• url-to-pdf-api: https://github.com/yale8848/url-to-pdf-api
• puppeteer: 1.6.0

• docker pull yale8848/url-to-pdf-api-docker:v1.1
• mkdir logs && docker run --name url2pdf -d --restart=always -p 9000:9000 -v pwd/logs:/root/logs yale8848/url-to-pdf-api-docker:v1.1

docker run --name url2pdf -d --restart=always -p 9000:9000 -e SINGLE_PROCESS=true -e SHOW_LOG=true -v pwd/logs:/root/logs yale8848/url-to-pdf-api-docker:v1.1

• docker build -t url2pdf .

• docker container localtime is set to 'Asia/Shanghai'
• and some chinese fonts in container
• modify url-to-pdf-api code

  • if page use Mathjax, render page after Mathjax render finish

  • add parse pdf header and footer temp in page (pdf.displayHeaderFooter=true and pdf.margin.*=1cm)

    <pdf-render-header style="display:none">
        <div><img src="b/https://github.com/yale8848/ase64" /></div>
    </pdf-render-header>
    
    <pdf-render-footer style="display:none">
        <style type="text/css"> 
            .footer { 
                font-size: 10px;
                position: relative;
                width: 100%;
                text-align: center;
                top: 5px;
            }
        </style>
        <div class="footer">
            <span class="pageNumber"></span>/<span class="totalPages"></span>
        </div>
    </pdf-render-footer>
    

  • js code start call render

     window._waitRendPdftoStart_ = 0; //call most early
    
    
     window._waitRendPdftoStart_ = 1; //  call start render; timeout 20 s
    
    

• use dumb-init to start docker container to avoid zombie chrome process (https://github.com/GoogleChrome/puppeteer/blob/master/docs/troubleshooting.md#running-puppeteer-in-docker)