Java 在pdf中通过关键字定位

 

package com.util;
import java.io.File;
import java.io.FileInputStream;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;


import com.itextpdf.awt.geom.Rectangle2D.Float;
import com.itextpdf.text.pdf.PdfDictionary;
import com.itextpdf.text.pdf.PdfName;
import com.itextpdf.text.pdf.PdfReader;
import com.itextpdf.text.pdf.parser.ContentByteUtils;
import com.itextpdf.text.pdf.parser.ImageRenderInfo;
import com.itextpdf.text.pdf.parser.PdfContentStreamProcessor;
import com.itextpdf.text.pdf.parser.RenderListener;
import com.itextpdf.text.pdf.parser.TextRenderInfo;


public class PdfKeywordFinder {


public static void main(String[] args) throws IOException {
File pdfFile = new File("D:\\下载.pdf");
byte[] pdfData = new byte[(int) pdfFile.length()];
FileInputStream inputStream = null;
try {
   inputStream = new FileInputStream(pdfFile);
   inputStream.read(pdfData);
} catch (IOException e) {
   throw e;
} finally {
 if (inputStream != null) {
  try {
    inputStream.close();
   } catch (IOException e) {
  }
}
}


String keyword = "甲方（签章）：";


List positions = findKeywordPostions(pdfData, keyword);


System.out.println("total:" + positions.size());
if (positions != null && positions.size() > 0) {
  for (float[] position : positions) {
   System.out.print("pageNum: " + (int) position[0]);
   System.out.print("\tx: " + position[1]);
   System.out.println("\ty: " + position[2]);
 }
 }
}


/**
* findKeywordPostions
* 
* @param pdfData
* @param keyword
* @return List : float[0]:pageNum float[1]:x float[2]:y
* @throws IOException
*/
public static List findKeywordPostions(byte[] pdfData, String keyword) throws IOException {
List result = new ArrayList<>();
List pdfPageContentPositions = getPdfContentPostionsList(pdfData);


for (PdfPageContentPositions pdfPageContentPosition : pdfPageContentPositions) {
List charPositions = findPositions(keyword, pdfPageContentPosition);
if (charPositions == null || charPositions.size() < 1) {
   continue;
}
   result.addAll(charPositions);
}
   return result;
}


private static List getPdfContentPostionsList(byte[] pdfData) throws IOException {
PdfReader reader = new PdfReader(pdfData);


List result = new ArrayList<>();


int pages = reader.getNumberOfPages();
for (int pageNum = 1; pageNum <= pages; pageNum++) {
float width = reader.getPageSize(pageNum).getWidth();
float height = reader.getPageSize(pageNum).getHeight();


PdfRenderListener pdfRenderListener = new PdfRenderListener(pageNum, width, height);



//解析pdf，定位位置
PdfContentStreamProcessor processor = new PdfContentStreamProcessor(pdfRenderListener);
PdfDictionary pageDic = reader.getPageN(pageNum);
PdfDictionary resourcesDic = pageDic.getAsDict(PdfName.RESOURCES);
try {
    processor.processContent(ContentByteUtils.getContentBytesForPage(reader, pageNum), resourcesDic);
} catch (IOException e) {
   reader.close();
throw e;
}


String content = pdfRenderListener.getContent();
List charPositions = pdfRenderListener.getcharPositions();


List positionsList = new ArrayList<>();
for (CharPosition charPosition : charPositions) {
float[] positions = new float[] { charPosition.getPageNum(), charPosition.getX(), charPosition.getY() };
positionsList.add(positions);
}


PdfPageContentPositions pdfPageContentPositions = new PdfPageContentPositions();
pdfPageContentPositions.setContent(content);
pdfPageContentPositions.setPostions(positionsList);


result.add(pdfPageContentPositions);
}
reader.close();
return result;
}


private static List findPositions(String keyword, PdfPageContentPositions pdfPageContentPositions) {


List result = new ArrayList<>();


String content = pdfPageContentPositions.getContent();
List charPositions = pdfPageContentPositions.getPositions();


for (int pos = 0; pos < content.length();) {
int positionIndex = content.indexOf(keyword, pos);
if (positionIndex == -1) {
break;
}
float[] postions = charPositions.get(positionIndex);
result.add(postions);
pos = positionIndex + 1;
}
return result;
}

 

private static class PdfPageContentPositions {
private String content;
private List positions;


public String getContent() {
return content;
}


public void setContent(String content) {
this.content = content;
}


public List getPositions() {
return positions;
}


public void setPostions(List positions) {
this.positions = positions;
}
}

 

private static class PdfRenderListener implements RenderListener {
private int pageNum;
private float pageWidth;
private float pageHeight;
private StringBuilder contentBuilder = new StringBuilder();
private List charPositions = new ArrayList<>();


public PdfRenderListener(int pageNum, float pageWidth, float pageHeight) {
this.pageNum = pageNum;
this.pageWidth = pageWidth;
this.pageHeight = pageHeight;
}


@Override
public void beginTextBlock() {


}


@Override
public void renderText(TextRenderInfo renderInfo) {
List characterRenderInfos = renderInfo.getCharacterRenderInfos();
for (TextRenderInfo textRenderInfo : characterRenderInfos) {
String word = textRenderInfo.getText();
if (word.length() > 1) {
word = word.substring(word.length() - 1, word.length());
}
Float rectangle = textRenderInfo.getAscentLine().getBoundingRectange();
double x = rectangle.getMinX();
double y = rectangle.getMaxY();


float xPercent = Math.round(x / pageWidth * 10000) / 10000f;
float yPercent = Math.round((1 - y / pageHeight) * 10000) / 10000f;// 淇濈暀鍥涗綅灏忔暟


CharPosition charPosition = new CharPosition(pageNum, xPercent, yPercent);
charPositions.add(charPosition);
contentBuilder.append(word);
}
}


@Override
public void endTextBlock() {


}


@Override
public void renderImage(ImageRenderInfo renderInfo) {


}


public String getContent() {
return contentBuilder.toString();
}


public List getcharPositions() {
return charPositions;
}
}


private static class CharPosition {
private int pageNum = 0;
private float x = 0;
private float y = 0;


public CharPosition(int pageNum, float x, float y) {
this.pageNum = pageNum;
this.x = x;
this.y = y;
}


public int getPageNum() {
return pageNum;
}


public float getX() {
return x;
}


public float getY() {
return y;
}


@Override
public String toString() {
return "[pageNum=" + this.pageNum + ",x=" + this.x + ",y=" + this.y + "]";
}
}
}