Create PDF from Webpage in Python

So, are you here because you need to convert web pages to PDFs and do it fast while being able to maintain page styles and JavaScrtip rendered content?

Recently I faced this task. I quickly understood that using something like pdfkit was not an option for me. It does not give full control over the contents of the page and the way the result PDF would look like. Also, if the page that you want to convert to PDF is rendered via JavaScript you won’t see the same result as you would through your browser. And pdfkit is kinda slow 🤷🏻‍♂️.

My solution includes using Selenium to achieve fast and controllable PDF generation. If that detail does not scare you, let's start.

Prerequisites:
You will need to have Chrome browser installed on your machine, it can be Mac OS, Windows, or Linux.
Also, you need to install webdriver-manager and selenium packages. webdriver-manager takes care of fetching the latest version of chromedriver, so don’t need to worry about it. This approach works for years in production for me with no issues with chromedriver and chrome being incompatible.

pip install webdriver-manager selenium

This approach is based on using Chrome`s built-in print function and invoking it through browser API.

I will provide a full code snippet and explain what is going on:

All the heavy lifting is done by _generate_pdfs function, which

1. Iterates over the URLs
2. Calls _get_pdf_from_url on them. This function uses build-in Chrome devtools and invokes Page.printToPDF on them by calling _send_devtools helper function.
3. Returns PDF file as bytes.

Example use:

pdf_file = PdfGenerator(['https://medium.com']).main()
# save pdf to file
with open('medium_site.pdf', "wb") as outfile:
    outfile.write(pdf_file[0].getbuffer())

After you’ve set everything up, you can start configuring PDF generator for your needs.

1. You can adjust print options by changing print_options class attribute. Check printToPDF Chrome devtools documentation for all the options.
2. You can change chromedriver options within the main . In the example code, I’ve set webdriver_options.add_argument(‘ — headless’) so that chromedriver would open in headless mode(not actually opening the browser)
3. You can change chromedriver properties within _get_pdf_from_url options. For example, you can set the screen resolution self.driver.set_window_size(1920, 1080)
4. You can adjust the timeout between opening the page and creating a PDF. For example, to create a PDF of medium.com, I had to set time.sleep(2) to 1 second, to allow JavaScript to render the page.

You can use _get_pdf_from_url separately by building your code around it, even change options on the fly depending on the URL.

This PDF generator works quite fast, for my use case I can generate 500 PDF reports in under 20 minutes on the production server running Ubuntu.