@Override
protected void outputCharacters(Element pElement, CharacterRun characterRun, String text) {
Element span = htmlDocumentFacade.document.createElement("span");
pElement.appendChild(span);
StringBuilder style = new StringBuilder();
BlockProperies blockProperies = this.blocksProperies.peek();
Triplet triplet = getCharacterRunTriplet(characterRun);
if (WordToHtmlUtils.isNotEmpty(triplet.fontName)
&& !WordToHtmlUtils.equals(triplet.fontName, blockProperies.pFontName)) {
style.append("font-family:" + triplet.fontName + ";");
}
if (characterRun.getFontSize() / 2 != blockProperies.pFontSize) {
style.append("font-size:" + characterRun.getFontSize() / 2 + "pt;");
}
if (triplet.bold) {
style.append("font-weight:bold;");
}
if (triplet.italic) {
style.append("font-style:italic;");
}
WordToHtmlUtils.addCharactersProperties(characterRun, style);
if (style.length() != 0) htmlDocumentFacade.addStyleClass(span, "s", style.toString());
Text textNode = htmlDocumentFacade.createText(text);
span.appendChild(textNode);
}
protected void processSection(HWPFDocument wordDocument, Section section, int sectionCounter) {
// TODO 解析章节
Element div = htmlDocumentFacade.createBlock();
htmlDocumentFacade.addStyleClass(div, "d", getSectionStyle(section));
// htmlDocumentFacade.body.appendChild( div );
pageContainer.appendChild(div);
processParagraphes(wordDocument, div, section, Integer.MIN_VALUE);
}
/** 转换文档 */
public static String convert(String filePath, String output) {
HtmlDocumentFacade facade = null;
// File docFile = new File(docPath);
try {
facade =
new HtmlDocumentFacade(
DocumentBuilderFactory.newInstance().newDocumentBuilder().newDocument());
String s1 = "background-color:gray";
facade.addStyleClass(facade.body, "body", s1);
} catch (Exception e) {
e.printStackTrace();
}
if (AbstractWordUtils.isEmpty(filePath) || AbstractWordUtils.isEmpty(output)) {
System.err.println("docPath OR output is empty. >>!quit");
return "";
}
System.out.println("Converting " + filePath);
System.out.println("Saving output to " + output);
DocConverter.output =
output
.substring(0, output.lastIndexOf(".html"))
.concat(File.separator)
.concat("images")
.concat(File.separator);
// get fileName
new File(DocConverter.output).mkdirs();
try {
Document doc = null;
if (facade == null) doc = DocConverter.process(new File(filePath));
else {
Document document = facade.getDocument();
Element window = document.createElement("div");
Element center = document.createElement("center");
center.appendChild(window);
facade.addStyleClass(
window,
"window",
"border:2px solid green;width:800px!important; margin:0 auto; background-color:#fff; text-align:left;");
facade.body.appendChild(center);
pageContainer = window;
doc = DocConverter.process(new File(filePath), facade);
}
FileWriter out = new FileWriter(output);
DOMSource domSource = new DOMSource(doc);
StreamResult streamResult = new StreamResult(out);
TransformerFactory tf = TransformerFactory.newInstance();
Transformer serializer = tf.newTransformer();
// TODO set encoding from a command argument
serializer.setOutputProperty(OutputKeys.ENCODING, "UTF-8");
serializer.setOutputProperty(OutputKeys.INDENT, "yes");
serializer.setOutputProperty(OutputKeys.METHOD, "html");
serializer.transform(domSource, streamResult);
out.close();
} catch (Exception e) {
e.printStackTrace();
}
return "";
}
/** 解析table */
protected void processTable(HWPFDocument hwpfDocument, Element flow, Table table) {
Element tableHeader = htmlDocumentFacade.createTableHeader();
Element tableBody = htmlDocumentFacade.createTableBody();
final int[] tableCellEdges = WordToHtmlUtils.buildTableCellEdgesArray(table);
final int tableRows = table.numRows();
int maxColumns = Integer.MIN_VALUE;
for (int r = 0; r < tableRows; r++) {
maxColumns = Math.max(maxColumns, table.getRow(r).numCells());
}
for (int r = 0; r < tableRows; r++) {
TableRow tableRow = table.getRow(r);
Element tableRowElement = htmlDocumentFacade.createTableRow();
StringBuilder tableRowStyle = new StringBuilder();
WordToHtmlUtils.addTableRowProperties(tableRow, tableRowStyle);
// index of current element in tableCellEdges[]
int currentEdgeIndex = 0;
final int rowCells = tableRow.numCells();
for (int c = 0; c < rowCells; c++) {
TableCell tableCell = tableRow.getCell(c);
if (tableCell.isVerticallyMerged() && !tableCell.isFirstVerticallyMerged()) {
currentEdgeIndex += getNumberColumnsSpanned(tableCellEdges, currentEdgeIndex, tableCell);
continue;
}
Element tableCellElement;
if (tableRow.isTableHeader()) {
tableCellElement = htmlDocumentFacade.createTableHeaderCell();
} else {
tableCellElement = htmlDocumentFacade.createTableCell();
}
StringBuilder tableCellStyle = new StringBuilder();
WordToHtmlUtils.addTableCellProperties(
tableRow,
tableCell,
r == 0,
r == tableRows - 1,
c == 0,
c == rowCells - 1,
tableCellStyle);
int colSpan = getNumberColumnsSpanned(tableCellEdges, currentEdgeIndex, tableCell);
currentEdgeIndex += colSpan;
if (colSpan == 0) continue;
if (colSpan != 1) tableCellElement.setAttribute("colspan", String.valueOf(colSpan));
final int rowSpan = getNumberRowsSpanned(table, tableCellEdges, r, c, tableCell);
if (rowSpan > 1) tableCellElement.setAttribute("rowspan", String.valueOf(rowSpan));
processParagraphes(hwpfDocument, tableCellElement, tableCell, table.getTableLevel());
if (!tableCellElement.hasChildNodes()) {
tableCellElement.appendChild(htmlDocumentFacade.createParagraph());
}
if (tableCellStyle.length() > 0)
htmlDocumentFacade.addStyleClass(
tableCellElement, tableCellElement.getTagName(), tableCellStyle.toString());
tableRowElement.appendChild(tableCellElement);
}
if (tableRowStyle.length() > 0)
tableRowElement.setAttribute(
"class", htmlDocumentFacade.getOrCreateCssClass("r", tableRowStyle.toString()));
if (tableRow.isTableHeader()) {
tableHeader.appendChild(tableRowElement);
} else {
tableBody.appendChild(tableRowElement);
}
}
final Element tableElement = htmlDocumentFacade.createTable();
tableElement.setAttribute(
"class",
htmlDocumentFacade.getOrCreateCssClass(
"t", "table-layout:fixed;border-collapse:collapse;border-spacing:0;"));
if (tableHeader.hasChildNodes()) {
tableElement.appendChild(tableHeader);
}
if (tableBody.hasChildNodes()) {
tableElement.appendChild(tableBody);
flow.appendChild(tableElement);
} else {
logger.log(
POILogger.WARN,
"Table without body starting at [",
Integer.valueOf(table.getStartOffset()),
"; ",
Integer.valueOf(table.getEndOffset()),
")");
}
}
protected void processSingleSection(HWPFDocument wordDocument, Section section) {
htmlDocumentFacade.addStyleClass(htmlDocumentFacade.body, "b", getSectionStyle(section));
processParagraphes(wordDocument, htmlDocumentFacade.body, section, Integer.MIN_VALUE);
}
protected void processParagraph(
HWPFDocument hwpfDocument,
Element parentElement,
int currentTableLevel,
Paragraph paragraph,
String bulletText) {
final Element pElement = htmlDocumentFacade.createParagraph();
parentElement.appendChild(pElement);
/*if(itemSymbol)
System.out.println(itemSymbol);*/
if (itemSymbol) {
Element span = htmlDocumentFacade.getDocument().createElement("span");
htmlDocumentFacade.addStyleClass(
span,
"itemSymbol",
"font-size:12.0pt;line-height:150%;font-family:Wingdings;mso-ascii-font-family:Wingdings;mso-hide:none;mso-ansi-language:EN-US;mso-fareast-language:ZH-CN;font-weight:normal;mso-bidi-font-weight:normal;font-style:normal;mso-bidi-font-style:normal;text-underline:windowtext none;text-decoration:none;background:transparent");
span.setTextContent("Ø");
pElement.appendChild(span);
itemSymbol = false;
}
StringBuilder style = new StringBuilder();
WordToHtmlUtils.addParagraphProperties(paragraph, style);
final int charRuns = paragraph.numCharacterRuns();
if (charRuns == 0) {
return;
}
{
final String pFontName;
final int pFontSize;
final CharacterRun characterRun = paragraph.getCharacterRun(0);
if ("".equals(paragraph.text().trim())) {
pElement.setTextContent(String.valueOf(UNICODECHAR_NO_BREAK_SPACE));
}
if (characterRun != null) {
Triplet triplet = getCharacterRunTriplet(characterRun);
pFontSize = characterRun.getFontSize() / 2;
pFontName = triplet.fontName;
WordToHtmlUtils.addFontFamily(pFontName, style);
WordToHtmlUtils.addFontSize(pFontSize, style);
} else {
pFontSize = -1;
pFontName = WordToHtmlUtils.EMPTY;
}
blocksProperies.push(new BlockProperies(pFontName, pFontSize));
}
try {
if (WordToHtmlUtils.isNotEmpty(bulletText)) {
if (bulletText.endsWith("\t")) {
/*
* We don't know how to handle all cases in HTML, but at
* least simplest case shall be handled
*/
final float defaultTab = TWIPS_PER_INCH / 2;
float firstLinePosition =
paragraph.getIndentFromLeft() + paragraph.getFirstLineIndent() + 20; // char have
// some space
float nextStop = (float) (Math.ceil(firstLinePosition / defaultTab) * defaultTab);
final float spanMinWidth = nextStop - firstLinePosition;
Element span = htmlDocumentFacade.getDocument().createElement("span");
htmlDocumentFacade.addStyleClass(
span,
"s",
"display: inline-block; text-indent: 0; min-width: "
+ (spanMinWidth / TWIPS_PER_INCH)
+ "in;");
pElement.appendChild(span);
Text textNode =
htmlDocumentFacade.createText(
bulletText.substring(0, bulletText.length() - 1)
+ UNICODECHAR_ZERO_WIDTH_SPACE
+ UNICODECHAR_NO_BREAK_SPACE);
span.appendChild(textNode);
} else {
Text textNode =
htmlDocumentFacade.createText(bulletText.substring(0, bulletText.length() - 1));
pElement.appendChild(textNode);
}
}
processCharacters(hwpfDocument, currentTableLevel, paragraph, pElement);
} finally {
blocksProperies.pop();
}
if (style.length() > 0) htmlDocumentFacade.addStyleClass(pElement, "p", style.toString());
WordToHtmlUtils.compactSpans(pElement);
return;
}
@SuppressWarnings("deprecation")
protected void processImage(
Element currentBlock, boolean inlined, Picture picture, String imageSourcePath) {
final int aspectRatioX = picture.getHorizontalScalingFactor();
final int aspectRatioY = picture.getVerticalScalingFactor();
StringBuilder style = new StringBuilder();
final float imageWidth;
final float imageHeight;
final float cropTop;
final float cropBottom;
final float cropLeft;
final float cropRight;
if (aspectRatioX > 0) {
imageWidth = picture.getDxaGoal() * aspectRatioX / 1000 / TWIPS_PER_INCH;
cropRight = picture.getDxaCropRight() * aspectRatioX / 1000 / TWIPS_PER_INCH;
cropLeft = picture.getDxaCropLeft() * aspectRatioX / 1000 / TWIPS_PER_INCH;
} else {
imageWidth = picture.getDxaGoal() / TWIPS_PER_INCH;
cropRight = picture.getDxaCropRight() / TWIPS_PER_INCH;
cropLeft = picture.getDxaCropLeft() / TWIPS_PER_INCH;
}
if (aspectRatioY > 0) {
imageHeight = picture.getDyaGoal() * aspectRatioY / 1000 / TWIPS_PER_INCH;
cropTop = picture.getDyaCropTop() * aspectRatioY / 1000 / TWIPS_PER_INCH;
cropBottom = picture.getDyaCropBottom() * aspectRatioY / 1000 / TWIPS_PER_INCH;
} else {
imageHeight = picture.getDyaGoal() / TWIPS_PER_INCH;
cropTop = picture.getDyaCropTop() / TWIPS_PER_INCH;
cropBottom = picture.getDyaCropBottom() / TWIPS_PER_INCH;
}
Element root;
if (cropTop != 0 || cropRight != 0 || cropBottom != 0 || cropLeft != 0) {
float visibleWidth = Math.max(0, imageWidth - cropLeft - cropRight);
float visibleHeight = Math.max(0, imageHeight - cropTop - cropBottom);
root = htmlDocumentFacade.createBlock();
htmlDocumentFacade.addStyleClass(
root,
"d",
"vertical-align:text-bottom;width:"
+ visibleWidth
+ "in;height:"
+ visibleHeight
+ "in;");
// complex
Element inner = htmlDocumentFacade.createBlock();
htmlDocumentFacade.addStyleClass(
inner,
"d",
"position:relative;width:"
+ visibleWidth
+ "in;height:"
+ visibleHeight
+ "in;overflow:hidden;");
root.appendChild(inner);
Element image = htmlDocumentFacade.createImage(imageSourcePath);
htmlDocumentFacade.addStyleClass(
image,
"i",
"position:absolute;left:-"
+ cropLeft
+ ";top:-"
+ cropTop
+ ";width:"
+ imageWidth
+ "in;height:"
+ imageHeight
+ "in;");
inner.appendChild(image);
style.append("overflow:hidden;");
} else {
root = htmlDocumentFacade.createImage(imageSourcePath);
root.setAttribute(
"style",
"width:" + imageWidth + "in;height:" + imageHeight + "in;vertical-align:text-bottom;");
}
currentBlock.appendChild(root);
}