| 1 |
<?php |
| 2 |
/** |
| 3 |
* @copyright Copyright (c) 2014 Carsten Brandt |
| 4 |
* @license https://github.com/cebe/markdown/blob/master/LICENSE |
| 5 |
* @link https://github.com/cebe/markdown#readme |
| 6 |
*/ |
| 7 |
|
| 8 |
namespace cebe\markdownparser; |
| 9 |
use ReflectionMethod; |
| 10 |
|
| 11 |
/** |
| 12 |
* A generic parser for markdown-like languages. |
| 13 |
* |
| 14 |
* @author Carsten Brandt <mail@cebe.cc> |
| 15 |
*/ |
| 16 |
abstract class Parser |
| 17 |
{ |
| 18 |
/** |
| 19 |
* @var integer the maximum nesting level for language elements. |
| 20 |
*/ |
| 21 |
public $maximumNestingLevel = 32; |
| 22 |
|
| 23 |
/** |
| 24 |
* @var array the current context the parser is in. |
| 25 |
* TODO remove in favor of absy |
| 26 |
*/ |
| 27 |
protected $context = []; |
| 28 |
/** |
| 29 |
* @var array these are "escapeable" characters. When using one of these prefixed with a |
| 30 |
* backslash, the character will be outputted without the backslash and is not interpreted |
| 31 |
* as markdown. |
| 32 |
*/ |
| 33 |
protected $escapeCharacters = [ |
| 34 |
'\\', // backslash |
| 35 |
]; |
| 36 |
|
| 37 |
private $_depth = 0; |
| 38 |
|
| 39 |
|
| 40 |
/** |
| 41 |
* Parses the given text considering the full language. |
| 42 |
* |
| 43 |
* This includes parsing block elements as well as inline elements. |
| 44 |
* |
| 45 |
* @param string $text the text to parse |
| 46 |
* @return string parsed markup |
| 47 |
*/ |
| 48 |
public function parse($text) |
| 49 |
{ |
| 50 |
$this->prepare(); |
| 51 |
|
| 52 |
if (ltrim($text) === '') { |
| 53 |
return ''; |
| 54 |
} |
| 55 |
|
| 56 |
$text = str_replace(["\r\n", "\n\r", "\r"], "\n", $text); |
| 57 |
|
| 58 |
$this->prepareMarkers($text); |
| 59 |
|
| 60 |
$absy = $this->parseBlocks(explode("\n", $text)); |
| 61 |
$markup = $this->renderAbsy($absy); |
| 62 |
|
| 63 |
$this->cleanup(); |
| 64 |
return $markup; |
| 65 |
} |
| 66 |
|
| 67 |
/** |
| 68 |
* Parses a paragraph without block elements (block elements are ignored). |
| 69 |
* |
| 70 |
* @param string $text the text to parse |
| 71 |
* @return string parsed markup |
| 72 |
*/ |
| 73 |
public function parseParagraph($text) |
| 74 |
{ |
| 75 |
$this->prepare(); |
| 76 |
|
| 77 |
if (ltrim($text) === '') { |
| 78 |
return ''; |
| 79 |
} |
| 80 |
|
| 81 |
$text = str_replace(["\r\n", "\n\r", "\r"], "\n", $text); |
| 82 |
|
| 83 |
$this->prepareMarkers($text); |
| 84 |
|
| 85 |
$absy = $this->parseInline($text); |
| 86 |
$markup = $this->renderAbsy($absy); |
| 87 |
|
| 88 |
$this->cleanup(); |
| 89 |
return $markup; |
| 90 |
} |
| 91 |
|
| 92 |
/** |
| 93 |
* This method will be called before `parse()` and `parseParagraph()`. |
| 94 |
* You can override it to do some initialization work. |
| 95 |
*/ |
| 96 |
protected function prepare() |
| 97 |
{ |
| 98 |
} |
| 99 |
|
| 100 |
/** |
| 101 |
* This method will be called after `parse()` and `parseParagraph()`. |
| 102 |
* You can override it to do cleanup. |
| 103 |
*/ |
| 104 |
protected function cleanup() |
| 105 |
{ |
| 106 |
} |
| 107 |
|
| 108 |
|
| 109 |
// block parsing |
| 110 |
|
| 111 |
private $_blockTypes; |
| 112 |
|
| 113 |
/** |
| 114 |
* @return array a list of block element types available. |
| 115 |
*/ |
| 116 |
protected function blockTypes() |
| 117 |
{ |
| 118 |
if ($this->_blockTypes === null) { |
| 119 |
// detect block types via "identify" functions |
| 120 |
$reflection = new \ReflectionClass($this); |
| 121 |
$this->_blockTypes = array_filter(array_map(function($method) { |
| 122 |
$name = $method->getName(); |
| 123 |
return strncmp($name, 'identify', 8) === 0 ? strtolower(substr($name, 8)) : false; |
| 124 |
}, $reflection->getMethods(ReflectionMethod::IS_PROTECTED))); |
| 125 |
|
| 126 |
sort($this->_blockTypes); |
| 127 |
} |
| 128 |
return $this->_blockTypes; |
| 129 |
} |
| 130 |
|
| 131 |
/** |
| 132 |
* Given a set of lines and an index of a current line it uses the registed block types to |
| 133 |
* detect the type of this line. |
| 134 |
* @param array $lines |
| 135 |
* @param integer $current |
| 136 |
* @return string name of the block type in lower case |
| 137 |
*/ |
| 138 |
protected function detectLineType($lines, $current) |
| 139 |
{ |
| 140 |
$line = $lines[$current]; |
| 141 |
$blockTypes = $this->blockTypes(); |
| 142 |
foreach($blockTypes as $blockType) { |
| 143 |
if ($this->{'identify' . $blockType}($line, $lines, $current)) { |
| 144 |
return $blockType; |
| 145 |
} |
| 146 |
} |
| 147 |
// consider the line a normal paragraph if no other block type matches |
| 148 |
return 'paragraph'; |
| 149 |
} |
| 150 |
|
| 151 |
/** |
| 152 |
* Parse block elements by calling `detectLineType()` to identify them |
| 153 |
* and call consume function afterwards. |
| 154 |
*/ |
| 155 |
protected function parseBlocks($lines) |
| 156 |
{ |
| 157 |
if ($this->_depth >= $this->maximumNestingLevel) { |
| 158 |
// maximum depth is reached, do not parse input |
| 159 |
return [['text', implode("\n", $lines)]]; |
| 160 |
} |
| 161 |
$this->_depth++; |
| 162 |
|
| 163 |
$blocks = []; |
| 164 |
|
| 165 |
// convert lines to blocks |
| 166 |
for ($i = 0, $count = count($lines); $i < $count; $i++) { |
| 167 |
$line = $lines[$i]; |
| 168 |
if ($line !== '' && rtrim($line) !== '') { // skip empty lines |
| 169 |
// identify a blocks beginning and parse the content |
| 170 |
list($block, $i) = $this->parseBlock($lines, $i); |
| 171 |
if ($block !== false) { |
| 172 |
$blocks[] = $block; |
| 173 |
} |
| 174 |
} |
| 175 |
} |
| 176 |
|
| 177 |
$this->_depth--; |
| 178 |
|
| 179 |
return $blocks; |
| 180 |
} |
| 181 |
|
| 182 |
/** |
| 183 |
* Parses the block at current line by identifying the block type and parsing the content |
| 184 |
* @param $lines |
| 185 |
* @param $current |
| 186 |
* @return array Array of two elements, the first element contains the block, |
| 187 |
* the second contains the next line index to be parsed. |
| 188 |
*/ |
| 189 |
protected function parseBlock($lines, $current) |
| 190 |
{ |
| 191 |
// identify block type for this line |
| 192 |
$blockType = $this->detectLineType($lines, $current); |
| 193 |
|
| 194 |
// call consume method for the detected block type to consume further lines |
| 195 |
return $this->{'consume' . $blockType}($lines, $current); |
| 196 |
} |
| 197 |
|
| 198 |
protected function renderAbsy($blocks) |
| 199 |
{ |
| 200 |
$output = ''; |
| 201 |
foreach ($blocks as $block) { |
| 202 |
array_unshift($this->context, $block[0]); |
| 203 |
$output .= $this->{'render' . $block[0]}($block); |
| 204 |
array_shift($this->context); |
| 205 |
} |
| 206 |
return $output; |
| 207 |
} |
| 208 |
|
| 209 |
/** |
| 210 |
* Consume lines for a paragraph |
| 211 |
* |
| 212 |
* @param $lines |
| 213 |
* @param $current |
| 214 |
* @return array |
| 215 |
*/ |
| 216 |
protected function consumeParagraph($lines, $current) |
| 217 |
{ |
| 218 |
// consume until newline |
| 219 |
$content = []; |
| 220 |
for ($i = $current, $count = count($lines); $i < $count; $i++) { |
| 221 |
if (ltrim($lines[$i]) !== '') { |
| 222 |
$content[] = $lines[$i]; |
| 223 |
} else { |
| 224 |
break; |
| 225 |
} |
| 226 |
} |
| 227 |
$block = [ |
| 228 |
'paragraph', |
| 229 |
'content' => $this->parseInline(implode("\n", $content)), |
| 230 |
]; |
| 231 |
return [$block, --$i]; |
| 232 |
} |
| 233 |
|
| 234 |
/** |
| 235 |
* Render a paragraph block |
| 236 |
* |
| 237 |
* @param $block |
| 238 |
* @return string |
| 239 |
*/ |
| 240 |
protected function renderParagraph($block) |
| 241 |
{ |
| 242 |
return '<p>' . $this->renderAbsy($block['content']) . "</p>\n"; |
| 243 |
} |
| 244 |
|
| 245 |
|
| 246 |
// inline parsing |
| 247 |
|
| 248 |
|
| 249 |
/** |
| 250 |
* @var array the set of inline markers to use in different contexts. |
| 251 |
*/ |
| 252 |
private $_inlineMarkers = []; |
| 253 |
|
| 254 |
/** |
| 255 |
* Returns a map of inline markers to the corresponding parser methods. |
| 256 |
* |
| 257 |
* This array defines handler methods for inline markdown markers. |
| 258 |
* When a marker is found in the text, the handler method is called with the text |
| 259 |
* starting at the position of the marker. |
| 260 |
* |
| 261 |
* Note that markers starting with whitespace may slow down the parser, |
| 262 |
* you may want to use [[renderText]] to deal with them. |
| 263 |
* |
| 264 |
* You may override this method to define a set of markers and parsing methods. |
| 265 |
* The default implementation looks for protected methods starting with `parse` that |
| 266 |
* also have an `@marker` annotation in PHPDoc. |
| 267 |
* |
| 268 |
* @return array a map of markers to parser methods |
| 269 |
*/ |
| 270 |
protected function inlineMarkers() |
| 271 |
{ |
| 272 |
$markers = []; |
| 273 |
// detect "parse" functions |
| 274 |
$reflection = new \ReflectionClass($this); |
| 275 |
foreach($reflection->getMethods(ReflectionMethod::IS_PROTECTED) as $method) { |
| 276 |
$methodName = $method->getName(); |
| 277 |
if (strncmp($methodName, 'parse', 5) === 0) { |
| 278 |
preg_match_all('/@marker ([^\s]+)/', $method->getDocComment(), $matches); |
| 279 |
foreach($matches[1] as $match) { |
| 280 |
$markers[$match] = $methodName; |
| 281 |
} |
| 282 |
} |
| 283 |
} |
| 284 |
return $markers; |
| 285 |
} |
| 286 |
|
| 287 |
/** |
| 288 |
* Prepare markers that are used in the text to parse |
| 289 |
* |
| 290 |
* Add all markers that are present in markdown. |
| 291 |
* Check is done to avoid iterations in parseInline(), good for huge markdown files |
| 292 |
* @param string $text |
| 293 |
*/ |
| 294 |
protected function prepareMarkers($text) |
| 295 |
{ |
| 296 |
$this->_inlineMarkers = []; |
| 297 |
foreach ($this->inlineMarkers() as $marker => $method) { |
| 298 |
if (strpos($text, $marker) !== false) { |
| 299 |
$m = $marker[0]; |
| 300 |
// put the longest marker first |
| 301 |
if (isset($this->_inlineMarkers[$m])) { |
| 302 |
reset($this->_inlineMarkers[$m]); |
| 303 |
if (strlen($marker) > strlen(key($this->_inlineMarkers[$m]))) { |
| 304 |
$this->_inlineMarkers[$m] = array_merge([$marker => $method], $this->_inlineMarkers[$m]); |
| 305 |
continue; |
| 306 |
} |
| 307 |
} |
| 308 |
$this->_inlineMarkers[$m][$marker] = $method; |
| 309 |
} |
| 310 |
} |
| 311 |
} |
| 312 |
|
| 313 |
/** |
| 314 |
* Parses inline elements of the language. |
| 315 |
* |
| 316 |
* @param string $text the inline text to parse. |
| 317 |
* @return array |
| 318 |
*/ |
| 319 |
protected function parseInline($text) |
| 320 |
{ |
| 321 |
if ($this->_depth >= $this->maximumNestingLevel) { |
| 322 |
// maximum depth is reached, do not parse input |
| 323 |
return [['text', $text]]; |
| 324 |
} |
| 325 |
$this->_depth++; |
| 326 |
|
| 327 |
$markers = implode('', array_keys($this->_inlineMarkers)); |
| 328 |
|
| 329 |
$paragraph = []; |
| 330 |
|
| 331 |
while (!empty($markers) && ($found = strpbrk($text, $markers)) !== false) { |
| 332 |
|
| 333 |
$pos = strpos($text, $found); |
| 334 |
|
| 335 |
// add the text up to next marker to the paragraph |
| 336 |
if ($pos !== 0) { |
| 337 |
$paragraph[] = ['text', substr($text, 0, $pos)]; |
| 338 |
} |
| 339 |
$text = $found; |
| 340 |
|
| 341 |
$parsed = false; |
| 342 |
foreach ($this->_inlineMarkers[$text[0]] as $marker => $method) { |
| 343 |
if (strncmp($text, $marker, strlen($marker)) === 0) { |
| 344 |
// parse the marker |
| 345 |
array_unshift($this->context, $method); |
| 346 |
list($output, $offset) = $this->$method($text); |
| 347 |
array_shift($this->context); |
| 348 |
|
| 349 |
$paragraph[] = $output; |
| 350 |
$text = substr($text, $offset); |
| 351 |
$parsed = true; |
| 352 |
break; |
| 353 |
} |
| 354 |
} |
| 355 |
if (!$parsed) { |
| 356 |
$paragraph[] = ['text', substr($text, 0, 1)]; |
| 357 |
$text = substr($text, 1); |
| 358 |
} |
| 359 |
} |
| 360 |
|
| 361 |
$paragraph[] = ['text', $text]; |
| 362 |
|
| 363 |
$this->_depth--; |
| 364 |
|
| 365 |
return $paragraph; |
| 366 |
} |
| 367 |
|
| 368 |
/** |
| 369 |
* Parses escaped special characters. |
| 370 |
* @marker \ |
| 371 |
*/ |
| 372 |
protected function parseEscape($text) |
| 373 |
{ |
| 374 |
if (isset($text[1]) && in_array($text[1], $this->escapeCharacters)) { |
| 375 |
return [['text', $text[1]], 2]; |
| 376 |
} |
| 377 |
return [['text', $text[0]], 1]; |
| 378 |
} |
| 379 |
|
| 380 |
/** |
| 381 |
* This function renders plain text sections in the markdown text. |
| 382 |
* It can be used to work on normal text sections for example to highlight keywords or |
| 383 |
* do special escaping. |
| 384 |
*/ |
| 385 |
protected function renderText($block) |
| 386 |
{ |
| 387 |
return $block[1]; |
| 388 |
} |
| 389 |
} |
| 390 |
|
