一、前言
如果不想網頁上的文章被複制(沒錯,說的就是某點),如果想實作文檔不需要下載下傳下來就能線上預覽檢視(常見于文檔付費下載下傳網站、郵箱附件預覽),該怎麼做?常見的做法就是将他們轉化成圖檔。
以下代碼分别提供基于aspose、pdfbox、spire來實作來實作txt、word、pdf、ppt、word等檔案轉圖檔的需求。
1、aspose
Aspose 是一家緻力于.Net ,Java,SharePoint,JasperReports和SSRS元件的提供商,數十個國家的數千機構都有用過aspose元件,建立、編輯、轉換或渲染 Office、OpenOffice、PDF、圖像、ZIP、CAD、XPS、EPS、PSD 和更多檔案格式。注意aspose是商用元件,未經授權導出檔案裡面都是是水印(尊重版權,遠離破解版)。
需要在項目的pom檔案裡添加如下依賴
<dependency>
<groupId>com.aspose</groupId>
<artifactId>aspose-words</artifactId>
<version>23.1</version>
</dependency>
<dependency>
<groupId>com.aspose</groupId>
<artifactId>aspose-pdf</artifactId>
<version>23.1</version>
</dependency>
<dependency>
<groupId>com.aspose</groupId>
<artifactId>aspose-cells</artifactId>
<version>23.1</version>
</dependency>
<dependency>
<groupId>com.aspose</groupId>
<artifactId>aspose-slides</artifactId>
<version>23.1</version>
</dependency>
2 、poi + pdfbox
因為aspose和spire雖然好用,但是都是是商用元件,是以這裡也提供使用開源庫操作的方式的方式。
POI是Apache軟體基金會用Java編寫的免費開源的跨平台的 Java API,Apache POI提供API給Java程式對Microsoft Office格式檔案讀和寫的功能。
Apache PDFBox是一個開源Java庫,支援PDF文檔的開發和轉換。 使用此庫,您可以開發用于建立,轉換和操作PDF文檔的Java程式。
需要在項目的pom檔案裡添加如下依賴
<dependency>
<groupId>org.apache.pdfbox</groupId>
<artifactId>pdfbox</artifactId>
<version>2.0.4</version>
</dependency>
<dependency>
<groupId>org.apache.poi</groupId>
<artifactId>poi</artifactId>
<version>5.2.0</version>
</dependency>
<dependency>
<groupId>org.apache.poi</groupId>
<artifactId>poi-ooxml</artifactId>
<version>5.2.0</version>
</dependency>
<dependency>
<groupId>org.apache.poi</groupId>
<artifactId>poi-scratchpad</artifactId>
<version>5.2.0</version>
</dependency>
<dependency>
<groupId>org.apache.poi</groupId>
<artifactId>poi-excelant</artifactId>
<version>5.2.0</version>
</dependency>
3 spire
spire一款專業的Office程式設計元件,涵蓋了對Word、Excel、PPT、PDF等檔案的讀寫、編輯、檢視功能。spire提供免費版本,但是存在隻能導出前3頁以及隻能導出前500行的限制,隻要達到其一就會觸發限制。需要超出前3頁以及隻能導出前500行的限制的這需要購買付費版(尊重版權,遠離破解版)。這裡使用免費版進行示範。
spire在添加pom之前還得先添加maven倉庫來源
<repository>
<id>com.e-iceblue</id>
<name>e-iceblue</name>
<url>https://repo.e-iceblue.cn/repository/maven-public/</url>
</repository>
接着在項目的pom檔案裡添加如下依賴
免費版:
<dependency>
<groupId>e-iceblue</groupId>
<artifactId>spire.office.free</artifactId>
<version>5.3.1</version>
</dependency>
付費版版:
<dependency>
<groupId>e-iceblue</groupId>
<artifactId>spire.office</artifactId>
<version>5.3.1</version>
</dependency>
二、将檔案轉換成圖檔,并生成到本地
1、将word檔案轉成圖檔
(1)使用aspose
public static void wordToImage(String wordPath, String imagePath) throws Exception {
Document doc = new Document(wordPath);
File file = new File(wordPath);
String filename = file.getName();
String pathPre = imagePath + File.separator + filename.substring(0, filename.lastIndexOf("."));
for (int i = 0; i < doc.getPageCount(); i++) {
Document extractedPage = doc.extractPages(i, 1);
String path = pathPre + (i + 1) + ".png";
extractedPage.save(path, SaveFormat.PNG);
}
}
驗證結果:
(2)使用pdfbox
word轉圖檔沒找到特别好的免費方案,隻能先轉pdf,再轉圖檔。。。
public void wordToImage(String wordPath, String imagePath) throws Exception {
imagePath = FileUtil.getNewFileFullPath(wordPath, imagePath, "png");
try(FileInputStream fileInputStream = new FileInputStream(wordPath);
ByteArrayOutputStream byteArrayOutputStream = new ByteArrayOutputStream()){
XWPFDocument document = new XWPFDocument(fileInputStream);
PdfOptions pdfOptions = PdfOptions.create();
PdfConverter.getInstance().convert(document, byteArrayOutputStream, pdfOptions);
document.close();
ByteArrayInputStream byteArrayInputStream = new ByteArrayInputStream(byteArrayOutputStream.toByteArray());
PDDocument doc = PDDocument.load(byteArrayInputStream);
PDFRenderer renderer = new PDFRenderer(doc);
for (int i = 0; i < doc.getNumberOfPages(); i++) {
BufferedImage image = renderer.renderImageWithDPI(i, 144); // Windows native DPI
String pathname = imagePath + (i + 1) + ".png";
ImageIO.write(image, "PNG", new File(pathname));
}
doc.close();
}
}
驗證結果:
(3)使用spire
public void wordToImage(String wordPath, String imagePath) throws Exception {
File file = new File(wordPath);
String filename = file.getName();
String pathPre = imagePath + File.separator + filename.substring(0, filename.lastIndexOf("."));
//加載Word文檔
Document document = new Document();
document.loadFromFile(wordPath);
//将Word文檔轉換為圖檔
BufferedImage[] images = document.saveToImages(0, document.getPageCount()-1, ImageType.Bitmap);
//儲存圖檔
for (int i = 0; i < images.length; i++) {
String pathname = pathPre + (i + 1) + ".png";
ImageIO.write(images[i], "PNG", new File(pathname));
}
}
驗證結果:
因為使用的是免費版,是以隻能生成前三頁。。。有超過三頁需求的可以選擇付費版本。
2、将txt檔案轉成圖檔(同word檔案轉成圖檔)
(1)使用aspose
public static void txtToImage(String txtPath, String imagePath) throws Exception {
wordToImage(txtPath, imagePath);
}
驗證:
public static void main(String[] args) throws Exception {
FileConvertUtil.wordToImage("D:\\書籍\\電子書\\其它\\《山海經》異獸圖.doc", "D:\\test\\word");
}
驗證結果:
3、将pdf檔案轉圖檔
(1)使用aspose
public static void pdfToImage(String pdfPath, String imagePath) throws Exception {
File file = new File(pdfPath);
String filename = file.getName();
String pathPre = imagePath + File.separator + filename.substring(0, filename.lastIndexOf("."));
PDDocument doc = PDDocument.load(file);
PDFRenderer renderer = new PDFRenderer(doc);
for (int i = 0; i < doc.getNumberOfPages(); i++) {
BufferedImage image = renderer.renderImageWithDPI(i, 144); // Windows native DPI
String pathname = pathPre + (i + 1) + ".png";
ImageIO.write(image, "PNG", new File(pathname));
}
doc.close();
}
驗證:
public static void main(String[] args) throws Exception {
FileConvertUtil.pdfToImage("D:\\書籍\\電子書\\其它\\自然哲學的數學原理.pdf", "D:\\test\\pdf");
}
驗證結果:
(2)使用pdfbox
public void pdfToImage(String pdfPath, String imagePath) throws Exception {
String pathPre = FileUtil.getNewMultiFileFullPathPre(pdfPath, imagePath);
PDDocument doc = PDDocument.load(new File(pdfPath));
PDFRenderer renderer = new PDFRenderer(doc);
for (int i = 0; i < doc.getNumberOfPages(); i++) {
BufferedImage image = renderer.renderImageWithDPI(i, 144); // Windows native DPI
String pathname = pathPre + (i + 1) + ".png";
ImageIO.write(image, "PNG", new File(pathname));
}
doc.close();
}
驗證結果:
[外鍊圖檔轉存失敗,源站可能有防盜鍊機制,建議将圖檔儲存下來直接上傳(img-q7MYkt4t-1688054753134)(D:\文章\預覽\預覽1轉圖檔\pic\pdf轉圖檔pdfbox.png)]
(3)使用spire
public void pdfToImage(String pdfPath, String imagePath) throws Exception {
PdfDocument pdf = new PdfDocument();
pdf.loadFromFile(pdfPath);
File file = new File(pdfPath);
String filename = file.getName();
String pathPre = imagePath + File.separator + filename.substring(0, filename.lastIndexOf("."));
for (int i = 0; i < pdf.getPages().getCount(); i++) {
BufferedImage image = pdf.saveAsImage(i);
String pathname = pathPre + (i + 1) + ".png";
ImageIO.write(image, "png", new File(pathname));
}
}
驗證結果:
因為使用的是免費版,是以隻有前三頁是正常的。。。有超過三頁需求的可以選擇付費版本。
4、将ppt檔案轉圖檔
(1)使用aspose
public void pptToImage(String pptPath, String imagePath) throws Exception {
File file = new File(pptPath);
String filename = file.getName();
String pathPre = imagePath + File.separator + filename.substring(0, filename.lastIndexOf("."));
Presentation presentation = new Presentation(pptPath);
for (int i = 0; i < presentation.getSlides().size(); i++) {
ISlide slide = presentation.getSlides().get_Item(i);
BufferedImage image = slide.getThumbnail(1f, 1f);
String path = pathPre + (i + 1) + ".png";
ImageIO.write(image, "png", new File(path));
}
}
驗證結果:
(2)使用pdfbox
public void pptToImage(String pptPath, String imagePath) throws Exception {
File file = new File(pptPath);
String filename = file.getName().substring(0, file.getName().lastIndexOf("."));
List<BufferedImage> images = pptToBufferedImages(pptPath);
String dicPath = imagePath + File.separator + filename;
File dic = new File(dicPath);
if (!dic.exists()) {
dic.mkdir();
}
for (int i = 0; i < images.size(); i++) {
BufferedImage image = images.get(i);
String path = dicPath+ File.separator + filename + (i + 1) + ".png";
ImageIO.write(image, "png", new File(path));
}
}
驗證結果:
(3)使用spire
驗證結果:
免費版ppt轉圖檔生成前10頁,有進步。。。有超過10頁需求的可以選擇付費版本。
三、利用多線程提升檔案寫入本地的效率
在将牛頓大大的長達669頁的巨作《自然哲學的數學原理》時發現執行時間較長,執行花了140,281ms。但其實這種IO密集型的操作是通過使用多線程的方式來提升效率的,于是針對這點,我又寫了一版多線程的版本。
同步執行導出 自然哲學的數學原理.pdf 耗時:
優化後的代碼如下:
public static void pdfToImageAsync(String pdfPath, String imagePath) throws Exception {
long old = System.currentTimeMillis();
File file = new File(pdfPath);
PDDocument doc = PDDocument.load(file);
PDFRenderer renderer = new PDFRenderer(doc);
int pageCount = doc.getNumberOfPages();
int numCores = Runtime.getRuntime().availableProcessors();
ExecutorService executorService = Executors.newFixedThreadPool(numCores);
for (int i = 0; i < pageCount; i++) {
int finalI = i;
executorService.submit(() -> {
try {
BufferedImage image = renderer.renderImageWithDPI(finalI, 144); // Windows native DPI
String filename = file.getName();
filename = filename.substring(0, filename.lastIndexOf("."));
String pathname = imagePath + File.separator + filename + (finalI + 1) + ".png";
ImageIO.write(image, "PNG", new File(pathname));
} catch (Exception ex) {
ex.printStackTrace();
}
});
}
executorService.shutdown();
executorService.awaitTermination(Long.MAX_VALUE, TimeUnit.NANOSECONDS);
doc.close();
long now = System.currentTimeMillis();
System.out.println("pdfToImage 多線程 轉換完成..用時:" + (now - old) + "ms");
}
多線程執行導出 自然哲學的數學原理.pdf 耗時如下:
從上圖可以看到本次執行隻花了24045ms,隻花了原先差不多六分之一的時間,極大地提升了執行效率。除了pdf,word、txt轉圖檔也可以做這樣的多線程改造:
//将word轉成圖檔(多線程)
public static void wordToImageAsync(String wordPath, String imagePath) throws Exception {
Document doc = new Document(wordPath);
File file = new File(wordPath);
String filename = file.getName();
String pathPre = imagePath + File.separator + filename.substring(0, filename.lastIndexOf("."));
int numCores = Runtime.getRuntime().availableProcessors();
ExecutorService executorService = Executors.newFixedThreadPool(numCores);
for (int i = 0; i < doc.getPageCount(); i++) {
int finalI = i;
executorService.submit(() -> {
try {
Document extractedPage = doc.extractPages(finalI, 1);
String path = pathPre + (finalI + 1) + ".png";
extractedPage.save(path, SaveFormat.PNG);
} catch (Exception ex) {
ex.printStackTrace();
}
});
}
}
//将txt轉成圖檔(多線程)
public static void txtToImageAsync(String txtPath, String imagePath) throws Exception {
wordToImageAsync(txtPath, imagePath);
}
四、将檔案轉換成圖檔流
有的時候我們轉成圖檔後并不需要在本地生成圖檔,而是需要将圖檔傳回或者上傳到圖檔伺服器,這時候就需要将轉換後的圖檔轉成流傳回以友善進行傳輸,代碼示例如下:
1、将word檔案轉成圖檔流
(1)使用aspose
public static List<byte[]> wordToImageStream(String wordPath) throws Exception {
Document doc = new Document(wordPath);
List<byte[]> list = new ArrayList<>();
for (int i = 0; i < doc.getPageCount(); i++) {
try(ByteArrayOutputStream outputStream = new ByteArrayOutputStream()){
Document extractedPage = doc.extractPages(i, 1);
extractedPage.save(outputStream, SaveFormat.*PNG*);
list.add(outputStream.toByteArray());
}
}
return list;
}
(2)使用pdfbox
public List<byte[]> wordToImageStream(String wordPath) throws Exception {
List<BufferedImage> images = new ArrayList<>();
try(FileInputStream fileInputStream = new FileInputStream(wordPath);
ByteArrayOutputStream byteArrayOutputStream = new ByteArrayOutputStream()){
XWPFDocument document = new XWPFDocument(fileInputStream);
PdfOptions pdfOptions = PdfOptions.create();
PdfConverter.getInstance().convert(document, byteArrayOutputStream, pdfOptions);
document.close();
ByteArrayInputStream byteArrayInputStream = new ByteArrayInputStream(byteArrayOutputStream.toByteArray());
PDDocument doc = PDDocument.load(byteArrayInputStream);
PDFRenderer renderer = new PDFRenderer(doc);
for (int i = 0; i < doc.getNumberOfPages(); i++) {
BufferedImage image = renderer.renderImageWithDPI(i, 144); // Windows native DPI
images.add(image);
}
doc.close();
}
return images.stream().map(image-> {
try {
return FileUtil.imageToByte(image);
} catch (IOException e) {
throw new RuntimeException(e);
}
}).collect(Collectors.toList());
}
(3)使用spire
public List<byte[]> wordToImageStream(String wordPath) throws Exception {
Document document = new Document();
document.loadFromFile(wordPath);
BufferedImage[] bufferedImages = document.saveToImages(ImageType.Bitmap);
return FileUtil.toByteArrays(bufferedImages);
}
2、将txt檔案轉成圖檔流
(1)使用aspose
public static List<byte[]> txtToImageStream(String txtPath) throws Exception {
return *wordToImagetream*(txtPath);
}
3、将pdf轉成圖檔流
(1)使用aspose
public static List<byte[]> pdfToImageStream(String pdfPath) throws Exception {
File file = new File(pdfPath);
PDDocument doc = PDDocument.*load*(file);
PDFRenderer renderer = new PDFRenderer(doc);
List<byte[]> list = new ArrayList<>();
for (int i = 0; i < doc.getNumberOfPages(); i++) {
try(ByteArrayOutputStream outputStream = new ByteArrayOutputStream()) {
BufferedImage image = renderer.renderImageWithDPI(i, 144); // Windows native DPI
ImageIO.*write*(image, "PNG", outputStream);
list.add(outputStream.toByteArray());
}
}
doc.close();
return list;
}
(2)使用pdfbox
public List<byte[]> pdfToImageStream(String pdfPath) throws Exception {
File file = new File(pdfPath);
PDDocument doc = PDDocument.load(file);
PDFRenderer renderer = new PDFRenderer(doc);
List<byte[]> list = new ArrayList<>();
for (int i = 0; i < doc.getNumberOfPages(); i++) {
try (ByteArrayOutputStream outputStream = new ByteArrayOutputStream()) {
BufferedImage image = renderer.renderImageWithDPI(i, 144); // Windows native DPI
ImageIO.write(image, "PNG", outputStream);
list.add(outputStream.toByteArray());
}
}
doc.close();
return list;
}
(3)使用spire
public List<byte[]> pdfToImageStream(String pdfPath) throws Exception {
PdfDocument pdf = new PdfDocument();
pdf.loadFromFile(pdfPath);
File file = new File(pdfPath);
String filename = file.getName();
List<byte[]> list = new ArrayList<>();
for (int i = 0; i < pdf.getPages().getCount(); i++) {
BufferedImage image = pdf.saveAsImage(i);
list.add(FileUtil.imageToByte(image));
}
return list;
}
4、将ppt檔案轉圖檔流
(1)使用aspose
public List<byte[]> pptToImageStream(String pptPath) throws IOException {
List<byte[]> list = new ArrayList<>();
Presentation presentation = new Presentation(pptPath);
for (int i = 0; i < presentation.getSlides().size(); i++) {
ISlide slide = presentation.getSlides().get_Item(i);
BufferedImage image = slide.getThumbnail(1f, 1f);
byte[] bytes = FileUtil.imageToByte(image);
list.add(bytes);
}
return list;
}
(2)使用pdfbox
public List<byte[]> pptToImageStream(String pptPath) throws IOException {
List<BufferedImage> images = pptToBufferedImages(pptPath);
if(CollectionUtils.isEmpty(images)){
return null;
}
return images.stream().map(image-> {
try {
return FileUtil.imageToByte(image);
} catch (IOException e) {
throw new RuntimeException(e);
}
}).collect(Collectors.toList());
}
(3)使用spire
public List<byte[]> pptToImageStream(String pptPath) throws Exception {
List<byte[]> list = new ArrayList<>();
Presentation presentation = new Presentation();
presentation.loadFromFile(pptPath);
for (int i = 0; i < presentation.getSlides().getCount(); i++) {
BufferedImage image = presentation.getSlides().get(i).saveAsImage();
list.add(FileUtil.imageToByte(image));
}
return list;
}
總結
将檔案轉成圖檔實作預覽的這種方式的優點是:
1、圖檔線上預覽控件比較多,也比較成熟,前端起來比較友善
2、文檔轉成圖檔後能有效減少文檔内容被複制的情況
3、浏覽器也天然支援
這種方式的缺點是:
1、文檔往往都不隻一頁,所有同城的做法将文檔的每一頁都生成一張圖檔,是以前後端都需要考慮處理多張圖檔的問題
2、如果圖檔都以base64的格式傳回給前端,會造成傳回體過大的問題,如果傳回有加日志還會存在日志體較長,增加日志伺服器的問題。
3、因為base64的格式直接傳回傳回體過長,好一點的做法現将圖檔上傳到圖檔伺服器,隻傳回圖檔的url,這樣解決了圖檔傳回體過長的問題,但要先将多張圖檔先上傳到圖檔伺服器,這樣會不可避免的拖慢接口的傳回速度,尤其是在文檔頁數較多的時候,同時也會增加圖檔伺服器的壓力。
解決多圖檔展示問題的解決方案:
應該如何解決多圖檔展示問題呢,其實很簡單,可以參考開源元件kkfileview解決多圖檔展示問題的(既然都參考了為什麼不直接拿來用[奸笑])的做法,即将生成的多張圖檔全都放到一個html頁面裡,用html保持樣式并實作多張圖檔展示,再将html傳回。
kkfileview展示效果如下:
下圖是kkfileview傳回的html代碼,從html代碼我們可以看到kkfileview其實是将檔案(txt檔案除外)每頁的内容都轉成了圖檔,然後将這些圖檔都嵌入到一個html裡,再傳回給使用者一個html頁面。