as prefix, because the | ||
| * cut text still contains the corresponding end tags. | ||
| */ | ||
| public function adapt_cut_point($tags, &$offset) { | ||
| $print = array(); | ||
| //Move the offset, such that no tag is split | ||
| foreach ($tags as $tag) { | ||
| if ($offset <= $tag['full_s'] || $offset >= $tag['full_e']) | ||
| { | ||
| continue; //We are outside the enclosing tags. | ||
| } | ||
| if ($offset < $tag['full_s'] + mb_strlen($tag['s_tag'])) | ||
| { | ||
| //Start tag would be cut. Move offset behind start-tag | ||
| $start_tag = $tag['s_tag']; | ||
| $offset = $tag['full_s'] + mb_strlen($start_tag); | ||
| break; //Tags are sequential. Cannot happen again. | ||
| } | ||
| if ($offset >= $tag['full_e'] - mb_strlen($tag['e_tag'])) | ||
| { | ||
| //End tag would be cut, move $offset behind the end-tag. | ||
| //A pure endtag without previous content also does not make sense, | ||
| //therefore, this is also skipped. (>= in the condition above) | ||
| $offset = $tag['full_e']; | ||
| break; //Tags are sequential. Cannot happen again. | ||
| } | ||
| } | ||
| //Collect the start tags, which must be added to the beginning, such that | ||
| //their is no end-tag with missing start tag | ||
| $start_tags = array(); | ||
| foreach ($tags as $tag) { | ||
| if ($offset > $tag['full_s'] && $offset < $tag['full_e']) | ||
| { | ||
| $start_tag = $tag['s_tag']; | ||
| $pos = $tag['full_s']; //keep start order | ||
| $start_tags[$pos] = $start_tag; | ||
| } | ||
| } | ||
| ksort($start_tags); | ||
| return implode('', $start_tags); | ||
| } | ||
|
|
||
| /** Find the start and end of text enclosed in html tags. | ||
| * The result is an array containing all tagged texts (alos if nested) | ||
| * in the following form: | ||
| * array( | ||
| * array('full' => full tag enclosed text, | ||
| * 'full_s' => offset, where full text starts | ||
| * 'full_e' => offset, where full text ends | ||
| * 's_tag' => complete start tag (including attributes) | ||
| * 'e_tag' => complete end tag | ||
| * )) | ||
| * In case of self closing tags, e_tag is empty. | ||
| */ | ||
| public function parse_tags($tagged_text) { | ||
| $tag_pattern = "/<([\w]+)([^>]*?)(?:([\s]*\/>)|(?:(>)(?:(?:(?:[^<]*?|<\!\-\-.*?\-\->)|(?R))*)(<\/\\1[\s]*>)))/xsmu"; | ||
| $result[] = array('full' => ' ' . $tagged_text, 'full_s' => -1); | ||
| for($i = 0; $i < count($result); $i++) { | ||
| $full = $result[$i]['full']; | ||
| $offset = $result[$i]['full_s']; | ||
| $t_count = preg_match_all($tag_pattern, mb_substr($full,1), $matches, PREG_OFFSET_CAPTURE); | ||
| if ($t_count) { | ||
| $j = 0; | ||
| foreach($matches[0] as $match) { | ||
| $mb_offset = mb_strlen(substr(mb_substr($full,1), 0, $match[1])); | ||
| $result[] = array( | ||
| 'full' => $match[0], | ||
| 'full_s' => $mb_offset + 1 + $offset, | ||
| 'full_e' => $mb_offset + 1 + mb_strlen($match[0]) + $offset, | ||
| 's_tag' => '<' . $matches[1][$j][0] . $matches[2][$j][0] . $matches[3][$j][0] . $matches[4][$j][0], | ||
| 'e_tag' => $matches[5][$j][0], | ||
| ); | ||
| $j++; | ||
| } | ||
| } | ||
| } | ||
| //Remove the first full text entry | ||
| return array_slice($result,1); | ||
| } | ||
|
|
||
| /** Format the telegram input by adding bbCodes according to the formatting information, | ||
| * which telegram sends as so called entities. | ||
| */ | ||
| public function format_input($text, $entities) | ||
| { | ||
| /* Split the text, at every point where a formatting starts or ends into an array. | ||
| * Therefore we collect at first all splitpoints, and remove duplicates. | ||
| */ | ||
| $split_points[] = 0; | ||
| foreach ($entities as $entity) | ||
| { | ||
| $split_points[] = $entity->offset; | ||
| $split_points[] = $entity->offset + $entity->length; | ||
| } | ||
| $split_points = array_unique($split_points); | ||
| rsort($split_points); | ||
| $chunks = array(); | ||
| foreach ($split_points as $point) | ||
| { | ||
| $chunks[$point] = mb_substr($text, $point); | ||
| $text = mb_substr($text, 0, $point); | ||
| } | ||
| ksort($chunks); | ||
| //Sort by end of formatting, such that in case of overlapping formats, the opening tag | ||
| //for the format, that gets closed last is placed at first. | ||
| usort($entities, function($a, $b) | ||
| { | ||
| return (($a->offset + $a->length) < ($b->offset + $b->length)) ? -1 : 1; | ||
| }); | ||
| foreach ($entities as $entity) | ||
| { | ||
| $bbcode = $this->get_bbcode($entity->type); | ||
| if (!$bbcode) | ||
| { | ||
| continue; | ||
| } | ||
| $chunks[$entity->offset] = $bbcode . $chunks[$entity->offset]; | ||
| } | ||
| for ($i = count($entities) - 1; $i >= 0; $i--) | ||
| { | ||
| $entity = $entities[$i]; | ||
| $bbcode = $this->get_bbcode($entity->type, false); | ||
| if (!$bbcode) | ||
| { | ||
| continue; | ||
| } | ||
| $bbcode_start = $this->get_bbcode($entity->type); | ||
| $end = $entity->offset + $entity->length; | ||
| if (strpos($chunks[$end], $bbcode_start) === 0) | ||
| { | ||
| //Remove ending tag immediatly followed by starting tag | ||
| $chunks[$end] = substr($chunks[$end], strlen($bbcode_start)); | ||
| } else | ||
| { | ||
| $chunks[$end] = $bbcode . $chunks[$end]; | ||
| } | ||
| } | ||
| //Remove non printable whitespace, which may have been included, when user copies | ||
| //a part of a post, where the whitespace was added. (See telegrami_api->htmlentitiesForTelegram) | ||
| $text = implode('', $chunks); | ||
| $text = str_replace("/\u{200B}", "/", $text); | ||
| return $text; | ||
| } | ||
|
|
||
| private function get_bbcode($format_type, $start = true) | ||
| { | ||
| switch ($format_type) | ||
| { | ||
| case 'bold': return $start ? '[b]' : '[/b]'; | ||
| case 'italic': return $start ? '[i]' : '[/i]'; | ||
| case 'underline': return $start ? '[u]' : '[/u]'; | ||
| case 'code': return $start ? '[code]' : '[/code]'; | ||
| case 'pre': return $start ? '[code]' : '[/code]'; | ||
| case 'strikethrough': return $start ? '<del>' : '</del>'; | ||
| case 'url': return ''; //No need for BBCode | ||
| default: return false; | ||
| } | ||
| } | ||
|
|
||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.