最近,我一直在研究一个程序,该程序可以将TeX生成的PDF转换为某种形式的文本,该文本保留一些语义上有意义的样式信息,例如下标和上标。
调试时,PDFTextStripper
类似乎可能发生了非常异常的事情。
这是我的TeXUtil
类,可以完成大部分工作。
import com.google.common.base.CharMatcher;
import org.apache.pdfbox.pdmodel.font.PDFont;
import org.json.JSONException;
import org.json.JSONObject;
import java.io.IOException;
import java.util.Hashtable;
import java.util.Stack;
public class TeXUtil {
private Stack<SSStatus> ssstatus;
private boolean accentMode;
private String fs;
private boolean mathMode;
private SymbolDB db;
private Hashtable<String, String> maccDict;
float endY;//Positions
float endX;
float Y;
int height;//Height
//boolean test;
public TeXUtil() throws IOException {
ssstatus = new Stack<SSStatus>();
fs = "rm";
accentMode = false;//as in the state of being right after an accent
mathMode = false;
db = new SymbolDB();
maccDict = new Hashtable<String, String>();
maccDict.put("\\vec","\\vec");
maccDict.put("\\widehat","\\widehat");
maccDict.put("\\widetilde","\\widetilde");
maccDict.put("\\^","\\hat");
maccDict.put("\\v","\\check");
maccDict.put("\\u","\\breve");
maccDict.put("\\`","\\grave");
maccDict.put("\\~","\\tilde");
maccDict.put("\\=","\\bar");
maccDict.put("\\.","\\dot");
maccDict.put("\\","\\ddot");
maccDict.put("\\'","\\acute");
endY = 0;
endX = 0;
Y = 0;
height = 0;
//test = false;
System.out.println("TeXUtil initialized!");
}
private static String fontShortName(PDFont font) {
String[] segments = font.getName().split("\\+");
return segments[segments.length - 1];
}
private static int fontHeight(PDFont font) {
CharMatcher matcher = CharMatcher.inRange('0', '9');
return Integer.parseInt(matcher.retainFrom(fontShortName(font)));
}
private static String fontClass(PDFont font) {
CharMatcher matcher = CharMatcher.inRange('A', 'Z');
return (matcher.retainFrom(fontShortName(font))).toLowerCase();
}
private String textToTeX(String shortFontName, int code) throws JSONException {
JSONObject info = db.getInfo(shortFontName, code);
return info.getString("value");
}
public String fullTextToTeX(PDFont font, int code, float newEndX, float newY, float newEndY){
String shortFontName = fontClass(font);
try {
JSONObject info = db.getInfo(shortFontName, code);
String teXCode = info.getString("value");
StringBuilder preamble1 = new StringBuilder("");
StringBuilder preamble2 = new StringBuilder("");
StringBuilder postamble = new StringBuilder("");
boolean text = info.getBoolean("text");
boolean math = info.getBoolean("math");
boolean tacc = info.getBoolean("tacc");
boolean macc = info.getBoolean("macc");
String newFont = info.getString("font");
int newHeight = fontHeight(font);
//Font change, rm is seen as having no font
if (!newFont.equals(fs)) {
if (!fs.equals("rm"))
preamble1.insert(0, '}');
if (!newFont.equals("rm")) {
preamble2.append('\\');
preamble2.append(newFont);
preamble2.append('{');
}
preamble1.insert(0, " fs = " + fs + " nFs = " + newFont + "\n");
fs = newFont;
}
if (height == 0) {
//preamble2.append(" Meow! am = " + accentMode + " fs = " + fs + " mm = " + mathMode + "\n");
}
//Subscripts/Superscripts
if (height > newHeight && newEndX > endX) {//New subscript/superscript
if (newEndY < endY) {//New superscript
//ssstatus.push(SSStatus.SUP);
preamble2.insert(0, "^{");
}
else if (newY > Y) {//New subscript
//ssstatus.push(SSStatus.SUB);
preamble2.insert(0, "_{");
}
//else {
// System.out.println("Please investigate the situation: texcode = " + teXCode + "endY = " + endY + " Y=" + Y + " endX=" + endX + " newEndY=" + newEndY + " newY=" + newY + " newEndX= " + newEndX);
//}
}
else if (height < newHeight && height != 0) {
//ssstatus.pop();
preamble1.append('}');
}
height = newHeight;
endX = newEndX;
endY = newEndY;
Y = newY;
//Enter or leave math mode
if (mathMode && !math && !macc) {
mathMode = false;
preamble1.append('$');
}
else if (!mathMode && !text && !tacc) {
mathMode = true;
preamble2.insert(0,'$');
}
//Accents
if (accentMode) {//If accent mode is ever entered we need to leave it at once
postamble.append('}');
accentMode = false;
}
if ((mathMode && macc) || (!mathMode && tacc)) {//Right now assume that anything that can be an accent is an accent
postamble.append('{');
if (mathMode)
teXCode = maccDict.get(teXCode);
accentMode = true;
}
if (teXCode.charAt(0) == '\\')
return preamble1.toString() + preamble2.toString() + teXCode + ' ' + postamble.toString();
else
return preamble1.toString() + preamble2.toString() + teXCode + postamble.toString();
}
catch(JSONException e) {
return "\\" + shortFontName + "{" + code + "}";
}
}
}
这是主要的课程。
import org.apache.pdfbox.pdmodel.PDDocument;
import org.apache.pdfbox.pdmodel.font.PDFont;
import org.apache.pdfbox.text.PDFTextStripper;
import org.apache.pdfbox.text.TextPosition;
import com.google.common.base.CharMatcher;
import org.json.*;
import java.io.File;
import java.io.IOException;
import java.util.List;
import java.util.Stack;
public class TEX2TXT {
public static void main(String args[]) throws IOException {
TeXUtil util = new TeXUtil();
//Loading an existing document
File file = new File("/Users/CatLover/Documents/Tex/Examples/c4.pdf");
PDDocument document = PDDocument.load(file);
//Instantiate PDFTextStripper class
PDFTextStripper pdfStripper = new PDFTextStripper() {
protected void writeString(String text, List<TextPosition> textPositions) throws IOException {
TeXUtil util = new TeXUtil();
StringBuilder builder = new StringBuilder();
for(TextPosition position: textPositions) {
float Y = position.getY();
float endY = position.getEndY();
float endX = position.getEndX();
PDFont font = position.getFont();
int[] codes = position.getCharacterCodes();
for(int code: codes) {
builder.append(util.fullTextToTeX(font, code, endX, Y, endY));
}
}
writeString(builder.toString());
}
};
//Retrieving text from PDF document
String text = pdfStripper.getText(document);
System.out.println(text);
//Closing the document
document.close();
}
真正奇怪的是,每次出现单词之间的任何空白时,都会构造
TeXUtil
,而TeXUtil()
仅应调用一次。我不确定为什么会这样。由于PDF是由LaTeX生成的,因此LaTeX不会在PDF中放置空格字符,而是在字符之间保留空格以隐式表示空格,这可能会影响PDFBox的工作方式。 最佳答案
您正在TeXUtil
子类的PDFTextStripper
方法的第一行中构造一个新的writeString
。如果仅删除该行,它仍应能够引用main方法中定义的util
(尽管取决于所使用的Java版本,您可能必须将其设置为final
)。