Sometimes you don't want just standard web links in your web presentation but a preview image showing the page you're linking to. That is where this package comes in. It generates – by way of calling the external Chrome browser – an image of the web page a given URL addresses. Those image files are stored locally and may be used as often as you want without additional external network traffic.
You can use Go to install this package for you:
go get github.com/mwat56/screenshot
After that you can import it the usual Go way to use the library.
There are only two functions you have to worry about:
// `SetImageDir` sets the directory to use for storing the generated
// screenshot images.
//
// If `aDirectory` is empty or invalid the system's temp directory is used.
//
// `aDirectory` The directory to store the generated images.
func SetImageDir(aDirectory string) { ... }
This function should be called before any other one to make sure the generated screenshots end up where you want them to be. The default is the system's temp directory (e.g. /tmp under GNU/Linux).
To actually create the screenshot image you'd call:
// CreateImage generates an image of `aURL` and stores it in `ImageDir()`,
// returning the file name of the saved image or an error in case of problems.
//
// `aURL` The address of the web page to process.
func CreateImage(aURL string) (string, error) { … }
The returned string is the name of the generated image file (without its path). If you combine it with the directory returned by ImageDir() you get the complete path/filename to locally access the image.
Generating a screenshot image usually takes between one and five seconds, depending on the actual web-page in question; however, it can take considerably longer. To avoid hanging the program the CreateImage() function uses a timeout of half a minute.
And, finally, not all web-pages can be rendered properly and turned into an image. In case of errors (like network-errors or problem while storing the image file) CreateImage() returns an empty filename and an error.
There are a couple more functions (mostly property GETters and SETters) which you will probably barely need; for details refer to the source code documentation.
The Go library controlling a headless instance of the Chrome browser
is required for this package to work.
Under Linux this browser is usually part of your distribution (as chromium-browser).
To resize the screenshot if required by the ImageHeight()/ImageWidth() values the
must be part of your Go installation (if not, run: go get -u golang.org/x/image/draw).
In the source code's sub-directory app/ there's a demo program (screenshot.go) allowing you to generate a screenshot image of an URL given on the commandline.
To run it call e.g.
#> cd app
#> go build screenshot.go
#> ./screenshot
It will show you all available commandline options e.g.:
Usage: ./screenshot [OPTIONS]
-bc
allow the browser to handle web cookies (default false)
-be
skip sites with Certificate errors (default false)
-bm
let browser emulate a mobile device (default false)
-bs
let browser show scrollbars if available (default false)
-bt int
max. time (seconds) allowed to process a single web page (default 32)
-ia
accept the respective other image format (default true)
-id string
directory for storing the screenshot image (default "/tmp")
-ih int
max. height of the screenshot image (default 768)
-io
overwrite an existing image (default false)
-iq int
quality of the screenshot image (default 75)
-is float
the browser's scale factor for the screenshot image (default 0.00)
-iw int
max. width of the screenshot image (default 896)
-ja string
name of text-file that contains sites better avoiding JavaScript
(default "/home/matthias/devel/Go/src/github.com/mwat56/screenshot/app/hostsavoidjs.list")
-jn string
name of text-file that contains sites needing JavaScript
(default "/home/matthias/devel/Go/src/github.com/mwat56/screenshot/app/hostsneedjs.list")
-jp navigator.platform
Identifier the JavaScript navigator.platform should use (default "Linux x86_64")
-js
allow browser's use of JavaScript (default false)
-ju string
description of the UserAgent the browser should report
(default "Mozilla/5.0 (X11; Linux x86_64; rv:89.0) Gecko/20100101 Firefox/89.0")
-u string
(*required*) the URL for the browser's screenshot
-v verbose (default false)
As noted before you'll only need the -u string option, obviously.
You can use this program to generate screenshot images "by hand" and fiddle with the various commandline options to see what difference it makes if you change them.
Prior to this a few years back I wrote the pageview package which used the external wkhtmltoimage program; and in most cases it worked just fine. However, once in a while wkhtmltoimage produced a segmentation fault (core dumped) – reproducible. For a while I thought I could live with it, but over time it happened more often (i.e. with additional URLs). Fiddling around with various commandline options provided no improvement.
In the end I started to look around, searching for alternative approaches – short of writing my own URL retrieval and rendering system. That's when I found ChromeDP and hence this package came into existence.
Copyright © 2022, 2024 M.Watermann, 10247 Berlin, Germany
All rights reserved
EMail : <support@mwat.de>
This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version.
This software is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
You should have received a copy of the GNU General Public License along with this program. If not, see the GNU General Public License for details.
