/**
* Validate TMX Format,and pilot to Body XMLElement
*
* @param vg
* @throws TmxReadException ;
*/
private void validateTmxAndParseHeader(VTDGen vg) throws TmxReadException {
VTDNav vn = vg.getNav();
AutoPilot ap = new AutoPilot(vn);
String rootPath = "/tmx";
vu = new VTDUtils();
try {
vu.bind(vn);
ap.selectXPath(rootPath);
if (ap.evalXPath() == -1) {
throw new TmxReadException(Messages.getString("document.TmxReader.validateTmxFileError"));
}
ap.resetXPath();
ap.selectXPath("/tmx/header");
if (ap.evalXPath() == -1) {
throw new TmxReadException(Messages.getString("document.TmxReader.validateTmxFileError"));
}
int id = vu.getVTDNav().getAttrVal("srclang");
if (id == -1) {
throw new TmxReadException(Messages.getString("document.TmxReader.validateTmxFileError"));
}
header.setSrclang(vu.getVTDNav().toString(id).trim());
if (vu.pilot("/tmx/body") == -1) {
throw new TmxReadException(Messages.getString("document.TmxReader.validateTmxFileError"));
}
// compute total tu number
this.totalTu = vu.getChildElementsCount();
} catch (VTDException e) {
logger.error("", e);
throw new TmxReadException(
Messages.getString("document.TmxReader.parseTmxFileError") + e.getMessage());
} finally {
vg.clear();
}
}
/**
* Requires vn to point to a relation element. From there it iterates over all matches of the ap.
* Puts an empty list for each referenced way.
*/
private List<Tuple2<WayRole, Long>> extractWayRefs(VTDNav vn, Hashtable<Long, List<Long>> ways)
throws NavException, XPathEvalException {
vn.push();
List<Tuple2<WayRole, Long>> refs = new ArrayList<>();
for (int j = MEMBER_WAY_PATH.evalXPath(); j != -1; j = MEMBER_WAY_PATH.evalXPath()) {
long ref = Long.parseLong(vn.toString(vn.getAttrVal("ref")));
String roleAsString = vn.toString(vn.getAttrVal("role")).toLowerCase();
WayRole role;
switch (roleAsString) {
case "inner":
role = WayRole.INNER;
break;
case "outer":
role = WayRole.OUTER;
break;
default:
continue;
}
refs.add(Tuple.tuple(role, ref));
ways.put(ref, Collections.emptyList());
}
MEMBER_WAY_PATH.resetXPath();
vn.pop();
return refs;
}
private void extractReferencedWays(
VTDNav vn, Hashtable<Long, List<Long>> ways, Hashtable<Long, Point2D> nodes)
throws XPathEvalException, NavException {
vn.push();
for (int i = WAY_PATH.evalXPath(); i != -1; i = WAY_PATH.evalXPath()) {
long id = Long.parseLong(vn.toString(vn.getAttrVal("id")));
// By checking that we already referenced the id we can reduce memory pressure
if (ways.containsKey(id)) {
ways.put(id, extractNodeRefs(vn, nodes));
}
}
vn.pop();
}
/**
* 创建所有批注节点
*
* @throws Exception
*/
private void createComments() throws Exception {
String xpath = "";
// 先生成全局批注的定义节点
if (fileCommentsList.size() > 0) {
String fileCommentId = CommonFunction.createUUID();
String fileCommentStr = "<sdl:cmt id=\"" + fileCommentId + "\" />";
xpath = "/xliff/file/header";
outputAP.selectXPath(xpath);
while (outputAP.evalXPath() != -1) {
outputXM.insertBeforeTail(fileCommentStr.getBytes("utf-8"));
commentMap.put(fileCommentId, fileCommentsList);
}
}
// 开始生成Comments节点
if (commentMap.size() == 0) {
return;
}
xpath = "/xliff/doc-info";
outputAP.selectXPath(xpath);
if (outputAP.evalXPath() != -1) {
StringBuffer commentSB = new StringBuffer();
commentSB.append("<cmt-defs>");
for (Entry<String, List<CommentBean>> entry : commentMap.entrySet()) {
String id = entry.getKey();
commentSB.append("<cmt-def id=\"" + id + "\">");
commentSB.append("<Comments xmlns=\"\">");
for (CommentBean bean : entry.getValue()) {
commentSB.append(
"<Comment severity=\""
+ bean.getSeverity()
+ "\" "
+ "user=\""
+ bean.getUser()
+ "\" date=\""
+ bean.getDate()
+ "\" version=\"1.0\">"
+ bean.getCommentText()
+ "</Comment>");
}
commentSB.append("</Comments>");
commentSB.append("</cmt-def>");
}
commentSB.append("</cmt-defs>");
outputXM.insertBeforeTail(commentSB.toString().getBytes("utf-8"));
}
}
private void readTuPropElement(TmxTU tu) throws VTDException {
VTDNav vn = vu.getVTDNav();
vn.push();
AutoPilot ap = new AutoPilot(vn);
ap.selectXPath("./prop");
while (ap.evalXPath() != -1) {
String content = vu.getElementContent();
if (content == null) {
continue;
}
int inx = vn.getAttrVal("type");
String typeValue = inx != -1 ? vn.toString(inx) : null;
if (typeValue == null) {
continue;
}
if (typeValue.equals(TmxContexts.PRE_CONTEXT_NAME)) {
tu.appendContext(TmxContexts.PRE_CONTEXT_NAME, content.trim());
} else if (typeValue.equals(TmxContexts.NEXT_CONTEXT_NAME)) {
tu.appendContext(TmxContexts.NEXT_CONTEXT_NAME, content.trim());
} else if (typeValue.equals("x-Context")) {
// Trados TMX file
String[] contexts = content.split(",");
if (contexts.length == 2) {
tu.appendContext(TmxContexts.PRE_CONTEXT_NAME, contexts[0].trim());
tu.appendContext(TmxContexts.NEXT_CONTEXT_NAME, contexts[1].trim());
}
} else {
TmxProp p = new TmxProp(typeValue, content);
tu.appendProp(p);
}
}
vn.pop();
}
/**
* Requires vn to point to a way element. From there it iterates over all matches of the ap. Puts
* a dummy point into nodes for every node it finds.
*/
private List<Long> extractNodeRefs(VTDNav vn, Hashtable<Long, Point2D> nodes)
throws NavException, XPathEvalException {
vn.push();
List<Long> refs = new ArrayList<>();
for (int j = NODE_REF_PATH.evalXPath(); j != -1; j = NODE_REF_PATH.evalXPath()) {
long ref = Long.parseLong(vn.toString(j + 1));
refs.add(ref);
nodes.put(ref, new Point2D(0, 0));
}
NODE_REF_PATH.resetXPath();
vn.pop();
return refs;
}
private List<List<Long>> extractWaysOfBuildings(VTDNav vn, Hashtable<Long, Point2D> nodes)
throws XPathEvalException, NavException {
vn.push();
List<List<Long>> ways = new ArrayList<>();
for (int i = BUILDING_WAY_PATH.evalXPath(); i != -1; i = BUILDING_WAY_PATH.evalXPath()) {
// The lambda will put in a dummy value for each encountered node,
// so that we know later which nodes we need to parse.
ways.add(extractNodeRefs(vn, nodes));
}
vn.pop();
BUILDING_WAY_PATH.resetXPath();
return ways;
}
private void readTuElementAttribute(TmxTU tu) throws VTDException {
VTDNav vn = vu.getVTDNav();
vn.push();
AutoPilot apAttributes = new AutoPilot(vu.getVTDNav());
apAttributes.selectXPath("@*");
int inx = -1;
while ((inx = apAttributes.evalXPath()) != -1) {
String name = vn.toString(inx);
inx = vn.getAttrVal(name);
String value = inx != -1 ? vn.toString(inx) : "";
// tuid, o-encoding, datatype, usagecount, lastusagedate, creationtool, creationtoolversion,
// creationdate,
// creationid, changedate, segtype, changeid, o-tmf, srclang.
if (name.equals("tuid")) {
tu.setTuId(value);
} else if (name.equals("creationtool")) {
tu.setCreationTool(value);
} else if (name.equals("creationtoolversion")) {
tu.setCreationToolVersion(value);
} else if (name.equals("creationdate")) {
tu.setCreationDate(value);
} else if (name.equals("creationid")) {
tu.setCreationUser(value);
} else if (name.equals("changedate")) {
tu.setChangeDate(value);
} else if (name.equals("changeid")) {
tu.setChangeUser(value);
} else {
tu.appendAttribute(name, value);
}
}
vn.pop();
}
/**
* 获取 tmxfile 中的所有语言
*
* @return
*/
public List<String> getLangs() {
List<String> langs = new LinkedList<String>();
langs.add(LanguageUtils.convertLangCode(header.getSrclang()));
VTDNav vn = vu.getVTDNav();
vn.push();
AutoPilot ap = new AutoPilot(vn);
try {
ap.selectXPath("/tmx/body/tu/tuv");
String lang;
int index = -1;
while (ap.evalXPath() != -1) {
index = vn.getAttrVal("xml:lang");
if (index == -1) {
index = vn.getAttrVal("lang"); // version 1.1
if (index == -1) {
continue;
}
}
lang = LanguageUtils.convertLangCode(vn.toRawString(index));
if (!langs.contains(lang)) {
langs.add(lang);
}
}
} catch (Exception e) {
e.printStackTrace();
}
vn.pop();
return langs;
}
private List<List<Tuple2<WayRole, Long>>> extractWayRefsOfMultipolygons(
VTDNav vn, Hashtable<Long, List<Long>> ways) throws NavException, XPathEvalException {
vn.push();
List<List<Tuple2<WayRole, Long>>> multipolygons = new ArrayList<>();
for (int i = BUILDING_MULTIPOLYGON_PATH.evalXPath();
i != -1;
i = BUILDING_MULTIPOLYGON_PATH.evalXPath()) {
// For an explanation for the lambda see extractWaysOfBuildings
multipolygons.add(extractWayRefs(vn, ways));
}
vn.pop();
BUILDING_WAY_PATH.resetXPath();
return multipolygons;
}
private void extractReferencedNodes(VTDNav vn, Hashtable<Long, Point2D> nodes)
throws XPathEvalException, NavException {
vn.push();
for (int i = NODE_PATH.evalXPath(); i != -1; i = NODE_PATH.evalXPath()) {
long id = Long.parseLong(vn.toString(vn.getAttrVal("id")));
// By checking that we already referenced the id we can reduce memory pressure
if (nodes.containsKey(id)) {
nodes.put(
id,
new Point2D(
Double.parseDouble(vn.toString(vn.getAttrVal("lon"))) * (1 << 10),
Double.parseDouble(vn.toString(vn.getAttrVal("lat"))) * (1 << 10)));
}
}
// Make sure we have all referenced nodes extracted
// assert !nodes.containsValue(DUMMY_POINT);
vn.pop();
}
/**
* 获取 R8 xliff文件的所有批注信息
*
* @param vn
* @param tgtbeBean
*/
private void getNotes(VTDNav vn, TuMrkBean tgtbeBean) throws Exception {
vn.push();
List<CommentBean> segCommentList = new LinkedList<CommentBean>();
AutoPilot ap = new AutoPilot(vn);
String xpath = "./note";
ap.selectXPath(xpath);
int atttIdx = -1;
CommentBean bean;
while (ap.evalXPath() != -1) {
boolean isCurrent = true;
if ((atttIdx = vn.getAttrVal("hs:apply-current")) != -1) {
if ("no".equalsIgnoreCase(vn.toString(atttIdx))) {
isCurrent = false;
}
}
String user = "";
String date = "";
String commentText = "";
// R8 xliff 文件中没有提示级别一属性，故此处皆为供参考
String severity = "Low";
if ((atttIdx = vn.getAttrVal("from")) != -1) {
user = vn.toString(atttIdx);
}
if (vn.getText() != -1) {
String r8NoteText = vn.toString(vn.getText());
if (r8NoteText.indexOf(":") != -1) {
date = r8NoteText.substring(0, r8NoteText.indexOf(":"));
commentText = r8NoteText.substring(r8NoteText.indexOf(":") + 1, r8NoteText.length());
} else {
commentText = r8NoteText;
}
}
bean = new CommentBean(user, date, severity, commentText, true);
if (isCurrent) {
segCommentList.add(new CommentBean(user, date, severity, commentText, true));
} else {
if (!fileCommentsList.contains(bean)) {
fileCommentsList.add(bean);
}
}
}
tgtbeBean.setCommentList(segCommentList);
vn.pop();
}
public TmxReaderEvent read() {
TmxTU tu = null;
try {
if (tuAp.evalXPath() != -1) {
tu = new TmxTU();
readTuElementAttribute(tu);
readTuNoteElement(tu);
readTuPropElement(tu);
readTuTuvElement(tu);
} else {
return new TmxReaderEvent(null, TmxReaderEvent.END_FILE);
}
} catch (VTDException e) {
return new TmxReaderEvent(null, TmxReaderEvent.READ_EXCEPTION);
}
return new TmxReaderEvent(tu, TmxReaderEvent.NORMAL_READ);
}
private void readTuTuvElement(TmxTU tu) throws VTDException {
VTDNav vn = vu.getVTDNav();
vn.push();
AutoPilot ap = new AutoPilot(vn);
ap.selectXPath("./tuv");
// TUV 节点下的Note,Prop节点暂时不处理，所以此处暂时不解析
while (ap.evalXPath() != -1) {
int inx = vn.getAttrVal("xml:lang");
inx = inx == -1 ? vn.getAttrVal("lang") : inx;
String lang = inx != -1 ? vn.toString(inx) : null;
if (lang == null) {
continue;
}
vn.push();
if (vu.pilot("./seg") != -1) {
String fullText = vu.getElementContent().trim();
String pureText = DocUtils.getTmxTbxPureText(vu).trim();
if (fullText == null || pureText == null || fullText.equals("") || pureText.equals("")) {
// fix Bug #2928 by Jason SQLite--导入TMX异常, 导入程序正常退出，但是未完全导入所有内容，此处在continue时应该先调用vn.pop()
vn.pop();
continue;
}
TmxSegement segment = new TmxSegement();
segment.setLangCode(Utils.convertLangCode(lang));
if (tmxFilter == null) segment.setFullText(fullText);
else {
String text = tmxFilter.clearString(fullText);
segment.setFullText(text);
}
segment.setPureText(pureText);
if (lang.equalsIgnoreCase(header.getSrclang())) {
tu.setSource(segment);
} else {
tu.appendSegement(segment);
}
}
vn.pop();
}
vn.pop();
}
private void readTuNoteElement(TmxTU tu) throws VTDException {
VTDNav vn = vu.getVTDNav();
vn.push();
AutoPilot ap = new AutoPilot(vn);
ap.selectXPath("./note");
while (ap.evalXPath() != -1) {
String fragment = vu.getElementFragment();
TmxNote note = new TmxNote();
note.setContent(fragment);
int inx = vn.getAttrVal("xml:lang");
String value = inx != -1 ? vn.toString(inx) : null;
if (value != null) {
note.setXmlLang(value);
}
inx = vn.getAttrVal("o-encoding");
value = inx != -1 ? vn.toString(inx) : null;
if (value != null) {
note.setXmlLang(value);
}
tu.appendNote(note);
}
vn.pop();
}
/**
* 分析xliff文件的每一个 trans-unit 节点
*
* @throws Exception
*/
private void ananysisXlfTU(IProgressMonitor monitor) throws Exception {
if (monitor == null) {
monitor = new NullProgressMonitor();
}
AutoPilot ap = new AutoPilot(xlfVN);
AutoPilot childAP = new AutoPilot(xlfVN);
VTDUtils vu = new VTDUtils(xlfVN);
String xpath = "count(/xliff/file/body//trans-unit)";
ap.selectXPath(xpath);
int totalTuNum = (int) ap.evalXPathToNumber();
if (totalTuNum > 500) {
workInterval = totalTuNum / 500;
}
int matchWorkUnit =
totalTuNum % workInterval == 0
? (totalTuNum / workInterval)
: (totalTuNum / workInterval) + 1;
monitor.beginTask("", matchWorkUnit);
xpath = "/xliff/file/body//trans-unit";
String srcXpath = "./source";
String tgtXpath = "./target";
ap.selectXPath(xpath);
int attrIdx = -1;
// trans-unit的id，对应sdl文件的占位符如%%%1%%% 。
String segId = "";
TuMrkBean srcBean = null;
TuMrkBean tgtBean = null;
int traversalTuIndex = 0;
while (ap.evalXPath() != -1) {
traversalTuIndex++;
if ((attrIdx = xlfVN.getAttrVal("id")) == -1) {
continue;
}
srcBean = new TuMrkBean();
tgtBean = new TuMrkBean();
segId = xlfVN.toString(attrIdx);
// 处理source节点
xlfVN.push();
childAP.selectXPath(srcXpath);
if (childAP.evalXPath() != -1) {
String srcContent = vu.getElementContent();
srcContent = srcContent == null ? "" : srcContent;
srcBean.setContent(srcContent);
srcBean.setSource(true);
}
xlfVN.pop();
// 处理target节点
String status = ""; // 状态，针对target节点，空字符串为未翻译
xlfVN.push();
tgtBean.setSource(false);
String tgtContent = null;
childAP.selectXPath(tgtXpath);
if (childAP.evalXPath() != -1) {
tgtContent = vu.getElementContent();
if ((attrIdx = xlfVN.getAttrVal("state")) != -1) {
status = xlfVN.toString(attrIdx);
}
}
tgtContent = tgtContent == null ? "" : tgtContent;
tgtBean.setContent(tgtContent);
xlfVN.pop();
// 处理批注
getNotes(xlfVN, tgtBean);
// 判断是否处于锁定状态
if ((attrIdx = xlfVN.getAttrVal("translate")) != -1) {
if ("no".equalsIgnoreCase(xlfVN.toString(attrIdx))) {
tgtBean.setLocked(true);
}
}
// 判断是否处于批准状态，若是签发，就没有必要判断了，因为签发了的一定就批准了的
if (!"signed-off".equalsIgnoreCase(status)) {
if ((attrIdx = xlfVN.getAttrVal("approved")) != -1) {
if ("yes".equalsIgnoreCase(xlfVN.toString(attrIdx))) {
status = "approved"; // 批准
}
}
}
tgtBean.setStatus(status);
replaceSegment(segId, srcBean, tgtBean);
monitorWork(monitor, traversalTuIndex, false);
}
monitorWork(monitor, traversalTuIndex, true);
}
/**
* 替换掉骨架文件中的占位符
*
* @param segId
* @param srcBean
* @param tgtbeBean
*/
private void replaceSegment(String segId, TuMrkBean srcBean, TuMrkBean tgtbeBean)
throws Exception {
String segStr = "%%%" + segId + "%%%";
String srcXpath = "/xliff/file/body//trans-unit/seg-source//mrk[text()='" + segStr + "']";
// 先处理源文
outputAP.selectXPath(srcXpath);
if (outputAP.evalXPath() != -1) {
int textIdx = outputVN.getText();
outputXM.updateToken(textIdx, srcBean.getContent().getBytes("utf-8"));
}
// 处理译文
String tgtXpath = "/xliff/file/body//trans-unit/target//mrk[text()='" + segStr + "']";
outputAP.selectXPath(tgtXpath);
if (outputAP.evalXPath() != -1) {
String content = tgtbeBean.getContent();
if (tgtbeBean.getCommentList().size() > 0) {
String uuId = CommonFunction.createUUID();
commentMap.put(uuId, tgtbeBean.getCommentList());
content =
"<mrk mtype=\"x-sdl-comment\" sdl:cid=\""
+ uuId
+ "\">"
+ tgtbeBean.getContent()
+ "</mrk>";
}
int textIdx = outputVN.getText();
outputXM.updateToken(textIdx, content.getBytes("utf-8"));
// 开始处理状态
int attrIdx = -1;
if ((attrIdx = outputVN.getAttrVal("mid")) != -1) {
boolean needLocked = false;
String mid = outputVN.toString(attrIdx);
// 下面根据mid找到对应的sdl:seg节点，这个节点里面存放的有每个文本段的状态
String xpath = "ancestor::trans-unit/sdl:seg-defs/sdl:seg[@id='" + mid + "']";
outputAP.selectXPath(xpath);
if (outputAP.evalXPath() != -1) {
// 先判断是否锁定
if (tgtbeBean.isLocked()) {
if ((attrIdx = outputVN.getAttrVal("locked")) != -1) {
if (!"true".equals(outputVN.toString(attrIdx))) {
outputXM.updateToken(attrIdx, "true");
}
} else {
needLocked = true;
}
} else {
if ((attrIdx = outputVN.getAttrVal("locked")) != -1) {
if ("true".equals(outputVN.toString(attrIdx))) {
outputXM.updateToken(attrIdx, "false");
}
}
}
// 下面根据R8的状态。修改sdl的状态。
String conf = "";
String status = tgtbeBean.getStatus();
if ("new".equals(status)) {
conf = "Draft";
} else if ("translated".equals(status)) {
conf = "Translated";
} else if ("approved".equals(status)) {
conf = "ApprovedTranslation";
} else if ("signed-off".equals(status)) {
conf = "ApprovedSignOff";
}
if ("".equals(conf)) {
if ((attrIdx = outputVN.getAttrVal("conf")) != -1) {
outputXM.updateToken(attrIdx, "");
}
} else {
if ((attrIdx = outputVN.getAttrVal("conf")) != -1) {
if (!conf.equals(outputVN.toString(attrIdx))) {
outputXM.updateToken(attrIdx, conf);
}
} else {
String attributeStr = "";
if (needLocked) {
attributeStr = " locked=\"true\" ";
}
attributeStr += " conf=\"" + conf + "\" ";
outputXM.insertAttribute(attributeStr.getBytes("utf-8"));
needLocked = false;
}
}
if (needLocked) {
outputXM.insertAttribute(" locked=\"true\" ".getBytes("utf-8"));
}
}
}
}
}