问题
I am using pdf-clown with pdfclown-0.2.0-HEAD.jar.I have written below code for highlighting search the keyword in Chinese language pdf file and same code is working fine with english pdf file.
import java.awt.Color;
import java.awt.Desktop;
import java.awt.geom.Rectangle2D;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.UnsupportedEncodingException;
import java.net.URL;
import java.nio.charset.Charset;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Date;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.concurrent.TimeUnit;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import java.io.BufferedInputStream;
import java.io.File;
import org.pdfclown.documents.Page;
import org.pdfclown.documents.contents.ITextString;
import org.pdfclown.documents.contents.TextChar;
import org.pdfclown.documents.contents.colorSpaces.DeviceRGBColor;
import org.pdfclown.documents.interaction.annotations.TextMarkup;
import org.pdfclown.documents.interaction.annotations.TextMarkup.MarkupTypeEnum;
import org.pdfclown.files.SerializationModeEnum;
import org.pdfclown.util.math.Interval;
import org.pdfclown.util.math.geom.Quad;
import org.pdfclown.tools.TextExtractor;
public class pdfclown2 {
private static int count;
public static void main(String[] args) throws IOException {
highlight("ebook.pdf","C:\\Users\\Downloads\\6.pdf");
System.out.println("OK");
}
private static void highlight(String inputPath, String outputPath) throws IOException {
URL url = new URL(inputPath);
InputStream in = new BufferedInputStream(url.openStream());
org.pdfclown.files.File file = null;
try {
file = new org.pdfclown.files.File("C:\\Users\\Desktop\\pdf\\test123.pdf");
Map<String, String> m = new HashMap<String, String>();
m.put("亿元或","hi");
m.put("收入亿来","hi");
System.out.println("map size"+m.size());
long startTime = System.currentTimeMillis();
// 2. Iterating through the document pages...
TextExtractor textExtractor = new TextExtractor(true, true);
for (final Page page : file.getDocument().getPages()) {
Map<Rectangle2D, List<ITextString>> textStrings = textExtractor.extract(page);
for (Map.Entry<String, String> entry : m.entrySet()) {
Pattern pattern;
String serachKey = entry.getKey();
final String translationKeyword = entry.getValue();
/*
if ((serachKey.contains(")") && serachKey.contains("("))
|| (serachKey.contains("(") && !serachKey.contains(")"))
|| (serachKey.contains(")") && !serachKey.contains("(")) || serachKey.contains("?")
|| serachKey.contains("*") || serachKey.contains("+")) {s
pattern = Pattern.compile(Pattern.quote(serachKey), Pattern.CASE_INSENSITIVE);
}
else*/
pattern = Pattern.compile(serachKey, Pattern.CASE_INSENSITIVE);
// 2.1. Extract the page text!
//System.out.println(textStrings.toString().indexOf(entry.getKey()));
// 2.2. Find the text pattern matches!
final Matcher matcher = pattern.matcher(TextExtractor.toString(textStrings));
// 2.3. Highlight the text pattern matches!
textExtractor.filter(textStrings, new TextExtractor.IIntervalFilter() {
public boolean hasNext() {
// System.out.println(matcher.find());
// if(key.getMatchCriteria() == 1){
if (matcher.find()) {
return true;
}
/*
* } else if(key.getMatchCriteria() == 2) { if
* (matcher.hitEnd()) { count++; return true; } }
*/
return false;
}
public Interval<Integer> next() {
return new Interval<Integer>(matcher.start(), matcher.end());
}
public void process(Interval<Integer> interval, ITextString match) {
// Defining the highlight box of the text pattern
// match...
System.out.println(match);
/* List<Quad> highlightQuads = new ArrayList<Quad>();
{
Rectangle2D textBox = null;
for (TextChar textChar : match.getTextChars()) {
Rectangle2D textCharBox = textChar.getBox();
if (textBox == null) {
textBox = (Rectangle2D) textCharBox.clone();
} else {
if (textCharBox.getY() > textBox.getMaxY()) {
highlightQuads.add(Quad.get(textBox));
textBox = (Rectangle2D) textCharBox.clone();
} else {
textBox.add(textCharBox);
}
}
}
textBox.setRect(textBox.getX(), textBox.getY(), textBox.getWidth(), textBox.getHeight());
highlightQuads.add(Quad.get(textBox));
}*/
List<Quad> highlightQuads = new ArrayList<Quad>();
List<TextChar> textChars = match.getTextChars();
Rectangle2D firstRect = textChars.get(0).getBox();
Rectangle2D lastRect = textChars.get(textChars.size()-1).getBox();
Rectangle2D rect = firstRect.createUnion(lastRect);
highlightQuads.add(Quad.get(rect).get(rect));
// subtype can be Highlight, Underline, StrikeOut, Squiggly
new TextMarkup(page, highlightQuads, translationKeyword, MarkupTypeEnum.Highlight);
}
public void remove() {
throw new UnsupportedOperationException();
}
});
}
}
SerializationModeEnum serializationMode = SerializationModeEnum.Standard;
file.save(new java.io.File(outputPath), serializationMode);
System.out.println("file created");
long endTime = System.currentTimeMillis();
System.out.println("seconds take for execution is:"+(endTime-startTime)/1000);
} catch (Exception e) {
e.printStackTrace();
}
finally{
in.close();
}
}
}
Kindly provide your inputs to highlight specific search keyword for non english pdf files.
I am serching the keyword in below text which is in chinese langauage.
普双套习近平修宪普京利用双套车绕开宪法装班要走普京
enter image description here
回答1:
Your PDF Clown version
The PDF Clown version you retrieved here from Tymate's maven repository on github has been pushed there April 23rd, 2015. The final (as of now) check-in to the PDF Clown subversion source code repository TRUNK on sourceforge, on the other hand, is from May 27th, 2015. There actually are some 30 checkins after April 23rd, 2015. Thus, you definitely do not use the most current version of this apparently dead PDF library project.
Using the current 0.2.0 snapshot
I tested your code with the 0.2.0 development version compiled from that trunk and the result indeed is different:
It is better insofar as the highlights have the width of the sought character and are located nearer to the actual character position. There still is a bug, though, as the second and third match highlights are somewhat off.
Fixing the bug
The remaining problem actually is not related to the language of the text. It simply is a bug in the processing of one type of the PDF text drawing commands, so it can be observed in documents with text in arbitrary languages. Due to the fact that these commands nowadays are used very seldom only, though, the bug is hardly ever observed, let alone reported. Your PDF, on the other hand, makes use of that kind of text drawing commands.
The bug is in the ShowText class (package org.pdfclown.documents.contents.objects). At the end of the scan method the text line matrix in the graphics state is updated like this if the ShowText instance actually is a ShowTextToNextLine instance derived from it:
if(textScanner == null)
{
state.setTm(tm);
if(this instanceof ShowTextToNextLine)
{state.setTlm((AffineTransform)tm.clone());}
}
The text line matrix here is set to the text matrix after the move to the next line and the drawing of the text. This is wrong, it must instead be set to text matrix right after the move to the next line before the drawing of the text.
This can be fixed e.g. like this:
if(textScanner == null)
{
state.setTm(tm);
if(this instanceof ShowTextToNextLine)
state.getTlm().concatenate(new AffineTransform(1, 0, 0, 1, 0, -state.getLead()));
}
With this change in place the result looks like this:
来源：https://stackoverflow.com/questions/48395771/pdf-clown-not-highlighting-specific-search-keyword