baidu-indexing-push.php

<?php
/**
 * Sitemap/Feed 最新URL提取工具
 * 用于从sitemap或RSS/Atom feed中提取最新的10个URL地址
 */

class SitemapParser {
    private $sitemapUrl;
    
    public function __construct($sitemapUrl) {
        $this->sitemapUrl = $sitemapUrl;
    }
    
    /**
     * 获取并解析sitemap
     */
    public function getLatestUrls($limit = 10) {
        // 获取sitemap内容
        $content = $this->fetchSitemap();
        if (!$content) {
            throw new Exception("无法获取sitemap内容");
        }
        
        // 解析XML
        $urls = $this->parseSitemap($content);
        
        // 按日期排序并获取最新的URL
        usort($urls, function($a, $b) {
            $dateA = isset($a['lastmod']) ? strtotime($a['lastmod']) : 0;
            $dateB = isset($b['lastmod']) ? strtotime($b['lastmod']) : 0;
            return $dateB - $dateA;
        });
        
        // 返回最新的URL
        return array_slice($urls, 0, $limit);
    }
    
    /**
     * 获取sitemap内容
     */
    private function fetchSitemap() {
        $ch = curl_init();
        curl_setopt($ch, CURLOPT_URL, $this->sitemapUrl);
        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
        curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
        curl_setopt($ch, CURLOPT_TIMEOUT, 30);
        curl_setopt($ch, CURLOPT_USERAGENT, 'Sitemap Parser 1.0');
        
        $content = curl_exec($ch);
        $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
        curl_close($ch);
        
        if ($httpCode !== 200) {
            return false;
        }
        
        return $content;
    }
    
    /**
     * 解析sitemap XML
     */
    private function parseSitemap($content) {
        $urls = [];
        $xml = simplexml_load_string($content);
        
        if ($xml === false) {
            throw new Exception("无效的XML格式");
        }
        
        // 处理索引sitemap (包含多个sitemap文件)
        if (isset($xml->sitemap)) {
            foreach ($xml->sitemap as $sitemap) {
                if (isset($sitemap->loc)) {
                    try {
                        $subContent = $this->fetchSitemap((string)$sitemap->loc);
                        if ($subContent) {
                            $subUrls = $this->parseSitemap($subContent);
                            $urls = array_merge($urls, $subUrls);
                        }
                    } catch (Exception $e) {
                        // 忽略子sitemap错误
                        continue;
                    }
                }
            }
        }
        // 处理普通sitemap (包含URL列表)
        else if (isset($xml->url)) {
            foreach ($xml->url as $url) {
                $urlData = [
                    'loc' => (string)$url->loc,
                ];
                
                if (isset($url->lastmod)) {
                    $urlData['lastmod'] = (string)$url->lastmod;
                }
                
                if (isset($url->changefreq)) {
                    $urlData['changefreq'] = (string)$url->changefreq;
                }
                
                if (isset($url->priority)) {
                    $urlData['priority'] = (string)$url->priority;
                }
                
                $urls[] = $urlData;
            }
        }
        
        return $urls;
    }
}

class FeedParser {
    private $feedUrl;
    
    public function __construct($feedUrl) {
        $this->feedUrl = $feedUrl;
    }
    
    /**
     * 获取并解析feed，返回最新的URL
     */
    public function getLatestUrls($limit = 10) {
        // 获取feed内容
        $content = $this->fetchFeed();
        if (!$content) {
            throw new Exception("无法获取feed内容");
        }
        
        // 解析feed
        $urls = $this->parseFeed($content);
        
        // 按日期排序并获取最新的URL
        usort($urls, function($a, $b) {
            $dateA = isset($a['timestamp']) ? $a['timestamp'] : 0;
            $dateB = isset($b['timestamp']) ? $b['timestamp'] : 0;
            return $dateB - $dateA;
        });
        
        // 返回最新的URL
        return array_slice($urls, 0, $limit);
    }
    
    /**
     * 获取feed内容
     */
    private function fetchFeed() {
        $ch = curl_init();
        curl_setopt($ch, CURLOPT_URL, $this->feedUrl);
        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
        curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
        curl_setopt($ch, CURLOPT_TIMEOUT, 30);
        curl_setopt($ch, CURLOPT_USERAGENT, 'Feed Parser 1.0');
        
        $content = curl_exec($ch);
        $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
        curl_close($ch);
        
        if ($httpCode !== 200) {
            return false;
        }
        
        return $content;
    }
    
    /**
     * 解析feed内容
     */
    private function parseFeed($content) {
        $urls = [];
        
        // 尝试解析为RSS
        $rss = @simplexml_load_string($content);
        if ($rss && isset($rss->channel->item)) {
            foreach ($rss->channel->item as $item) {
                $urlData = [
                    'loc' => (string)$item->link,
                    'title' => (string)$item->title,
                    'timestamp' => isset($item->pubDate) ? strtotime((string)$item->pubDate) : time()
                ];
                
                if (isset($item->description)) {
                    $urlData['description'] = (string)$item->description;
                }
                
                $urls[] = $urlData;
            }
            return $urls;
        }
        
        // 尝试解析为Atom
        $atom = @simplexml_load_string($content);
        if ($atom && isset($atom->entry)) {
            foreach ($atom->entry as $entry) {
                $urlData = [
                    'loc' => (string)$entry->link['href'],
                    'title' => (string)$entry->title,
                    'timestamp' => isset($entry->updated) ? strtotime((string)$entry->updated) : 
                                  (isset($entry->published) ? strtotime((string)$entry->published) : time())
                ];
                
                if (isset($entry->summary)) {
                    $urlData['description'] = (string)$entry->summary;
                }
                
                $urls[] = $urlData;
            }
            return $urls;
        }
        
        throw new Exception("无效的Feed格式");
    }
}

// Web界面处理
if ($_SERVER['REQUEST_METHOD'] === 'GET') {
    showForm();
} else if ($_SERVER['REQUEST_METHOD'] === 'POST') {
    if (isset($_POST['action']) && $_POST['action'] === 'push') {
        handlePush();
    } else {
        handlePost();
    }
}

function showForm() {
?>
<!DOCTYPE html>
<html>
<head>
    <meta charset="UTF-8">
    <title>Sitemap/Feed 最新URL提取工具</title>
    <style>
        body { font-family: Arial, sans-serif; margin: 40px; background-color: #f5f5f5; }
        .container { max-width: 800px; margin: 0 auto; background: white; padding: 30px; border-radius: 8px; box-shadow: 0 2px 10px rgba(0,0,0,0.1); }
        h1 { color: #333; text-align: center; }
        .form-group { margin-bottom: 20px; }
        label { display: block; margin-bottom: 5px; font-weight: bold; }
        input[type="url"] { width: 100%; padding: 12px; border: 1px solid #ddd; border-radius: 4px; box-sizing: border-box; }
        button { background-color: #007cba; color: white; padding: 12px 24px; border: none; border-radius: 4px; cursor: pointer; font-size: 16px; }
        button:hover { background-color: #005a87; }
        .result { margin-top: 30px; }
        .url-item { padding: 10px; border-bottom: 1px solid #eee; }
        .url-link { color: #007cba; text-decoration: none; }
        .url-link:hover { text-decoration: underline; }
        .lastmod { color: #666; font-size: 14px; margin-top: 5px; }
        .error { color: #d63638; background: #fcf0f1; padding: 15px; border-radius: 4px; margin: 20px 0; }
        .success { color: #008a20; background: #edfaef; padding: 15px; border-radius: 4px; margin: 20px 0; }
        .push-section { margin-top: 30px; padding: 20px; background: #f9f9f9; border-radius: 4px; }
        .push-form { margin-top: 15px; }
        .push-form input[type="text"] { width: 100%; padding: 8px; margin: 10px 0; }
        .push-button { background-color: #333; color: white; padding: 10px 20px; border: none; border-radius: 4px; cursor: pointer; }
        .push-button:hover { background-color: #555; }
        .push-result { margin-top: 20px; padding: 15px; border-radius: 4px; }
        .push-success { background: #edfaef; color: #008a20; }
        .push-error { background: #fcf0f1; color: #d63638; }
        .source-type { margin-bottom: 15px; }
        .source-type label { display: inline-block; margin-right: 20px; font-weight: normal; }
        .source-type input { margin-right: 5px; }
        .description { color: #666; font-size: 14px; margin-top: 5px; }
    </style>
</head>
<body>
    <div class="container">
        <h1>Sitemap/Feed 最新URL提取工具</h1>
        <form method="post">
            <div class="source-type">
                <label><input type="radio" name="source_type" value="sitemap" checked> Sitemap</label>
                <label><input type="radio" name="source_type" value="feed"> RSS/Atom Feed</label>
            </div>
            
            <div class="form-group">
                <label for="source_url">请输入地址:</label>
                <input type="url" id="source_url" name="source_url" placeholder="https://example.com/sitemap.xml 或 https://example.com/feed" required>
            </div>
            <button type="submit">获取最新URL</button>
        </form>
    </div>
</body>
</html>
<?php
}

function handlePost() {
    $sourceType = $_POST['source_type'] ?? 'sitemap';
    $sourceUrl = $_POST['source_url'] ?? '';
    
    if (empty($sourceUrl)) {
        echo '<div class="error">请输入有效的地址</div>';
        showForm();
        return;
    }
    
    try {
        if ($sourceType === 'sitemap') {
            $parser = new SitemapParser($sourceUrl);
        } else {
            $parser = new FeedParser($sourceUrl);
        }
        
        $latestUrls = $parser->getLatestUrls(10);
        
        ?>
        <!DOCTYPE html>
        <html>
        <head>
            <meta charset="UTF-8">
            <title>结果 - 最新URL提取工具</title>
            <style>
                body { font-family: Arial, sans-serif; margin: 40px; background-color: #f5f5f5; }
                .container { max-width: 800px; margin: 0 auto; background: white; padding: 30px; border-radius: 8px; box-shadow: 0 2px 10px rgba(0,0,0,0.1); }
                h1 { color: #333; text-align: center; }
                .back-link { margin-bottom: 20px; }
                a { color: #007cba; text-decoration: none; }
                a:hover { text-decoration: underline; }
                .result { margin-top: 20px; }
                .url-item { padding: 15px; border-bottom: 1px solid #eee; }
                .url-item:last-child { border-bottom: none; }
                .url-link { color: #007cba; text-decoration: none; font-weight: bold; }
                .url-link:hover { text-decoration: underline; }
                .lastmod { color: #666; font-size: 14px; margin-top: 5px; }
                .no-results { color: #666; text-align: center; padding: 20px; }
                .push-section { margin-top: 30px; padding: 20px; background: #f9f9f9; border-radius: 4px; }
                .push-form { margin-top: 15px; }
                .push-form input[type="text"] { width: 100%; padding: 8px; margin: 10px 0; }
                .push-button { background-color: #333; color: white; padding: 10px 20px; border: none; border-radius: 4px; cursor: pointer; }
                .push-button:hover { background-color: #555; }
                .push-result { margin-top: 20px; padding: 15px; border-radius: 4px; }
                .push-success { background: #edfaef; color: #008a20; }
                .push-error { background: #fcf0f1; color: #d63638; }
                .source-info { background: #eef7fa; padding: 10px; border-radius: 4px; margin-bottom: 20px; }
                .description { color: #666; font-size: 14px; margin-top: 5px; }
            </style>
        </head>
        <body>
            <div class="container">
                <div class="back-link">
                    <a href="<?php echo $_SERVER['PHP_SELF']; ?>">&larr; 返回</a>
                </div>
                
                <h1>最新更新的URL</h1>
                
                <div class="source-info">
                    来源类型: <?php echo $sourceType === 'sitemap' ? 'Sitemap' : 'RSS/Atom Feed'; ?><br>
                    来源地址: <?php echo htmlspecialchars($sourceUrl); ?>
                </div>
                
                <div class="result">
                    <?php if (empty($latestUrls)): ?>
                        <div class="no-results">未找到任何URL</div>
                    <?php else: ?>
                        <?php foreach ($latestUrls as $index => $url): ?>
                            <div class="url-item">
                                <div>
                                    <a href="<?php echo htmlspecialchars($url['loc']); ?>" target="_blank" class="url-link">
                                        <?php echo htmlspecialchars($url['title'] ?? $url['loc']); ?>
                                    </a>
                                </div>
                                <?php if (isset($url['lastmod']) || isset($url['timestamp'])): ?>
                                    <div class="lastmod">
                                        <?php 
                                        if (isset($url['lastmod'])) {
                                            echo '最后修改时间: ' . date('Y-m-d H:i:s', strtotime($url['lastmod']));
                                        } elseif (isset($url['timestamp'])) {
                                            echo '发布时间: ' . date('Y-m-d H:i:s', $url['timestamp']);
                                        }
                                        ?>
                                    </div>
                                <?php endif; ?>
                                
                                <?php if (isset($url['description'])): ?>
                                    <div class="description">
                                        <?php echo htmlspecialchars(mb_substr(strip_tags($url['description']), 0, 150)) . '...'; ?>
                                    </div>
                                <?php endif; ?>
                            </div>
                        <?php endforeach; ?>
                    <?php endif; ?>
                </div>
                
                <?php if (!empty($latestUrls)): ?>
                <div class="push-section">
                    <h3>推送到百度</h3>
                    <p>是否要将以上URL推送到百度站长平台？</p>
                    <form method="post" class="push-form">
                        <input type="hidden" name="action" value="push">
                        <input type="hidden" name="source_type" value="<?php echo htmlspecialchars($sourceType); ?>">
                        <?php foreach ($latestUrls as $url): ?>
                            <input type="hidden" name="urls[]" value="<?php echo htmlspecialchars($url['loc']); ?>">
                        <?php endforeach; ?>
                        <div class="form-group">
                            <label for="baidu_token">百度推送密钥:</label>
                            <input type="text" id="baidu_token" name="baidu_token" placeholder="请输入百度站长平台的推送密钥" required>
                        </div>
                        <button type="submit" class="push-button">推送到百度</button>
                    </form>
                </div>
                <?php endif; ?>
            </div>
        </body>
        </html>
        <?php
    } catch (Exception $e) {
        ?>
        <!DOCTYPE html>
        <html>
        <head>
            <meta charset="UTF-8">
            <title>错误 - 最新URL提取工具</title>
            <style>
                body { font-family: Arial, sans-serif; margin: 40px; background-color: #f5f5f5; }
                .container { max-width: 800px; margin: 0 auto; background: white; padding: 30px; border-radius: 8px; box-shadow: 0 2px 10px rgba(0,0,0,0.1); }
                .back-link { margin-bottom: 20px; }
                a { color: #007cba; text-decoration: none; }
                a:hover { text-decoration: underline; }
                .error { color: #d63638; background: #fcf0f1; padding: 15px; border-radius: 4px; }
            </style>
        </head>
        <body>
            <div class="container">
                <div class="back-link">
                    <a href="<?php echo $_SERVER['PHP_SELF']; ?>">&larr; 返回重试</a>
                </div>
                
                <div class="error">
                    <strong>处理过程中发生错误:</strong><br>
                    <?php echo htmlspecialchars($e->getMessage()); ?>
                </div>
            </div>
        </body>
        </html>
        <?php
    }
}

function handlePush() {
    $urls = $_POST['urls'] ?? [];
    $token = $_POST['baidu_token'] ?? '';
    $sourceType = $_POST['source_type'] ?? 'sitemap';
    
    if (empty($token)) {
        echo '<div class="error">请输入百度推送密钥</div>';
        return;
    }
    
    if (empty($urls)) {
        echo '<div class="error">没有可推送的URL</div>';
        return;
    }
    
    // 提取站点域名
    $site = '';
    if (!empty($urls[0])) {
        $parsedUrl = parse_url($urls[0]);
        $site = $parsedUrl['scheme'] . '://' . $parsedUrl['host'];
    }
    
    $api = "http://data.zz.baidu.com/urls?site={$site}&token={$token}";
    
    $ch = curl_init();
    $options = array(
        CURLOPT_URL => $api,
        CURLOPT_POST => true,
        CURLOPT_RETURNTRANSFER => true,
        CURLOPT_POSTFIELDS => implode("\n", $urls),
        CURLOPT_HTTPHEADER => array('Content-Type: text/plain'),
    );
    curl_setopt_array($ch, $options);
    $result = curl_exec($ch);
    $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
    curl_close($ch);
    
    // 显示推送结果
    ?>
    <!DOCTYPE html>
    <html>
    <head>
        <meta charset="UTF-8">
        <title>推送结果 - 最新URL提取工具</title>
        <style>
            body { font-family: Arial, sans-serif; margin: 40px; background-color: #f5f5f5; }
            .container { max-width: 800px; margin: 0 auto; background: white; padding: 30px; border-radius: 8px; box-shadow: 0 2px 10px rgba(0,0,0,0.1); }
            h1 { color: #333; text-align: center; }
            .back-link { margin-bottom: 20px; }
            a { color: #007cba; text-decoration: none; }
            a:hover { text-decoration: underline; }
            .push-result { margin-top: 20px; padding: 15px; border-radius: 4px; }
            .push-success { background: #edfaef; color: #008a20; }
            .push-error { background: #fcf0f1; color: #d63638; }
            .url-list { margin-top: 20px; }
            .url-item { padding: 5px 0; border-bottom: 1px solid #eee; }
            .source-info { background: #eef7fa; padding: 10px; border-radius: 4px; margin-bottom: 20px; }
        </style>
    </head>
    <body>
        <div class="container">
            <div class="back-link">
                <a href="<?php echo $_SERVER['PHP_SELF']; ?>">&larr; 返回首页</a>
            </div>
            
            <h1>百度推送结果</h1>
            
            <div class="source-info">
                来源类型: <?php echo $sourceType === 'sitemap' ? 'Sitemap' : 'RSS/Atom Feed'; ?>
            </div>
            
            <?php if ($httpCode !== 200): ?>
                <div class="push-result push-error">
                    <strong>推送失败:</strong><br>
                    HTTP状态码: <?php echo $httpCode; ?>
                </div>
            <?php else: 
                $response = json_decode($result, true);
                ?>
                <div class="push-result <?php echo isset($response['error']) ? 'push-error' : 'push-success'; ?>">
                    <?php if (isset($response['error'])): ?>
                        <strong>推送失败:</strong><br>
                        错误代码: <?php echo htmlspecialchars($response['error']); ?><br>
                        错误信息: <?php echo htmlspecialchars($response['message']); ?>
                    <?php else: ?>
                        <strong>推送成功!</strong><br>
                        成功推送: <?php echo $response['success'] ?? 'N/A'; ?> 条URL<br>
                        剩余配额: <?php echo $response['remain'] ?? 'N/A'; ?>
                    <?php endif; ?>
                </div>
            <?php endif; ?>
            
            <div class="url-list">
                <h3>推送的URL列表:</h3>
                <?php foreach ($urls as $url): ?>
                    <div class="url-item"><?php echo htmlspecialchars($url); ?></div>
                <?php endforeach; ?>
            </div>
        </div>
    </body>
    </html>
    <?php
}
?>