概述
业务开发中,经常会有HTML页面截图,或打印另存为PDF文件的需求。本文即是HTML页面截图需求的技术调研过程的成文。不想看长篇大论的同学,可以直接看Selenium部分,本人最后也是采取此方案。
html2canvas
直接上代码:
const canvas = function () {
html2canvas($("#chart"), {
onrendered: function (canvas) {
// 将id为“class11”部分的代码转换为canvas
$("#class11").html(canvas);
const type = 'png';
// 将图片转换为png类型的流
const imgData = canvas.toDataURL('png');
const saveFile = function (data, filename) {
const save_link = document.createElementNS('http://www.w3.org/1999/xhtml', 'a');
save_link.href = data;
save_link.download = filename;
const event = document.createEvent('MouseEvents');
event.initMouseEvent('click', true, false, window, 0, 0, 0,
0, 0, false, false, false, false, 0, null);
save_link.dispatchEvent(event);
};
const filename = 'html2canvasSavePic' + (new Date()).getTime() + '.' + type;
// 下载文件
saveFile(imgData, filename);
}
});
};
优缺点
html2canvas可以将HTML代码块进行截取,并生成快照形式的canvas,然后利用html5的下载功能。
优点:前台技术,实现比较容易。
缺点:如果是使用H5,则只能在IE9+的版本上使用。需引用jQuery.js和html2canvas.js。
PhantomJS
PhantomJS下载地址 直接上代码:
/**
* 图片保存目录
*/
private static final String TEMP_PATH = "E:\\tmp\\img";
private static final String BLANK = " ";
/**
* 可执行文件路径
*/
private static final String BIN_PATH = "D:\\phantomjs-2.1.1-windows\\bin\\phantomjs.exe";
/**
* js路径
*/
private static final String JS_PATH = "D:\\phantomjs-2.1.1-windows\\examples\\rasterize.js";
public static void main(String[] args) throws IOException, InterruptedException, SAXException {
String url = "http://www.samr.gov.cn/ggjgs/sjdt/gzdt/202001/t20200115_310511.html";
// String url = "http://www.cqn.com.cn/cj/content/2018-06/12/content_5907410.htm";
printUrlScreen2jpg(url);
}
public static void printUrlScreen2jpg(String url) throws IOException, InterruptedException {
// 图片路径
String imagePath = TEMP_PATH + "/" + System.currentTimeMillis() + ".png";
// Java中使用Runtime和Process类运行外部程序
Process process = Runtime.getRuntime().exec(cmd(imagePath, url));
Thread.sleep(1000);
InputStream inputStream = process.getInputStream();
BufferedReader reader = new BufferedReader(new InputStreamReader(inputStream));
while (reader.readLine() != null) {
close(process, reader);
}
}
/**
* 执行cmd命令
*/
private static String cmd(String imagePath, String url) {
return BIN_PATH + BLANK + JS_PATH + BLANK + url + BLANK + imagePath;
}
/**
* 关闭命令
*/
private static void close(Process process, BufferedReader bufferedReader) throws IOException {
if (bufferedReader != null) {
bufferedReader.close();
}
if (process != null) {
process.destroy();
}
}
优缺点
优点:效率高,截图清晰,最重要的是能够截完整的页面,而不是当前可见页面。
缺点:对于运行环境有小小的要求,需要有PhantomJS可执行环境;很容易解决,算不上缺点。
但是,对于部分网站,如上面代码片段里面的http://www.samr.gov.cn/ggjgs/sjdt/gzdt/202001/t20200115_310511.html,截图失败:
很明显,被拦截。
解决思路:添加User-Agent信息,考虑在rasterize.js
脚本追加,但是看不到相关的代码片段。有待进一步调研。
DjNativeSwing
原生Java实现,几乎不会被网站屏蔽,至少我没碰到过。
Maven引入依赖,第三个是64位操作系统需要添加:
<dependency>
<groupId>com.hynnet</groupId>
<artifactId>DJNativeSwing</artifactId>
<version>1.0.0</version>
</dependency>
<dependency>
<groupId>com.hynnet</groupId>
<artifactId>DJNativeSwing-SWT</artifactId>
<version>1.0.0</version>
</dependency>
<dependency>
<groupId>org.eclipse.swt.org.eclipse.swt.win32.win32.x86_64.4.3.swt</groupId>
<artifactId>org.eclipse.swt.win32.win32.x86_64</artifactId>
<version>4.3</version>
</dependency>
直接上代码:
public class PicSnapshotUtils extends JPanel {
/**
* 行分隔符
*/
private final static String LS = System.getProperty("line.separator", "/n");
/**
* 当网页超出目标大小时 截取
*/
private static final int MAX_WIDTH = 2000;
private static final int MAX_HEIGHT = 1400;
public static void main(String[] args) throws IOException, InterruptedException, SAXException {
String url = "http://www.samr.gov.cn/ggjgs/sjdt/gzdt/202001/t20200115_310511.html";
printUrlScreen2jpg("1122.jpg", url, 1400, 900);
}
/**
* @param file 预生成的图片全路径
* @param url 网页地址
*/
public PicSnapshotUtils(final String file, final String url, final String withResult) {
super(new BorderLayout());
JPanel webBrowserPanel = new JPanel(new BorderLayout());
final JWebBrowser webBrowser = new JWebBrowser((NSOption) null);
webBrowser.setBarsVisible(false);
webBrowser.navigate(url);
webBrowserPanel.add(webBrowser, BorderLayout.CENTER);
add(webBrowserPanel, BorderLayout.CENTER);
JPanel panel = new JPanel(new FlowLayout());
webBrowser.addWebBrowserListener(new WebBrowserAdapter() {
// 监听加载进度
@Override
public void loadingProgressChanged(WebBrowserEvent e) {
// 当加载完毕时
if (e.getWebBrowser().getLoadingProgress() == 100) {
String result = (String) webBrowser.executeJavascriptWithResult(withResult);
if (StrUtil.isEmpty(result)) {
return;
}
int index = result.indexOf(":");
NativeComponent nativeComponent = webBrowser.getNativeComponent();
Dimension originalSize = nativeComponent.getSize();
Dimension imageSize = new Dimension(Integer.parseInt(result.substring(0, index)),
Integer.parseInt(result.substring(index + 1)));
imageSize.width = Math.max(originalSize.width, imageSize.width + 50);
imageSize.height = Math.max(originalSize.height, imageSize.height + 50);
nativeComponent.setSize(imageSize);
BufferedImage image = new BufferedImage(imageSize.width, imageSize.height, BufferedImage.TYPE_INT_RGB);
nativeComponent.paintComponent(image);
nativeComponent.setSize(originalSize);
// 当网页超出目标大小时
if (imageSize.width > MAX_WIDTH || imageSize.height > MAX_HEIGHT) {
// 截图部分图形
// image = image.getSubimage(0, 0, MAX_WIDTH, MAX_HEIGHT);
// 此部分为使用缩略图
int width = image.getWidth(), height = image.getHeight();
AffineTransform tx = new AffineTransform();
tx.scale((double) MAX_WIDTH / width, (double) MAX_HEIGHT / height);
AffineTransformOp op = new AffineTransformOp(tx, AffineTransformOp.TYPE_NEAREST_NEIGHBOR);
image = op.filter(image, null);
}
try {
// 输出图像
ImageIO.write(image, "jpg", new File(file));
} catch (IOException ex) {
ex.printStackTrace();
}
// 退出操作
System.exit(0);
}
}
});
add(panel, BorderLayout.SOUTH);
}
/**
* 以javascript脚本获得网页全屏后大小
*/
private static String getScreenWidthHeight() {
return "var width = 0;" + LS +
"var height = 0;" + LS +
"if(document.documentElement) {" + LS +
" width = Math.max(width, document.documentElement.scrollWidth);" + LS +
" height = Math.max(height, document.documentElement.scrollHeight);" + LS +
"}" + LS +
"if(self.innerWidth) {" + LS +
" width = Math.max(width, self.innerWidth);" + LS +
" height = Math.max(height, self.innerHeight);" + LS +
"}" + LS +
"if(document.body.scrollWidth) {" + LS +
" width = Math.max(width, document.body.scrollWidth);" + LS +
" height = Math.max(height, document.body.scrollHeight);" + LS +
"}" + LS +
"return width + ':' + height;";
}
private static void printUrlScreen2jpg(final String file, final String url, final int width, final int height) {
NativeInterface.open();
SwingUtilities.invokeLater(() -> {
String withResult = "var width = " + width + ";var height = " + height + ";return width +':' + height;";
if (width == 0 || height == 0) {
withResult = getScreenWidthHeight();
}
JFrame frame = new JFrame("网页截图");
frame.getContentPane().add(new PicSnapshotUtils(file, url, withResult), BorderLayout.CENTER);
// TODO:加载指定页面,最大保存为640x480的截图
frame.setSize(640, 480);
// 仅初始化,但不显示
frame.invalidate();
frame.pack();
frame.setVisible(false);
});
NativeInterface.runEventPump();
}
}
优缺点
优点:大概率不会被拦截
缺点:不能截完整的页面,只能截可见(不滑动滚轮)的当前页面。Linux系统兼容性暂未测试。
另外,程序运行时,会输出报错日志,虽然截图是成功的:
Exception "java.lang.ClassNotFoundException: com/intellij/codeInsight/editorActions/FoldingData"while constructing DataFlavor for: application/x-java-jvm-local-objectref; class=com.intellij.codeInsight.editorActions.FoldingData
Exception "java.lang.ClassNotFoundException: com/intellij/codeInsight/editorActions/FoldingData"while constructing DataFlavor for: application/x-java-jvm-local-objectref; class=com.intellij.codeInsight.editorActions.FoldingData
Exception "java.lang.ClassNotFoundException: com/intellij/codeInsight/editorActions/ReferenceData"while constructing DataFlavor for: application/x-java-jvm-local-objectref; class=com.intellij.codeInsight.editorActions.ReferenceData
Exception "java.lang.ClassNotFoundException: com/intellij/codeInsight/editorActions/ReferenceData"while constructing DataFlavor for: application/x-java-jvm-local-objectref; class=com.intellij.codeInsight.editorActions.ReferenceData
Selenium
核心,另起一篇:Java+Selenium实现网页截图
html2image
引用依赖:
<dependency>
<groupId>com.github.xuwei-k</groupId>
<artifactId>html2image</artifactId>
<version>0.1.0</version>
</dependency>
优缺点
html2image是可以识别html标签并将html转换成图片的java项目。
优点:后台转换,故对浏览器的版本基本没有要求。
缺点:对样式的识别不是很好,转换出来的图片比较简单,基本没有可以兼容的样式。
Cssbox
依赖:
<dependency>
<groupId>net.sf.cssbox</groupId>
<artifactId>cssbox</artifactId>
<version>5.0.0</version>
</dependency>
直接给代码:
public static void cssBox() throws IOException, InterruptedException, SAXException {
ImageRenderer render = new ImageRenderer();
render.setWindowSize(new Dimension(800, 1000), false);
FileOutputStream out = new FileOutputStream("cssbox.png");
Thread.sleep(10000);
render.renderURL("http://www.samr.gov.cn/ggjgs/sjdt/gzdt/202001/t20200115_310511.html", out, ImageRenderer.Type.PNG);
}
优缺点
优点:无
缺点:不支持引用的外部js、css;其次,自定义设置的宽度有时候不起作用,代码内部有一个默认的宽度是2400。控制台很多报错:
10:34:46.475 [main] ERROR org.fit.cssbox.layout.ContentImage - Unable to get image from: http://www.samr.gov.cn/images/ico39.png
Robot
和Selenium比较类似。未做过多调研。
Cobra
maven搜索到一个看起来比较像能实现网页截图的依赖,并不是放在maven中央仓库的,如果可以下载jar包也行。依赖jar包下载地址为:Tidalwave,实际上已经打不开,404。另一方面,该jar包久未更新,故而未继续调研下去。
xhtmlrenderer
有依赖启动问题。
结论
Selenium牛逼。
一开始网上找到的Java版本Selenium代码也有截图不全的问题,和Python Selenium API还不是一一对应,考虑Java调用Python脚本。未免不是一种思路,但是太麻烦。
参考
Java实现网页截屏功能-PhantomJSDjNativeSwing网页截图
使用Python+Selenium网页截图,解决截图不全问题
HTML转为图片-xhtmlrenderer
HTML转为图片-Cobra
HTML转为图片-Robot
Cssbox实现HTML转图片