Update scraper.js

scraper.js

@@ -1,6 +1,83 @@
 const puppeteer = require('puppeteer-core');
 
-async function scrape(query, mode, orientation) {
+// --- PINTEREST LOGIC ---
+async function scrapePinterest(browser, query) {
+    const page = await browser.newPage();
+    await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36');
+
+    const searchUrl = `https://www.pinterest.com/search/pins/?q=${encodeURIComponent(query)}&rs=typed`;
+    await page.goto(searchUrl, { waitUntil: 'networkidle2', timeout: 30000 });
+
+    // Scroll to find more pins
+    await page.evaluate(async () => {
+        for (let i = 0; i < 3; i++) {
+            window.scrollBy(0, window.innerHeight * 2);
+            await new Promise(r => setTimeout(r, 1500));
+        }
+    });
+
+    const pinUrls = await page.evaluate(() => {
+        return Array.from(document.querySelectorAll('a[href*="/pin/"]'))
+            .map(a => a.href).filter(href => href.includes('/pin/'));
+    });
+
+    const uniquePins = [...new Set(pinUrls)].slice(0, 40); // Check 40 pins
+
+    for (const pinUrl of uniquePins) {
+        const pPage = await browser.newPage();
+        try {
+            await pPage.goto(pinUrl, { waitUntil: 'domcontentloaded', timeout: 8000 });
+            const content = await pPage.content();
+            const videoRegex = /https:\/\/v1\.pinimg\.com\/videos\/mc\/[^\s"']+/g;
+            const matches = content.match(videoRegex);
+
+            if (matches) {
+                let rawUrl = matches[0].replace(/\\u002F/g, '/').replace(/[奖励"']/g, '');
+                const hashMatch = rawUrl.match(/([a-f0-9]{32})/);
+                if (hashMatch) {
+                    const h = hashMatch[1];
+                    // Print 720p URL
+                    process.stdout.write(`https://v1.pinimg.com/videos/mc/720p/${h.substring(0,2)}/${h.substring(2,4)}/${h.substring(4,6)}/${h}.mp4\n`);
+                }
+            }
+        } catch (e) {} 
+        await pPage.close();
+    }
+}
+
+// --- PEXELS LOGIC (FALLBACK) ---
+async function scrapePexels(browser, query, orientation) {
+    const page = await browser.newPage();
+    await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36');
+
+    // Convert "Any" to empty, otherwise use portrait/landscape
+    let orientationParam = "";
+    if (orientation.toLowerCase() === "portrait") orientationParam = "&orientation=portrait";
+    if (orientation.toLowerCase() === "landscape") orientationParam = "&orientation=landscape";
+
+    const searchUrl = `https://www.pexels.com/search/videos/${encodeURIComponent(query)}/?size=medium${orientationParam}`;
+    
+    console.error(`Fallback: Searching Pexels for ${orientation}...`);
+    await page.goto(searchUrl, { waitUntil: 'networkidle2', timeout: 30000 });
+
+    // Pexels puts direct MP4 links in <source> tags inside <article>
+    const videoUrls = await page.evaluate(() => {
+        const sources = Array.from(document.querySelectorAll('article video source[type="video/mp4"]'));
+        return sources.map(s => s.src);
+    });
+
+    // Remove duplicates and print
+    const uniqueUrls = [...new Set(videoUrls)];
+    uniqueUrls.forEach(url => process.stdout.write(url + "\n"));
+}
+
+// --- MAIN CONTROLLER ---
+(async () => {
+    const args = process.argv.slice(2);
+    const platform = args[0]; // 'pinterest' or 'pexels'
+    const query = args[1];
+    const orientation = args[2] || "Any";
+
     let browser;
     try {
         browser = await puppeteer.launch({
@@ -8,61 +85,16 @@ async function scrape(query, mode, orientation) {
             headless: 'new',
             args: ['--no-sandbox', '--disable-setuid-sandbox', '--disable-dev-shm-usage']
         });
-        const page = await browser.newPage();
-        await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36');
-
-        if (mode === "pinterest") {
-            const searchUrl = `https://www.pinterest.com/search/pins/?q=${encodeURIComponent(query)}&rs=typed`;
-            await page.goto(searchUrl, { waitUntil: 'networkidle2', timeout: 30000 });
-            await page.evaluate(async () => {
-                window.scrollBy(0, 1500);
-                await new Promise(r => setTimeout(r, 1000));
-            });
-            const pinUrls = await page.evaluate(() => {
-                return Array.from(document.querySelectorAll('a[href*="/pin/"]')).map(a => a.href);
-            });
-            const uniquePins = [...new Set(pinUrls)].slice(0, 30);
-
-            for (const pinUrl of uniquePins) {
-                const pPage = await browser.newPage();
-                try {
-                    await pPage.goto(pinUrl, { waitUntil: 'domcontentloaded', timeout: 10000 });
-                    const content = await pPage.content();
-                    const match = content.match(/https:\/\/v1\.pinimg\.com\/videos\/mc\/[^\s"']+/);
-                    if (match) {
-                        const hashMatch = match[0].match(/([a-f0-9]{32})/);
-                        if (hashMatch) {
-                            const h = hashMatch[1];
-                            console.log(`https://v1.pinimg.com/videos/mc/720p/${h.substring(0,2)}/${h.substring(2,4)}/${h.substring(4,6)}/${h}.mp4`);
-                        }
-                    }
-                } catch (e) {}
-                await pPage.close();
-            }
-        } 
-        else if (mode === "pexels") {
-            // Pexels supports orientation filter directly in the URL!
-            const pexelsOri = orientation.toLowerCase() === "any" ? "" : orientation.toLowerCase();
-            const searchUrl = `https://www.pexels.com/search/videos/${encodeURIComponent(query)}/?orientation=${pexelsOri}`;
-            await page.goto(searchUrl, { waitUntil: 'networkidle2', timeout: 30000 });
-
-            const videoUrls = await page.evaluate(() => {
-                // Find all video source tags on Pexels search page
-                return Array.from(document.querySelectorAll('video source'))
-                    .map(s => s.src)
-                    .filter(src => src.includes('.mp4') || src.includes('video-preview'));
-            });
-            
-            videoUrls.forEach(url => console.log(url));
+
+        if (platform === 'pinterest') {
+            await scrapePinterest(browser, query);
+        } else if (platform === 'pexels') {
+            await scrapePexels(browser, query, orientation);
         }
 
     } catch (error) {
-        process.stderr.write(error.message);
+        console.error(error.message);
     } finally {
         if (browser) await browser.close();
     }
-}
-
-const args = process.argv.slice(2);
-// args[0] = query, args[1] = mode, args[2] = orientation
-scrape(args[0], args[1] || "pinterest", args[2] || "any");
+})();