All Downloads are FREE. Search and download functionalities are using the official Maven repository.
Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance.
Project price only 1 $
You can buy this project and download/modify it how often you want.
nl.dedicon.pipeline.braille.step.BrailleToText Maven / Gradle / Ivy
package nl.dedicon.pipeline.braille.step;
import static nl.dedicon.pipeline.braille.step.CapitalMode.letter;
import static nl.dedicon.pipeline.braille.step.CapitalMode.lowercase;
import static nl.dedicon.pipeline.braille.step.CapitalMode.permanent;
import static nl.dedicon.pipeline.braille.step.CapitalMode.word;
import static nl.dedicon.pipeline.braille.step.NumericMode.decimalSign;
import static nl.dedicon.pipeline.braille.step.NumericMode.moneyZeros;
import static nl.dedicon.pipeline.braille.step.NumericMode.no;
import static nl.dedicon.pipeline.braille.step.NumericMode.start;
import static nl.dedicon.pipeline.braille.step.NumericMode.yes;
import org.w3c.dom.Document;
import org.w3c.dom.Node;
/**
* Braille to Text
*
* @author Paul Rambags
*/
public class BrailleToText {
private static final char
BRAILLE_CAPITAL = 0x2828, // '⠨'
BRAILLE_CAPITAL_PERMANENT = 0x2818, // '⠘'
BRAILLE_DIGIT = 0x283C, // '⠼'
BRAILLE_HYPHEN = 0x2824, // '⠤'
BRAILLE_RESTORE = 0x2820, // '⠠'
BRAILLE_SPACE = 0x2800, // '⠀'
SPACE = ' ';
private CapitalMode capitalMode = lowercase;
private NumericMode numericMode = no;
private boolean endsWithHyphen = false;
public void convert (Document document) {
Node body = Utils.getChild(document.getDocumentElement(), "body");
Node main = Utils.getFirstChildWithAttribute(body, "div", "id", "main");
while (main != null) {
Node volume = Utils.getFirstChildWithAttribute(main, "div", "class", "volume");
while (volume != null) {
Node page = Utils.getFirstChildWithAttribute(volume, "div", "class", "page");
while (page != null) {
Node braillePage = Utils.getFirstChildWithAttribute(page, "div", "class", "braille-page");
Node textPage = Utils.getFirstChildWithAttribute(page, "div", "class", "text-page");
Node brailleRow = Utils.getFirstChildWithAttribute(braillePage, "div", "class", "row");
Node textRow = Utils.getFirstChildWithAttribute(textPage, "div", "class", "row");
while (brailleRow != null && textRow != null) {
String braille = brailleRow.getTextContent();
String text = convert(braille);
textRow.setTextContent(text);
brailleRow = Utils.getNextSiblingWithAttribute(brailleRow, "div", "class", "row");
textRow = Utils.getNextSiblingWithAttribute(textRow, "div", "class", "row");
}
page = Utils.getNextSiblingWithAttribute(page, "div", "class", "page");
}
volume = Utils.getNextSiblingWithAttribute(volume, "div", "class", "volume");
}
main = Utils.getNextSiblingWithAttribute(main, "div", "id", "main");
}
}
// @todo Houd rekening met meerdere betekenissen van braille symbolen b.v. (123456) = é of %
private String convert (String braille) {
StringBuilder text = new StringBuilder();
for (char b : braille.toCharArray()) {
// text character
char t;
if (numericMode != no) {
t = Utils.convertBrailleNumeric(b);
} else {
t = Utils.convertBraille(b);
if (capitalMode != lowercase) {
t = String.valueOf(t).toUpperCase().charAt(0);
}
}
text.append(t);
// numeric mode
if (b == BRAILLE_DIGIT) {
numericMode = start;
} else {
switch (numericMode) {
case start:
if (Utils.isBrailleMinus(b) || Utils.isBrailleDigit(b)) {
numericMode = yes;
}
break;
case yes:
if (Utils.isBrailleDecimalSeparator(b)) {
numericMode = decimalSign;
} else if (!Utils.isBrailleDigit(b)) {
numericMode = no;
}
break;
case decimalSign:
if (Utils.isBrailleDigit(b)) {
numericMode = yes;
} else if (Utils.isBrailleMoneyZeros(b)) {
numericMode = moneyZeros;
} else {
numericMode = no;
}
break;
case moneyZeros:
if (!Utils.isBrailleMoneyZeros(b)) {
numericMode = no;
}
break;
case no:
default:
break;
}
}
// capital mode
if (b == BRAILLE_CAPITAL) {
capitalMode = letter;
} else if (b == BRAILLE_CAPITAL_PERMANENT) {
switch(capitalMode) {
case lowercase: capitalMode = word;
case letter: capitalMode = word;
case word: capitalMode = permanent;
case permanent: capitalMode = word;
}
} else if (b == BRAILLE_RESTORE) {
numericMode = no;
capitalMode = lowercase;
} else if (isSpace(b)) {
numericMode = no;
if (capitalMode != permanent) {
capitalMode = lowercase;
}
} else if (capitalMode == letter) {
capitalMode = lowercase;
}
// ends with hyphen
if (b == BRAILLE_HYPHEN) {
endsWithHyphen = true;
} else if (!isSpace(b)) {
endsWithHyphen = false;
}
}
// end of line
numericMode = no;
if (!endsWithHyphen) {
if (capitalMode != permanent) {
capitalMode = lowercase;
}
}
return text.toString();
}
private static boolean isSpace(char b) {
return b == BRAILLE_SPACE || b == SPACE;
}
}
