/* recursively
(i) create new Parts for each thing listed
in the relationships
(ii) add the new Part to the package
(iii) cross the PartName off unusedZipEntries
*/
private void addPartsFromRelationships(Base source, RelationshipsPart rp, ContentTypeManager ctm)
throws Docx4JException {
OpcPackage pkg = source.getPackage();
for (Relationship r : rp.getRelationships().getRelationship()) {
log.debug(
"\n For Relationship Id="
+ r.getId()
+ " Source is "
+ rp.getSourceP().getPartName()
+ ", Target is "
+ r.getTarget()
+ ", type: "
+ r.getType());
// This is usually the first logged comment for
// a part, so start with a line break.
try {
getPart(pkg, rp, r, ctm);
} catch (Exception e) {
throw new Docx4JException("Failed to add parts from relationships", e);
}
}
}
protected static void removeDefinedCustomXmlParts(
WordprocessingMLPackage wmlPackage, String xpathStorageItemId) {
List<PartName> partsToRemove = new ArrayList<PartName>();
RelationshipsPart relationshipsPart = wmlPackage.getMainDocumentPart().getRelationshipsPart();
List<Relationship> relationshipsList =
((relationshipsPart != null) && (relationshipsPart.getRelationships() != null)
? relationshipsPart.getRelationships().getRelationship()
: null);
Part part = null;
CustomXmlDataStoragePart dataPart = null;
if (relationshipsList != null) {
for (Relationship relationship : relationshipsList) {
if (Namespaces.CUSTOM_XML_DATA_STORAGE.equals(relationship.getType())) {
part = relationshipsPart.getPart(relationship);
if (IsPartToRemove(part, xpathStorageItemId)) {
partsToRemove.add(part.getPartName());
}
}
}
}
if (!partsToRemove.isEmpty()) {
for (int i = 0; i < partsToRemove.size(); i++) {
relationshipsPart.removePart(partsToRemove.get(i));
}
}
}
private OpcPackage process(HashMap<String, ByteArray> partByteArrays) throws Docx4JException {
long startTime = System.currentTimeMillis();
// 2. Create a new Package
// Eventually, you'll also be able to create an Excel package etc
// but only the WordML package exists at present
ContentTypeManager ctm = new ContentTypeManager();
try {
InputStream is = getInputStreamFromZippedPart(partByteArrays, "[Content_Types].xml");
ctm.parseContentTypesFile(is);
} catch (IOException e) {
throw new Docx4JException("Couldn't get [Content_Types].xml from ZipFile", e);
} catch (NullPointerException e) {
throw new Docx4JException("Couldn't get [Content_Types].xml from ZipFile", e);
}
// .. now find the name of the main part
String partName = "_rels/.rels";
RelationshipsPart rp = getRelationshipsPartFromZip(null, partByteArrays, partName);
if (rp == null) {
throw new Docx4JException("_rels/.rels appears to be missing from this package!");
}
String mainPartName = PackageRelsUtil.getNameOfMainPart(rp);
String pkgContentType = ctm.getContentType(new PartName("/" + mainPartName));
// 2. Create a new Package; this'll return the appropriate subclass
OpcPackage p = ctm.createPackage(pkgContentType);
log.info("Instantiated package of type " + p.getClass().getName());
p.setRelationships(rp);
rp.setSourceP(p); //
// 5. Now recursively
// (i) create new Parts for each thing listed
// in the relationships
// (ii) add the new Part to the package
// (iii) cross the PartName off unusedZipEntries
addPartsFromRelationships(partByteArrays, p, rp, ctm);
// 6. Check unusedZipEntries is empty
// if (log.isDebugEnabled()) {
// Iterator myVeryOwnIterator = unusedZipEntries.keySet().iterator();
// while(myVeryOwnIterator.hasNext()) {
// String key = (String)myVeryOwnIterator.next();
// log.info( key + " " + unusedZipEntries.get(key));
// }
// }
registerCustomXmlDataStorageParts(p);
long endTime = System.currentTimeMillis();
log.info("package read; elapsed time: " + Math.round((endTime - startTime)) + " ms");
return p;
}
/**
* Traverse the document, and return a map of all styles which are used directly in the document.
* (IE this does not include styles on which others are just BasedOn).
*
* @return
*/
public Set<String> getStylesInUse() {
org.docx4j.wml.Document wmlDocumentEl = (org.docx4j.wml.Document) this.getJaxbElement();
Body body = wmlDocumentEl.getBody();
List<Object> bodyChildren = body.getContent();
Set<String> stylesInUse = new HashSet<String>();
FontAndStyleFinder finder = new FontAndStyleFinder(null, null, stylesInUse);
finder.defaultCharacterStyle = this.getStyleDefinitionsPart().getDefaultCharacterStyle();
finder.defaultParagraphStyle = this.getStyleDefinitionsPart().getDefaultParagraphStyle();
new TraversalUtil(bodyChildren, finder);
finder.finish();
// Styles in headers, footers?
RelationshipsPart rp = this.getRelationshipsPart();
if (rp != null) {
for (Relationship r : rp.getRelationships().getRelationship()) {
Part part = rp.getPart(r);
if (part instanceof FooterPart) {
Ftr ftr = ((FooterPart) part).getJaxbElement();
finder.walkJAXBElements(ftr);
} else if (part instanceof HeaderPart) {
Hdr hdr = ((HeaderPart) part).getJaxbElement();
finder.walkJAXBElements(hdr);
}
}
}
// Styles in endnotes, footnotes?
if (this.getEndNotesPart() != null) {
log.debug("Looking at endnotes");
CTEndnotes endnotes = this.getEndNotesPart().getJaxbElement();
finder.walkJAXBElements(endnotes);
}
if (this.getFootnotesPart() != null) {
log.debug("Looking at footnotes");
CTFootnotes footnotes = this.getFootnotesPart().getJaxbElement();
finder.walkJAXBElements(footnotes);
}
// Comments
if (this.getCommentsPart() != null) {
log.debug("Looking at comments");
Comments comments = this.getCommentsPart().getJaxbElement();
finder.walkJAXBElements(comments);
}
return stylesInUse;
}
/**
* Create an image part from the provided byte array, attach it to the source part (eg the main
* document part, a header part etc), and return it.
*
* <p>Works for both docx and pptx.
*
* <p>Knowing the MIME type allows you to avoid ImageInfo, but you'll probably also need to know
* the image dimensions
*
* @param opcPackage
* @param sourcePart
* @param bytes
* @param mime MIME type eg image/png
* @return
* @throws Exception @Since 3.0.1
*/
public static BinaryPartAbstractImage createImagePart(
OpcPackage opcPackage, Part sourcePart, byte[] bytes, String mime) throws Exception {
String ext = mimeToExt(mime);
if (mime == null || ext == null) {
log.warn("Null or unknown mime type; image introspection required!");
return createImagePart(opcPackage, sourcePart, bytes);
}
ContentTypeManager ctm = opcPackage.getContentTypeManager();
// Ensure the relationships part exists
if (sourcePart.getRelationshipsPart() == null) {
RelationshipsPart.createRelationshipsPartForPart(sourcePart);
}
String proposedRelId = sourcePart.getRelationshipsPart().getNextId();
BinaryPartAbstractImage imagePart =
(BinaryPartAbstractImage)
ctm.newPartForContentType(
mime, createImageName(opcPackage, sourcePart, proposedRelId, ext), null);
log.debug(
"created part "
+ imagePart.getClass().getName()
+ " with name "
+ imagePart.getPartName().toString());
imagePart.setBinaryData(bytes);
imagePart.rels.add(sourcePart.addTargetPart(imagePart, proposedRelId));
return imagePart;
}
// private void addPartsFromRelationships(ZipFile zf, Base source, RelationshipsPart rp)
private void addPartsFromRelationships(
HashMap<String, ByteArray> partByteArrays,
Base source,
RelationshipsPart rp,
ContentTypeManager ctm)
throws Docx4JException {
OpcPackage pkg = source.getPackage();
// for (Iterator it = rp.iterator(); it.hasNext(); ) {
// Relationship r = (Relationship)it.next();
// log.info("For Relationship Id=" + r.getId() + " Source is "
// + r.getSource().getPartName()
// + ", Target is " + r.getTargetURI() );
// try {
//
// getPart(zf, pkg, rp, r);
//
// } catch (Exception e) {
// throw new Docx4JException("Failed to add parts from relationships", e);
// }
// }
for (Relationship r : rp.getRelationships().getRelationship()) {
log.debug(
"\n For Relationship Id="
+ r.getId()
+ " Source is "
+ rp.getSourceP().getPartName()
+ ", Target is "
+ r.getTarget()
+ ", type: "
+ r.getType());
// This is usually the first logged comment for
// a part, so start with a line break.
try {
getPart(partByteArrays, pkg, rp, r, ctm);
} catch (Exception e) {
throw new Docx4JException("Failed to add parts from relationships", e);
}
}
}
protected static Set<ContentAccessor> getParts(WordprocessingMLPackage srcPackage) {
Set<ContentAccessor> partList = new HashSet<ContentAccessor>();
partList.add(srcPackage.getMainDocumentPart());
// Add headers/footers
RelationshipsPart rp = srcPackage.getMainDocumentPart().getRelationshipsPart();
for (Relationship r : rp.getRelationships().getRelationship()) {
if (r.getType().equals(Namespaces.HEADER)) {
partList.add((HeaderPart) rp.getPart(r));
} else if (r.getType().equals(Namespaces.FOOTER)) {
partList.add((FooterPart) rp.getPart(r));
}
}
return partList;
}
private RelationshipsPart getRelationshipsPartFromZip(Base p, ZipFile zf, String partName)
throws Docx4JException {
// Document contents = null;
// try {
// contents = getDocumentFromZippedPart( zf, partName);
// } catch (Exception e) {
// e.printStackTrace();
// throw new Docx4JException("Error getting document from Zipped Part", e);
//
// }
// // debugPrint(contents);
// // TODO - why don't any of the part names in this document start with "/"?
// return new RelationshipsPart( p, new PartName("/" + partName), contents );
RelationshipsPart rp = null;
InputStream is = null;
try {
is = getInputStreamFromZippedPart(zf, partName);
// thePart = new RelationshipsPart( p, new PartName("/" + partName), is );
rp = new RelationshipsPart(new PartName("/" + partName));
rp.setSourceP(p);
rp.unmarshal(is);
} catch (Exception e) {
e.printStackTrace();
throw new Docx4JException("Error getting document from Zipped Part:" + partName, e);
} finally {
if (is != null) {
try {
is.close();
} catch (IOException exc) {
exc.printStackTrace();
}
}
}
return rp;
// debugPrint(contents);
// TODO - why don't any of the part names in this document start with "/"?
}
/**
* @param out
* @param resolvedPartUri
* @param part
* @throws Docx4JException
* @throws IOException
*/
public void savePart(ZipOutputStream out, Part part) throws Docx4JException, IOException {
// Drop the leading '/'
String resolvedPartUri = part.getPartName().getName().substring(1);
if (handled.get(resolvedPartUri) != null) {
log.debug(".. duplicate save avoided ..");
return;
}
if (part instanceof BinaryPart) {
log.debug(".. saving binary stuff");
saveRawBinaryPart(out, part);
} else {
log.debug(".. saving ");
saveRawXmlPart(out, part);
}
handled.put(resolvedPartUri, resolvedPartUri);
// recurse via this parts relationships, if it has any
RelationshipsPart rrp = part.getRelationshipsPart(false); // don't create
if (rrp != null) {
// log.debug("Found relationships " + rrp.getPartName() );
// Only save it if it actually has rels in it
if (rrp.getRelationships().getRelationship().size() > 0) {
String relPart = PartName.getRelationshipsPartName(resolvedPartUri);
// log.debug("Cf constructed name " + relPart );
saveRawXmlPart(out, rrp, relPart);
addPartsFromRelationships(out, rrp);
}
}
// else {
// log.debug("No relationships for " + resolvedPartUri );
// }
}
protected static boolean IsPartToRemove(Part part, String xpathStorageItemId) {
boolean ret = false;
RelationshipsPart relationshipsPart = part.getRelationshipsPart();
List<Relationship> relationshipsList =
((relationshipsPart != null) && (relationshipsPart.getRelationships() != null)
? relationshipsPart.getRelationships().getRelationship()
: null);
CustomXmlDataStoragePropertiesPart propertiesPart = null;
DatastoreItem datastoreItem = null;
if ((relationshipsList != null) && (!relationshipsList.isEmpty())) {
for (Relationship relationship : relationshipsList) {
if (Namespaces.CUSTOM_XML_DATA_STORAGE_PROPERTIES.equals(relationship.getType())) {
propertiesPart =
(CustomXmlDataStoragePropertiesPart) relationshipsPart.getPart(relationship);
break;
}
}
}
if (propertiesPart != null) {
datastoreItem = propertiesPart.getJaxbElement();
}
if (datastoreItem != null) {
if ((datastoreItem.getItemID() != null) && (datastoreItem.getItemID().length() > 0)) {
ret = datastoreItem.getItemID().equals(xpathStorageItemId);
}
if ((!ret)
&& (datastoreItem.getSchemaRefs() != null)
&& (datastoreItem.getSchemaRefs().getSchemaRef() != null)
&& (!datastoreItem.getSchemaRefs().getSchemaRef().isEmpty())) {
for (SchemaRef ref : datastoreItem.getSchemaRefs().getSchemaRef()) {
if (PART_TO_REMOVE_SCHEMA_TYPES.contains(ref.getUri())) {
ret = true;
break;
}
}
}
}
return ret;
}
private RelationshipsPart getRelationshipsPartFromZip(Base p, String partName)
throws Docx4JException {
RelationshipsPart rp = null;
InputStream is = null;
try {
is = partStore.loadPart(partName);
if (is == null) {
return null; // that's ok
}
rp = new RelationshipsPart(new PartName("/" + partName));
rp.setSourceP(p);
rp.unmarshal(is);
} catch (Exception e) {
e.printStackTrace();
throw new Docx4JException("Error getting document from Zipped Part:" + partName, e);
} finally {
IOUtils.closeQuietly(is);
}
return rp;
}
private PartName getNewPartName(String prefix, String suffix, RelationshipsPart rp)
throws InvalidFormatException {
PartName proposed = null;
int i = 1;
do {
if (i > 1) {
proposed = new PartName(prefix + i + suffix);
} else {
proposed = new PartName(prefix + suffix);
}
i++;
} while (rp.getRel(proposed) != null);
return proposed;
}
public static void visit(
WordprocessingMLPackage wmlPackage, boolean bodyOnly, Callback callback) {
MainDocumentPart mainDocument = null;
RelationshipsPart relPart = null;
List<Relationship> relList = null;
List<Object> elementList = null;
if ((wmlPackage != null) && (callback != null)) {
mainDocument = wmlPackage.getMainDocumentPart();
callback.walkJAXBElements(mainDocument.getJaxbElement().getBody());
if (!bodyOnly) {
relPart = mainDocument.getRelationshipsPart();
relList = relPart.getRelationships().getRelationship();
for (Relationship rs : relList) {
elementList = null;
if (Namespaces.HEADER.equals(rs.getType())) {
elementList = ((HeaderPart) relPart.getPart(rs)).getJaxbElement().getContent();
} else if (Namespaces.FOOTER.equals(rs.getType())) {
elementList = ((FooterPart) relPart.getPart(rs)).getJaxbElement().getContent();
} else if (Namespaces.ENDNOTES.equals(rs.getType())) {
// elementList = ((EndnotesPart) relPart.getPart(rs)).getContent();
elementList = new ArrayList();
elementList.addAll(((EndnotesPart) relPart.getPart(rs)).getJaxbElement().getEndnote());
} else if (Namespaces.FOOTNOTES.equals(rs.getType())) {
// elementList = ((FootnotesPart) relPart.getPart(rs)).getContent();
elementList = new ArrayList();
elementList.addAll(
((FootnotesPart) relPart.getPart(rs)).getJaxbElement().getFootnote());
} else if (Namespaces.COMMENTS.equals(rs.getType())) {
elementList = new ArrayList();
for (Comment comment :
((CommentsPart) relPart.getPart(rs)).getJaxbElement().getComment()) {
elementList.addAll(comment.getEGBlockLevelElts());
}
}
if ((elementList != null) && (!elementList.isEmpty())) {
log.debug("Processing target: " + rs.getTarget() + ", type: " + rs.getType());
callback.walkJAXBElements(elementList);
}
}
}
}
}
/**
* Create an image part from the provided filePath image, attach it to the source part (eg the
* main document part, a header part etc), and return it.
*
* <p>Works for both docx and pptx.
*
* @param opcPackage
* @param sourcePart
* @param filePath
* @return
* @throws Exception
*/
public static BinaryPartAbstractImage createImagePart(
OpcPackage opcPackage, Part sourcePart, File imageFile) throws Exception {
final byte[] locByte = new byte[1];
// We are in the case that image is not load (no byte Array) so isLoad is false
ImageInfo info = ensureFormatIsSupported(imageFile, locByte, false);
ContentTypeManager ctm = opcPackage.getContentTypeManager();
// Ensure the relationships part exists
if (sourcePart.getRelationshipsPart() == null) {
RelationshipsPart.createRelationshipsPartForPart(sourcePart);
}
String proposedRelId = sourcePart.getRelationshipsPart().getNextId();
String ext = info.getMimeType().substring(info.getMimeType().indexOf("/") + 1);
BinaryPartAbstractImage imagePart =
(BinaryPartAbstractImage)
ctm.newPartForContentType(
info.getMimeType(),
createImageName(opcPackage, sourcePart, proposedRelId, ext),
null);
log.debug(
"created part "
+ imagePart.getClass().getName()
+ " with name "
+ imagePart.getPartName().toString());
FileInputStream fis = new FileInputStream(imageFile);
imagePart.setBinaryData(fis);
imagePart.rels.add(sourcePart.addTargetPart(imagePart, proposedRelId));
imagePart.setImageInfo(info);
return imagePart;
}
/**
* Create an image part from the provided byte array, attach it to the source part (eg the main
* document part, a header part etc), and return it.
*
* <p>Works for both docx and pptx.
*
* <p>Note: this method creates a temp file (and attempts to delete it). That's because it uses
* org.apache.xmlgraphics
*
* @param opcPackage
* @param sourcePart
* @param bytes
* @return
* @throws Exception
*/
public static BinaryPartAbstractImage createImagePart(
OpcPackage opcPackage, Part sourcePart, byte[] bytes) throws Exception {
// Whatever image type this is, we're going to need
// to know its dimensions.
// For that we use ImageInfo, which can only
// load an image from a URI.
// So first, write the bytes to a temp file
File tmpImageFile = File.createTempFile("img", ".img");
FileOutputStream fos = new FileOutputStream(tmpImageFile);
fos.write(bytes);
fos.close();
log.debug("created tmp file: " + tmpImageFile.getAbsolutePath());
ImageInfo info = ensureFormatIsSupported(tmpImageFile, bytes, true);
// In the absence of an exception, tmpImageFile now contains an image
// Word will accept
ContentTypeManager ctm = opcPackage.getContentTypeManager();
// Ensure the relationships part exists
if (sourcePart.getRelationshipsPart() == null) {
RelationshipsPart.createRelationshipsPartForPart(sourcePart);
}
String proposedRelId = sourcePart.getRelationshipsPart().getNextId();
String ext = info.getMimeType().substring(info.getMimeType().indexOf("/") + 1);
// System.out.println(ext);
BinaryPartAbstractImage imagePart =
(BinaryPartAbstractImage)
ctm.newPartForContentType(
info.getMimeType(),
createImageName(opcPackage, sourcePart, proposedRelId, ext),
null);
log.debug(
"created part "
+ imagePart.getClass().getName()
+ " with name "
+ imagePart.getPartName().toString());
FileInputStream fis = new FileInputStream(tmpImageFile);
imagePart.setBinaryData(fis);
imagePart.rels.add(sourcePart.addTargetPart(imagePart, proposedRelId));
imagePart.setImageInfo(info);
// Delete the tmp file
// As per http://stackoverflow.com/questions/991489/i-cant-delete-a-file-in-java
// the following 3 lines are necessary, at least on Win 7 x64
// Also reported on Win XP, but in my testing, the files were deleting OK anyway.
fos = null;
fis = null;
if (Docx4jProperties.getProperty(
"docx4j.openpackaging.parts.WordprocessingML.BinaryPartAbstractImage.TempFiles.ForceGC",
true)) {
System.gc();
}
if (tmpImageFile.delete()) {
log.debug(".. deleted " + tmpImageFile.getAbsolutePath());
} else {
log.warn("Couldn't delete tmp file " + tmpImageFile.getAbsolutePath());
tmpImageFile.deleteOnExit();
// If that doesn't work, see "Clean Up Your Mess: Managing Temp Files in Java Apps"
// at devx.com
}
return imagePart;
}
/**
* Get a Part (except a relationships part), and all its related parts. This can be called
* directly from outside the library, in which case the Part will not be owned by a Package until
* the calling code makes it so.
*
* @param zf
* @param source
* @param unusedZipEntries
* @param pkg
* @param r
* @param resolvedPartUri
* @throws Docx4JException
* @throws InvalidFormatException
*/
private void getPart(OpcPackage pkg, RelationshipsPart rp, Relationship r, ContentTypeManager ctm)
throws Docx4JException, InvalidFormatException, URISyntaxException {
Base source = null;
String resolvedPartUri = null;
if (r.getType().equals(Namespaces.HYPERLINK)) {
// Could be Internal or External
// Example of Internal is w:drawing/wp:inline/wp:docPr/a:hlinkClick
log.info("Encountered (but not loading) hyperlink " + r.getTarget());
return;
} else if (r.getTargetMode() == null || !r.getTargetMode().equals("External")) {
// Usual case
source = rp.getSourceP();
resolvedPartUri =
URIHelper.resolvePartUri(rp.getSourceURI(), new URI(r.getTarget())).toString();
// Now drop leading "/'
resolvedPartUri = resolvedPartUri.substring(1);
// Now normalise it .. ie abc/def/../ghi
// becomes abc/ghi
// Maybe this isn't necessary with a zip file,
// - ZipFile class may be smart enough to do it.
// But it is certainly necessary in the JCR case.
// resolvedPartUri = (new java.net.URI(resolvedPartUri)).normalize().toString();
// log.info("Normalised, it is " + resolvedPartUri );
} else {
// EXTERNAL
if (loadExternalTargets && r.getType().equals(Namespaces.IMAGE)) {
// It could instead be, for example, of type hyperlink,
// and we don't want to try to fetch that
log.info("Loading external resource " + r.getTarget() + " of type " + r.getType());
BinaryPart bp = ExternalResourceUtils.getExternalResource(r.getTarget());
pkg.getExternalResources().put(bp.getExternalTarget(), bp);
} else {
log.info(
"Encountered (but not loading) external resource "
+ r.getTarget()
+ " of type "
+ r.getType());
}
return;
}
String relationshipType = r.getType();
Part part;
if (pkg.handled.get(resolvedPartUri) != null) {
// The source Part (or Package) might have a convenience
// method for this
part = pkg.getParts().getParts().get(new PartName("/" + resolvedPartUri));
if (source.setPartShortcut(part, relationshipType)) {
log.debug(
"Convenience method established from "
+ source.getPartName()
+ " to "
+ part.getPartName());
}
return;
}
part = getRawPart(ctm, resolvedPartUri, r); // will throw exception if null
// The source Part (or Package) might have a convenience
// method for this
if (source.setPartShortcut(part, relationshipType)) {
log.debug(
"Convenience method established from "
+ source.getPartName()
+ " to "
+ part.getPartName());
}
if (part instanceof BinaryPart || part instanceof DefaultXmlPart) {
// The constructors of other parts should take care of this...
part.setRelationshipType(relationshipType);
}
rp.loadPart(part, r);
pkg.handled.put(resolvedPartUri, resolvedPartUri);
// unusedZipEntries.put(resolvedPartUri, new Boolean(false));
RelationshipsPart rrp = getRelationshipsPart(part);
if (rrp != null) {
// recurse via this parts relationships, if it has any
addPartsFromRelationships(part, rrp, ctm);
String relPart = PartName.getRelationshipsPartName(part.getPartName().getName().substring(1));
// unusedZipEntries.put(relPart, new Boolean(false));
}
}
/* recursively
(i) get each Part listed in the relationships
(ii) add the Part to the zip file
(iii) traverse its relationship
*/
public void addPartsFromRelationships(ZipOutputStream out, RelationshipsPart rp)
throws Docx4JException {
// for (Iterator it = rp.iterator(); it.hasNext(); ) {
// Relationship r = (Relationship)it.next();
// log.info("For Relationship Id=" + r.getId() + " Source is " + r.getSource().getPartName() +
// ", Target is " + r.getTargetURI() );
for (Relationship r : rp.getRelationships().getRelationship()) {
log.debug(
"For Relationship Id="
+ r.getId()
+ " Source is "
+ rp.getSourceP().getPartName()
+ ", Target is "
+ r.getTarget());
if (r.getType().equals(Namespaces.HYPERLINK)) {
continue; // whether internal or external
}
if (r.getTargetMode() != null && r.getTargetMode().equals("External")) {
// ie its EXTERNAL
// As at 1 May 2008, we don't have a Part for these;
// there is just the relationship.
log.warn("Encountered external resource " + r.getTarget() + " of type " + r.getType());
// So
continue;
}
try {
// String resolvedPartUri = URIHelper.resolvePartUri(r.getSourceURI(), r.getTargetURI()
// ).toString();
String resolvedPartUri =
URIHelper.resolvePartUri(rp.getSourceURI(), new URI(r.getTarget())).toString();
// Now drop leading "/'
resolvedPartUri = resolvedPartUri.substring(1);
// Now normalise it .. ie abc/def/../ghi
// becomes abc/ghi
// Maybe this isn't necessary with a zip file,
// - ZipFile class may be smart enough to do it.
// But it is certainly necessary in the JCR case.
// target = (new java.net.URI(target)).normalize().toString();
// log.info("Normalised, it is " + target );
// Document contents = getDocumentFromZippedPart( zf, target);
if (!false) {
log.debug("Getting part /" + resolvedPartUri);
// Part part = p.getParts().get(new PartName("/" + resolvedPartUri));
Part part = rp.getPart(r);
// 2012 09 26: If the part is actually attached to
// a different package, using this you can still get it.
// Use this 'feature' at your own risk!
if (part == null) {
log.error("Part " + resolvedPartUri + " not found!");
} else {
log.debug(part.getClass().getName());
}
if (!part.getPackage().equals(p)) {
log.warn("Part " + resolvedPartUri + " is attached to some other package");
}
savePart(out, part);
}
} catch (Exception e) {
throw new Docx4JException(
"Failed to add parts from relationships of " + rp.getSourceP().getPartName(), e);
}
}
}
public OpcPackage get() throws Docx4JException {
long startTime = System.currentTimeMillis();
// 1. Get [Content_Types].xml
ContentTypeManager ctm = new ContentTypeManager();
InputStream is = null;
try {
is = partStore.loadPart("[Content_Types].xml");
ctm.parseContentTypesFile(is);
} catch (Docx4JException e) {
throw new Docx4JException("Couldn't get [Content_Types].xml from ZipFile", e);
} catch (NullPointerException e) {
throw new Docx4JException("Couldn't get [Content_Types].xml from ZipFile", e);
} finally {
IOUtils.closeQuietly(is);
}
// .. now find the name of the main part
String partName = "_rels/.rels";
RelationshipsPart rp = getRelationshipsPartFromZip(null, partName);
if (rp == null) {
throw new Docx4JException("_rels/.rels appears to be missing from this package!");
}
String mainPartName = PackageRelsUtil.getNameOfMainPart(rp);
PartName mainPartNameObj;
if (mainPartName.startsWith("/")) {
// OpenXML SDK 2.0 writes Target="/word/document.xml" (note leading "/")
mainPartNameObj = new PartName(mainPartName);
} else {
// Microsoft Word, docx4j etc write Target="word/document.xml"
mainPartNameObj = new PartName("/" + mainPartName);
}
String pkgContentType = ctm.getContentType(mainPartNameObj);
// 2. Create a new Package; this'll return the appropriate subclass
OpcPackage p = ctm.createPackage(pkgContentType);
log.info("Instantiated package of type " + p.getClass().getName());
p.setPartStore(partStore);
p.setRelationships(rp);
rp.setSourceP(p); //
// 5. Now recursively
// (i) create new Parts for each thing listed
// in the relationships
// (ii) add the new Part to the package
// (iii) cross the PartName off unusedZipEntries
addPartsFromRelationships(p, rp, ctm);
// 6.
registerCustomXmlDataStorageParts(p);
// partStore.finishLoad();
long endTime = System.currentTimeMillis();
log.info("package read; elapsed time: " + Math.round((endTime - startTime)) + " ms");
return p;
}
/**
* Traverse the document, looking for fonts which have been applied, either directly, or via a
* style.
*
* @return
*/
public Set<String> fontsInUse() {
log.info("fontsInUse..");
getPropertyResolver(); // this inits our virtual DocDefaults style
// Setup
Set<String> fontsDiscovered = new java.util.HashSet<String>();
// // Keep track of styles we encounter, so we can
// // inspect these for fonts
// Set<String> stylesInUse = new java.util.HashSet<String>();
//
// org.docx4j.wml.Styles styles = null;
// if (this.getStyleDefinitionsPart()!=null) {
// styles = (org.docx4j.wml.Styles)this.getStyleDefinitionsPart().getJaxbElement();
// }
// // It is convenient to have a HashMap of styles
// Map<String, Style> stylesDefined = new java.util.HashMap<String, Style>();
// if (styles!=null) {
// for (Iterator<Style> iter = styles.getStyle().iterator(); iter.hasNext();) {
// Style s = iter.next();
// stylesDefined.put(s.getStyleId(), s);
// }
// }
// // We need to know what fonts and styles are used in the document
org.docx4j.wml.Document wmlDocumentEl = (org.docx4j.wml.Document) this.getJaxbElement();
Body body = wmlDocumentEl.getBody();
List<Object> bodyChildren = body.getContent();
FontDiscoveryCharacterVisitor visitor = new FontDiscoveryCharacterVisitor(fontsDiscovered);
RunFontSelector runFontSelector =
new RunFontSelector(
(WordprocessingMLPackage) this.pack, visitor, RunFontActionType.DISCOVERY);
FontAndStyleFinder finder = new FontAndStyleFinder(runFontSelector, fontsDiscovered, null);
finder.defaultCharacterStyle = this.getStyleDefinitionsPart().getDefaultCharacterStyle();
finder.defaultParagraphStyle = this.getStyleDefinitionsPart().getDefaultParagraphStyle();
new TraversalUtil(bodyChildren, finder);
// finder.finish();
fontsDiscovered.add(runFontSelector.getDefaultFont());
// fonts in headers, footers?
RelationshipsPart rp = this.getRelationshipsPart();
if (rp != null) {
for (Relationship r : rp.getRelationships().getRelationship()) {
Part part = rp.getPart(r);
if (part instanceof FooterPart) {
Ftr ftr = ((FooterPart) part).getJaxbElement();
finder.walkJAXBElements(ftr);
} else if (part instanceof HeaderPart) {
Hdr hdr = ((HeaderPart) part).getJaxbElement();
finder.walkJAXBElements(hdr);
}
}
}
// Styles in endnotes, footnotes?
if (this.getEndNotesPart() != null) {
log.debug("Looking at endnotes");
CTEndnotes endnotes = this.getEndNotesPart().getJaxbElement();
finder.walkJAXBElements(endnotes);
}
if (this.getFootnotesPart() != null) {
log.debug("Looking at footnotes");
CTFootnotes footnotes = this.getFootnotesPart().getJaxbElement();
finder.walkJAXBElements(footnotes);
}
// Comments
if (this.getCommentsPart() != null) {
log.debug("Looking at comments");
Comments comments = this.getCommentsPart().getJaxbElement();
finder.walkJAXBElements(comments);
}
// Add fonts used in the styles we discovered
// .. 2013 03 10: no longer necessary
// Fonts can also be used in the numbering part
// For now, treat any font mentioned in that part as in use.
// Ideally, we'd only register fonts used in numbering levels
// that were actually used in the document
if (getNumberingDefinitionsPart() != null) {
Numbering numbering = getNumberingDefinitionsPart().getJaxbElement();
for (Numbering.AbstractNum abstractNumNode : numbering.getAbstractNum()) {
for (Lvl lvl : abstractNumNode.getLvl()) {
if (lvl.getRPr() != null && lvl.getRPr().getRFonts() != null) {
String fontName = lvl.getRPr().getRFonts().getAscii();
if (fontName != null) {
fontsDiscovered.add(fontName);
log.debug(
"Registered "
+ fontName
+ " for abstract list "
+ abstractNumNode.getAbstractNumId()
+ " lvl "
+ lvl.getIlvl());
}
}
}
}
}
return fontsDiscovered;
}
