|WEBTHUMB(8)||Linux System Manager's Manual||WEBTHUMB(8)|
webthumb -- web page snapshot-taker for Linux
Take a snapshot of our home page and reduce it to a thumbnail no larger than 100x100, preserving the original aspect ratio (enter this command as one line, of course):
webthumb http://www.boutell.com/ | pnmscale -xysize 100 100 | pnmtojpeg > thumb.jpg
Version 1.01, 12/31/2003. Version 1.01 adds locking, which forces webthumb tasks to run sequentially so that each task receives the correct screenshot.
You must have the following tools, which are often already installed and/or available as packages for your Linux distribution:
- The Xvfb virtual framebuffer X server. Sometimes not installed by default, but easily available as a package for Red Hat, Debian, et cetera.
- The Mozilla web browser.
- The netpbm image manipulation utilities. Almost always installed; it's easy to build them yourself or install them as a package.
- The Linux operating system. webthumb uses the /proc file system to determine whether processes are still running properly.
WHERE TO GET
Creates a PPM-format image of the first screenful of a web page and
writes this image to standard output. This is done using the Xvfb
virtual framebuffer X server, which provides an environment for the
mozilla web browser. To minimize CPU impact, the Xvfb and mozilla
processes are kept resident in memory and reused by future
webthumb. See the top of the Perl
source for useful settings.
The "chrome" of the browser is displayed. This could be fixed with a set of custom XUL files and the -chrome command line option. If you have run Mozilla or Netscape 4 interactively as the current Unix user before, Mozilla may display dialog boxes telling you about profile upgrades and so forth, which prevent the display of the desired page. This can be addressed by deleting the .mozilla and .netscape folders from the home directory of the current Unix user, if they do not contain files that are important to you. It would be nice if webthumb could set up your Mozilla preferences at first use in an appropriate way.
In an ideal world, mozilla would have a command line option like this:
mozilla -htmltops URL (THIS DOESN'T REALLY EXIST)
In order to produce a high-quality postscript version of a web page without the need for any X server at all, and this could then be piped through ghostscript to produce all manner of great things. But there is no such feature, so webthumb is a handy workaround.