tags with
tags, and remove fonts.
104 | $string = preg_replace("/
", $string);
105 | $string = preg_replace("/<\/?font[^>]*>/i", "", $string);
106 |
107 | // @see https://github.com/feelinglucky/php-readability/issues/7
108 | // - from http://stackoverflow.com/questions/7130867/remove-script-tag-from-html-content
109 | $string = preg_replace("# 's, commas, special classes, etc.
160 | $i = 0;
161 | while($paragraph = $allParagraphs->item($i++)) {
162 | $parentNode = $paragraph->parentNode;
163 | $contentScore = intval($parentNode->getAttribute(Readability::ATTR_CONTENT_SCORE));
164 | $className = $parentNode->getAttribute("class");
165 | $id = $parentNode->getAttribute("id");
166 |
167 | // Look for a special classname
168 | if (preg_match("/(comment|meta|footer|footnote)/i", $className)) {
169 | $contentScore -= 50;
170 | } else if(preg_match(
171 | "/((^|\\s)(section|post|hentry|entry[-]?(content|text|body)?|article[-]?(content|text|body)?)(\\s|$))/i",
172 | $className)) {
173 | $contentScore += 25;
174 | }
175 |
176 | // Look for a special ID
177 | if (preg_match("/(comment|meta|footer|footnote)/i", $id)) {
178 | $contentScore -= 50;
179 | } else if (preg_match(
180 | "/^(post|hentry|entry[-]?(content|text|body)?|article[-]?(content|text|body)?)$/i",
181 | $id)) {
182 | $contentScore += 25;
183 | }
184 |
185 | // Add a point for the paragraph found
186 | // Add points for any commas within this paragraph
187 | if (strlen($paragraph->nodeValue) > 10) {
188 | $contentScore += strlen($paragraph->nodeValue);
189 | }
190 |
191 | // 保存父元素的判定得分
192 | $parentNode->setAttribute(Readability::ATTR_CONTENT_SCORE, $contentScore);
193 |
194 | // 保存章节的父元素，以便下次快速获取
195 | array_push($this->parentNodes, $parentNode);
196 | }
197 |
198 | $topBox = null;
199 |
200 | // Assignment from index for performance.
201 | // See http://www.peachpit.com/articles/article.aspx?p=31567&seqNum=5
202 | for ($i = 0, $len = sizeof($this->parentNodes); $i < $len; $i++) {
203 | $parentNode = $this->parentNodes[$i];
204 | $contentScore = intval($parentNode->getAttribute(Readability::ATTR_CONTENT_SCORE));
205 | $orgContentScore = intval($topBox ? $topBox->getAttribute(Readability::ATTR_CONTENT_SCORE) : 0);
206 |
207 | // by raywill, 2016-9-2
208 | // for case: xxx yyy
209 | if ($parentNode && $topBox && $topBox->parentNode
210 | && $parentNode !== $topBox
211 | && $parentNode->parentNode === $topBox->parentNode
212 | && $this->scoreMatch($parentNode, $topBox)) { // trust same level
213 |
214 | $topScore = intval($topBox->getAttribute(Readability::ATTR_CONTENT_SCORE));
215 | $topBox = $topBox->parentNode;
216 | $topBox->setAttribute(Readability::ATTR_CONTENT_SCORE, $topScore + $contentScore);
217 | } else if ($contentScore && $contentScore > $orgContentScore) {
218 |
219 | $topBox = $parentNode;
220 | }
221 | }
222 |
223 | // 此时，$topBox 应为已经判定后的页面内容主元素
224 | return $topBox;
225 | }
226 |
227 | protected function scoreMatch($n1, $n2) {
228 | $n1Score = intval($n1->getAttribute(Readability::ATTR_CONTENT_SCORE));
229 | $n2Score = intval($n2->getAttribute(Readability::ATTR_CONTENT_SCORE));
230 | return ($n1Score > 0 && $n2Score > 0);
231 | }
232 |
233 | /**
234 | * 获取 HTML 页面标题
235 | *
236 | * @return String
237 | */
238 | public function getTitle() {
239 | $split_point = ' - ';
240 | $titleNodes = $this->DOM->getElementsByTagName("title");
241 |
242 | if ($titleNodes->length
243 | && $titleNode = $titleNodes->item(0)) {
244 | // @see http://stackoverflow.com/questions/717328/how-to-explode-string-right-to-left
245 | $title = trim($titleNode->nodeValue);
246 | $result = array_map('strrev', explode($split_point, strrev($title)));
247 | return sizeof($result) > 1 ? array_pop($result) : $title;
248 | }
249 |
250 | return null;
251 | }
252 |
253 |
254 | /**
255 | * Get Leading Image Url
256 | *
257 | * @return String
258 | */
259 | public function getLeadImageUrl($node) {
260 | $images = $node->getElementsByTagName("img");
261 |
262 | if ($images->length){
263 | $i = 0;
264 | while($leadImage = $images->item($i++)) {
265 | $imgsrc = $leadImage->getAttribute("src");
266 | $imgdatasrc = $leadImage->getAttribute("data-src");
267 | $imgsrclast = $imgsrc ? $imgsrc : $imgdatasrc;
268 | list($img['width'],$img['height'])=getimagesize($imgsrclast);
269 | if($img['width'] > 150 && $img['height'] >150){
270 | return $imgsrclast;
271 | }
272 |
273 | }
274 | }
275 |
276 | return null;
277 | }
278 |
279 |
280 | /**
281 | * 获取页面的主要内容（Readability 以后的内容）
282 | *
283 | * @return Array
284 | */
285 | public function getContent() {
286 | if (!$this->DOM) return false;
287 |
288 | // 获取页面标题
289 | $ContentTitle = $this->getTitle();
290 |
291 | // 获取页面主内容
292 | $ContentBox = $this->getTopBox();
293 |
294 | //Check if we found a suitable top-box.
295 | if($ContentBox === null)
296 | throw new RuntimeException(Readability::MESSAGE_CAN_NOT_GET);
297 |
298 | // 复制内容到新的 DOMDocument
299 | $Target = new DOMDocument;
300 | $Target->appendChild($Target->importNode($ContentBox, true));
301 |
302 | // 删除不需要的标签
303 | foreach ($this->junkTags as $tag) {
304 | $Target = $this->removeJunkTag($Target, $tag);
305 | }
306 |
307 | // 删除不需要的属性
308 | foreach ($this->junkAttrs as $attr) {
309 | $Target = $this->removeJunkAttr($Target, $attr);
310 | }
311 |
312 | $content = mb_convert_encoding($Target->saveHTML(), Readability::DOM_DEFAULT_CHARSET, "HTML-ENTITIES");
313 |
314 | // 多个数据，以数组的形式返回
315 | return Array(
316 | 'lead_image_url' => $this->getLeadImageUrl($Target),
317 | 'word_count' => mb_strlen(strip_tags($content), Readability::DOM_DEFAULT_CHARSET),
318 | 'title' => $ContentTitle ? $ContentTitle : null,
319 | 'content' => $content
320 | );
321 | }
322 |
323 | function __destruct() { }
324 | }
325 |
326 |
--------------------------------------------------------------------------------
/template/footer.inc.html:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
8 |
14 |
15 |
43 |
--------------------------------------------------------------------------------
/template/header.inc.html:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
--------------------------------------------------------------------------------
/template/index.html:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 | 無標題文檔
11 | 
12 |
15 |
68 |
69 |
70 |
71 |
--------------------------------------------------------------------------------
/template/reader.html:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 | 無標題文檔
12 |
13 |