DuckDuckGo URL Collector (Rewritten)

Collects URLs from DuckDuckGo with optional site: filtering and rate limiting, updated for modern DuckDuckGo

当前为 2025-07-28 提交的版本,查看 最新版本

您需要先安装一款用户脚本管理器扩展,例如 Tampermonkey 篡改猴Greasemonkey 油猴子Violentmonkey 暴力猴,才能安装此脚本。

您需要先安装一款用户脚本管理器扩展,例如 Tampermonkey 篡改猴,才能安装此脚本。

您需要先安装一款用户脚本管理器扩展,例如 Tampermonkey 篡改猴Violentmonkey 暴力猴,才能安装此脚本。

您需要先安装一款用户脚本管理器扩展,例如 Tampermonkey 篡改猴Userscripts ,才能安装此脚本。

您需要先安装一款用户脚本管理器扩展,例如 Tampermonkey 篡改猴,才能安装此脚本。

您需要先安装一款用户脚本管理器扩展后才能安装此脚本。

(我已经安装了用户脚本管理器,让我安装!)

您需要先安装一款用户样式管理器扩展,比如 Stylus,才能安装此样式。

您需要先安装一款用户样式管理器扩展,比如 Stylus,才能安装此样式。

您需要先安装一款用户样式管理器扩展,比如 Stylus,才能安装此样式。

您需要先安装一款用户样式管理器扩展后才能安装此样式。

您需要先安装一款用户样式管理器扩展后才能安装此样式。

您需要先安装一款用户样式管理器扩展后才能安装此样式。

(我已经安装了用户样式管理器,让我安装!)

// ==UserScript==
// @name        DuckDuckGo URL Collector (Rewritten)
// @namespace   http://tampermonkey.net/
// @version     1.4
// @description Collects URLs from DuckDuckGo with optional site: filtering and rate limiting, updated for modern DuckDuckGo
// @author      Ghosty-Tongue
// @match       *://duckduckgo.com/*
// @grant       GM_notification
// @license     MIT
// ==/UserScript==

(function() {
    'use strict';

    const collectedUrls = new Set();
    let isProcessing = false;
    let startTime, timerInterval;
    let targetSite = null;

    const searchFormInput = document.getElementById('search_form_input') ||
                            document.querySelector('form[data-testid="search-form"] input[type="text"]');

    if (searchFormInput && searchFormInput.value.includes('site:')) {
        const match = searchFormInput.value.match(/site:([^\s]+)/);
        if (match) {
            targetSite = match[1].toLowerCase();
        }
    }

    const banner = document.createElement('div');
    Object.assign(banner.style, {
        position: 'fixed',
        top: '90px',
        right: '10px',
        zIndex: '10001',
        backgroundColor: 'rgba(255, 165, 0, 0.9)',
        color: 'white',
        padding: '10px',
        borderRadius: '5px',
        display: 'none'
    });
    document.body.appendChild(banner);

    const style = document.createElement('style');
    style.textContent = `
        @keyframes rgbFlow {
            0% { background-position: 0% 50%; }
            100% { background-position: 100% 50%; }
        }
        .ddg-url-collector-btn.processing {
            animation: rgbFlow 1s linear infinite alternate;
        }
    `;
    document.head.appendChild(style);

    const timerDisplay = document.createElement('div');
    Object.assign(timerDisplay.style, {
        position: 'fixed',
        top: '50px',
        right: '10px',
        zIndex: '10000',
        color: 'white',
        backgroundColor: 'rgba(0,0,0,0.7)',
        padding: '5px 10px',
        borderRadius: '5px',
        fontFamily: 'Arial, sans-serif',
        fontSize: '14px'
    });
    document.body.appendChild(timerDisplay);

    function startTimer() {
        if (timerInterval) clearInterval(timerInterval);
        startTime = Date.now();
        timerInterval = setInterval(updateTimer, 1000);
        timerDisplay.textContent = '0s';
    }

    function updateTimer() {
        const elapsed = Math.floor((Date.now() - startTime) / 1000);
        timerDisplay.textContent = `${elapsed}s`;
    }

    function stopTimer() {
        clearInterval(timerInterval);
        const elapsed = Math.floor((Date.now() - startTime) / 1000);
        timerDisplay.textContent = `${elapsed}s (stopped)`;
    }

    function extractUrls() {
        const results = document.querySelectorAll('article[data-testid="result"]');
        let newUrlsCount = 0;
        results.forEach(result => {
            const link = result.querySelector('a[data-testid="result-title-a"]') ||
                         result.querySelector('a[data-testid="result-extras-url-link"]');

            if (link && link.href) {
                const url = link.href;
                try {
                    const urlDomain = new URL(url).hostname.toLowerCase();
                    if (targetSite) {
                        if (!urlDomain.includes(targetSite)) return;
                    }
                    if (!collectedUrls.has(url)) {
                        collectedUrls.add(url);
                        newUrlsCount++;
                    }
                } catch (e) {
                    console.warn('Invalid URL found:', url, e);
                }
            }
        });
        return newUrlsCount;
    }

    async function clickMoreResults() {
        isProcessing = true;
        btn.classList.add('processing');
        let batchCount = 0;
        let moreResultsButton;

        do {
            if (!isProcessing) break;

            moreResultsButton = document.getElementById('more-results') ||
                                document.querySelector('button[type="button"]#more-results') ||
                                Array.from(document.querySelectorAll('button, a')).find(el =>
                                    el.textContent.includes('More results') && el.offsetHeight > 0
                                );

            if (moreResultsButton) {
                moreResultsButton.click();
                window.scrollTo(0, document.body.scrollHeight);
                await new Promise(resolve => setTimeout(resolve, 3000));
                batchCount += extractUrls();

                banner.textContent = `Collecting... Found ${collectedUrls.size} URLs so far.`;
                banner.style.display = 'block';

                if (batchCount >= 420) {
                    banner.textContent = 'Taking 15s break to avoid limits...';
                    await new Promise(resolve => setTimeout(resolve, 15000));
                    banner.textContent = `Resuming collection. Found ${collectedUrls.size} URLs.`;
                    batchCount = 0;
                }

            } else {
                break;
            }
        } while (moreResultsButton && isProcessing);

        isProcessing = false;
        btn.classList.remove('processing');
        banner.style.display = 'none';

        GM_notification({
            title: 'DuckDuckGo URL Collector',
            text: `Collection ${isProcessing ? 'interrupted' : 'complete'}. Saved ${collectedUrls.size} URLs.`,
            timeout: 5000
        });
        saveUrls();
    }

    function saveUrls() {
        const blob = new Blob([Array.from(collectedUrls).join('\n')], {type: 'text/plain'});
        const url = URL.createObjectURL(blob);
        const a = document.createElement('a');
        a.href = url;
        a.download = `duckduckgo_urls_${new Date().toISOString().slice(0,10)}.txt`;
        document.body.appendChild(a);
        a.click();
        document.body.removeChild(a);
        URL.revokeObjectURL(url);
        stopTimer();
    }

    const btn = document.createElement('button');
    btn.textContent = '🦆';
    btn.classList.add('ddg-url-collector-btn');
    Object.assign(btn.style, {
        position: 'fixed',
        top: '10px',
        right: '10px',
        zIndex: '10000',
        padding: '12px 24px',
        background: 'linear-gradient(90deg, #ff0000, #00ff00, #0000ff, #ff0000)',
        backgroundSize: '300% 100%',
        animation: 'rgbFlow 5s linear infinite',
        color: 'white',
        border: 'none',
        borderRadius: '25px',
        cursor: 'pointer',
        fontFamily: 'Arial, sans-serif',
        fontWeight: 'bold',
        boxShadow: '0 4px 15px rgba(0,0,0,0.2)',
        transition: 'transform 0.2s, box-shadow 0.2s'
    });

    btn.addEventListener('mouseover', () => {
        btn.style.transform = 'scale(1.05)';
        btn.style.boxShadow = '0 6px 20px rgba(0,0,0,0.25)';
    });

    btn.addEventListener('mouseout', () => {
        btn.style.transform = 'scale(1)';
        btn.style.boxShadow = '0 4px 15px rgba(0,0,0,0.2)';
    });

    btn.addEventListener('click', () => {
        if (!isProcessing) {
            collectedUrls.clear();
            startTimer();
            clickMoreResults();
        } else {
            isProcessing = false;
            btn.classList.remove('processing');
            banner.style.display = 'none';
            stopTimer();
            saveUrls();
        }
    });

    document.body.appendChild(btn);

    extractUrls();
})();