网站首页  汉语字词  英语词汇  考试资料  写作素材  旧版资料

请输入您要查询的范文:

 

标题 php提取网页正文内容的例子
范文
    因为难点在于如何去识别并保留网页中的文章部分,而且删除其它无用的信息,并且要做到通用化,不能像火车头那样根据目标站来制定采集规则,因为搜索引擎结果中有各种的网页。
    抓回一个页面的数据,如何匹配出正文部分,郑晓在下班路上想了个思路是:
    1. 提取出body标签部分–>剔除所有链接–>剔除所有script、注释–>剔除所有空白标签(包括标签内不含中文的)–>获取结果。
    2. 直接匹配出非链接的、 符合在div、p、h标签中的中文部分???
    还是会有不少其它多余信息啊,比如底部信息等。。 如何搞?不知道大家有木有什么思路或建议?
    这个类是从网上找到的一个php实现的提取网页正文部分的算法,郑晓在本地也测试了下,准确率非常高。
    代码如下:
    <?php
    class Readability {
    // 保存判定结果的标记位名称
    const ATTR_CONTENT_SCORE = "contentScore";
    // DOM 解析类目前只支持 UTF-8 编码
    const DOM_DEFAULT_CHARSET = "utf-8";
    // 当判定失败时显示的内容
    const MESSAGE_CAN_NOT_GET = "Readability was unable to parse this page for content.";
    // DOM 解析类(PHP5 已内置)
    protected $DOM = null;
    // 需要解析的源代码
    protected $source = "";
    // 章节的父元素列表
    private $parentNodes = array();
    // 需要删除的标签
    // Note: added extra tags from
    private $junkTags = Array("style", "form", "iframe", "script", "button", "input", "textarea",
    "noscript", "select", "option", "object", "applet", "basefont",
    "bgsound", "blink", "canvas", "command", "menu", "nav", "datalist",
    "embed", "frame", "frameset", "keygen", "label", "marquee", "link");
    // 需要删除的属性
    private $junkAttrs = Array("style", "class", "onclick", "onmouseover", "align", "border", "margin");
    /**
    * 构造函数
    * @param $input_char 字符串的编码。默认 utf-8,可以省略
    */
    function __construct($source, $input_char = "utf-8") {
    $this->source = $source;
    // DOM 解析类只能处理 UTF-8 格式的字符
    $source = mb_convert_encoding($source, 'HTML-ENTITIES', $input_char);
    // 预处理 HTML 标签,剔除冗余的标签等
    $source = $this->preparSource($source);
    // 生成 DOM 解析类
    $this->DOM = new DOMDocument('1.0', $input_char);
    try {
    //libxml_use_internal_errors(true);
    // 会有些错误信息,不过不要紧 :^)
    if ( encoding="'.Readability::DOM_DEFAULT_CHARSET.'">'.$source)) {
    throw new Exception("Parse HTML Error!");
    }
    foreach ($this->DOM->childNodes as $item) {
    if ($item->nodeType == XML_PI_NODE) {
    $this->DOM->removeChild($item); // remove hack
    }
    }
    // insert proper
    $this->DOM->encoding = Readability::DOM_DEFAULT_CHARSET;
    } catch (Exception $e) {
    // ...
    }
    }
    /**
    * 预处理 HTML 标签,使其能够准确被 DOM 解析类处理
    *
    * @return String
    */
    private function preparSource($string) {
    // 剔除多余的 HTML 编码标记,避免解析出错
    preg_match("/charset=([\w|\-]+);?/", $string, $match);
    if (isset($match[1])) {
    $string = preg_replace("/charset=([\w|\-]+);?/", "", $string, 1);
    }
    // Replace all doubled-up <BR> tags with <P> tags, and remove fonts.
    $string = preg_replace("/<br\/?>[ \r\n\s]*<br\/?>/i", "</p><p>", $string);
    $string = preg_replace("/<\/?font[^>]*>/i", "", $string);
    // @see
    // - from
    $string = preg_replace("#<script(.*?)>(.*?)</script>#is", "", $string);
    return trim($string);
    }
    /**
    * 删除 DOM 元素中所有的 $TagName 标签
    *
    * @return DOMDocument
    */
    private function removeJunkTag($RootNode, $TagName) {
    $Tags = $RootNode->getElementsByTagName($TagName);
    //Note: always index 0, because removing a tag removes it from the results as well.
    while($Tag = $Tags->item(0)){
    $parentNode = $Tag->parentNode;
    $parentNode->removeChild($Tag);
    }
    return $RootNode;
    }
    /**
    * 删除元素中所有不需要的属性
    */
    private function removeJunkAttr($RootNode, $Attr) {
    $Tags = $RootNode->getElementsByTagName("*");
    $i = 0;
    while($Tag = $Tags->item($i++)) {
    $Tag->removeAttribute($Attr);
    }
    return $RootNode;
    }
    /**
    * 根据评分获取页面主要内容的盒模型
    * 判定算法来自:
    * 这里由郑晓博客转发
    * @return DOMNode
    */
    private function getTopBox() {
    // 获得页面所有的章节
    $allParagraphs = $this->DOM->getElementsByTagName("p");
    // Study all the paragraphs and find the chunk that has the best score.
    // A score is determined by things like: Number of <p>'s, commas, special classes, etc.
    $i = 0;
    while($paragraph = $allParagraphs->item($i++)) {
    $parentNode = $paragraph->parentNode;
    $contentScore = intval($parentNode->getAttribute(Readability::ATTR_CONTENT_SCORE));
    $className = $parentNode->getAttribute("class");
    $id = $parentNode->getAttribute("id");
    // Look for a special classname
    if (preg_match("/(comment|meta|footer|footnote)/i", $className)) {
    $contentScore -= 50;
    } else if(preg_match(
    "/((^|\\s)(post|hentry|entry[-]?(content|text|body)?|article[-]?(content|text|body)?)(\\s|$))/i",
    $className)) {
    $contentScore += 25;
    }
    // Look for a special ID
    if (preg_match("/(comment|meta|footer|footnote)/i", $id)) {
    $contentScore -= 50;
    } else if (preg_match(
    "/^(post|hentry|entry[-]?(content|text|body)?|article[-]?(content|text|body)?)$/i",
    $id)) {
    $contentScore += 25;
    }
    // Add a point for the paragraph found
    // Add points for any commas within this paragraph
    if (strlen($paragraph->nodeValue) > 10) {
    $contentScore += strlen($paragraph->nodeValue);
    }
    // 保存父元素的判定得分
    $parentNode->setAttribute(Readability::ATTR_CONTENT_SCORE, $contentScore);
    // 保存章节的父元素,以便下次快速获取
    array_push($this->parentNodes, $parentNode);
    }
    $topBox = null;
    // Assignment from index for performance.
    // See
    for ($i = 0, $len = sizeof($this->parentNodes); $i < $len; $i++) {
    $parentNode = $this->parentNodes[$i];
    $contentScore = intval($parentNode->getAttribute(Readability::ATTR_CONTENT_SCORE));
    $orgContentScore = intval($topBox ? $topBox->getAttribute(Readability::ATTR_CONTENT_SCORE) : 0);
    if ($contentScore && $contentScore > $orgContentScore) {
    $topBox = $parentNode;
    }
    }
    // 此时,$topBox 应为已经判定后的页面内容主元素
    return $topBox;
    }
    /**
    * 获取 HTML 页面标题
    *
    * @return String
    */
    public function getTitle() {
    $split_point = ' - ';
    $titleNodes = $this->DOM->getElementsByTagName("title");
    if ($titleNodes->length
    && $titleNode = $titleNodes->item(0)) {
    // @see
    $title = trim($titleNode->nodeValue);
    $result = array_map('strrev', explode($split_point, strrev($title)));
    return sizeof($result) > 1 ? array_pop($result) : $title;
    }
    return null;
    }
    /**
    * Get Leading Image Url
    *
    * @return String
    */
    public function getLeadImageUrl($node) {
    $images = $node->getElementsByTagName("img");
    if ($images->length && $leadImage = $images->item(0)) {
    return $leadImage->getAttribute("src");
    }
    return null;
    }
    /**
    * 获取页面的主要内容(Readability 以后的内容)
    *
    * @return Array
    */
    public function getContent() {
    if (!$this->DOM) return false;
    // 获取页面标题
    $ContentTitle = $this->getTitle();
    // 获取页面主内容
    $ContentBox = $this->getTopBox();
    //Check if we found a suitable top-box.
    if($ContentBox === null)
    throw new RuntimeException(Readability::MESSAGE_CAN_NOT_GET);
    // 复制内容到新的 DOMDocument
    $Target = new DOMDocument;
    $Target->appendChild($Target->importNode($ContentBox, true));
    // 删除不需要的标签
    foreach ($this->junkTags as $tag) {
    $Target = $this->removeJunkTag($Target, $tag);
    }
    // 删除不需要的属性
    foreach ($this->junkAttrs as $attr) {
    $Target = $this->removeJunkAttr($Target, $attr);
    }
    $content = mb_convert_encoding($Target->saveHTML(), Readability::DOM_DEFAULT_CHARSET, "HTML-ENTITIES");
    // 多个数据,以数组的形式返回
    return Array(
    'lead_image_url' => $this->getLeadImageUrl($Target),
    'word_count' => mb_strlen(strip_tags($content), Readability::DOM_DEFAULT_CHARSET),
    'title' => $ContentTitle ? $ContentTitle : null,
    'content' => $content
    );
    }
    function __destruct() { }
    }
    使用起来也非常简单,实例化时传入网页的html源码和相应的编码,然后直接调用其getContent方法即可返回提取到的正文部分,提取出的文章中可能还会含有少部分链接,可以自己后期再修改
随便看

 

在线学习网范文大全提供好词好句、学习总结、工作总结、演讲稿等写作素材及范文模板,是学习及工作的有利工具。

 

Copyright © 2002-2024 cuapp.net All Rights Reserved
更新时间:2025/5/18 0:17:01