PDF downloader

download-pdfs.js

const axios = require('axios');
const cheerio = require('cheerio');
const fs = require('fs');
const path = require('path');

async function downloadPDFs(url) {
    try {
        // Get webpage content
        const response = await axios.get(url);
        const $ = cheerio.load(response.data);
        
        // Find all PDF links
        const pdfLinks = new Set();
        $('a').each((i, link) => {
            const href = $(link).attr('href');
            if (href && href.toLowerCase().endsWith('.pdf')) {
                // Handle relative URLs
                const absoluteUrl = new URL(href, url).href;
                pdfLinks.add(absoluteUrl);
            }
        });

        console.log(`Found ${pdfLinks.size} PDF links`);
        
        // Create downloads directory if it doesn't exist
        const downloadDir = 'downloads';
        if (!fs.existsSync(downloadDir)) {
            fs.mkdirSync(downloadDir);
        }

        // Download each PDF
        for (const pdfUrl of pdfLinks) {
            try {
                const filename = path.basename(pdfUrl);
                const filepath = path.join(downloadDir, filename);
                
                console.log(`Downloading: ${filename}`);
                const response = await axios({
                    method: 'get',
                    url: pdfUrl,
                    responseType: 'stream'
                });

                const writer = fs.createWriteStream(filepath);
                response.data.pipe(writer);

                await new Promise((resolve, reject) => {
                    writer.on('finish', resolve);
                    writer.on('error', reject);
                });

                console.log(`Downloaded: ${filename}`);
            } catch (error) {
                console.error(`Failed to download ${pdfUrl}:`, error.message);
            }
        }
        
        console.log('All downloads completed!');
        
    } catch (error) {
        console.error('Error:', error.message);
    }
}

// Check if URL is provided as command line argument
const url = process.argv[2];
if (!url) {
    console.error('Please provide a URL as an argument');
    console.log('Usage: node script.js https://example.com');
    process.exit(1);
}

downloadPDFs(url);

package.json

{
  "dependencies": {
    "axios": "^1.7.8",
    "cheerio": "^1.0.0"
  }
}
 

usage.sh

node download-pdfs.js https://example.com