I'll take you step by step into how to make a service that takes screenshots of webpages and returns them as an image.
First, let's assume you have google-chrome or chromium-browser installed. Both should work the same way. These browsers have command line options that let you capture a screenshot in headless mode.
chromium-browser --headless
--disable-gpu
--no-sandbox
--screenshot=out.png
--window-size=1280,900
--virtual-time-budget=1000
--user-agent="Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/129.0.0.0 Safari/537.36"
https://cbc.ca/news
Url of the site to screenshot
| --headless | Put it in headless mode |
| --disable-gpu | Using GPU in the background sometimes gives us problems, so we disable it |
| --no-sandbox | The sandbox often gives problems with headless execution as well. |
| --screenshot="<filename>" | You can tell it to capture the screenshot to this file. |
| --window-size=width,height | Set the width and height of the window |
| --virtual-time-budget=<ms> | Wait this many milliseconds before taking the screenshot, to give the site time to execute frontend frameworks. |
| --useragent="<user agent>" | Set a custom user agent, since many sites will not work with the default one used in headless mode. |
| url |
Let's make it into a docker image that we can use. I won't go into the details, since I merely asked AI to do it for me. Synchronize the repo it made here:
git clone github.com/smhanov/screenshot-service cd screenshot-service make build make run
This will build and create your screenshotting service.Then you can get the screenshot of any web site by going to the url:
http://localhost:5000/screenshot?url=https://cbc.ca/news
It also accepts other parameters for the various arguments.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。