/**
* Convenience method, given a Graphic in a document, to get the byte[] representing the
* associated image
*
* @param wmlPkg
* @param graphic
* @return
*/
public static byte[] getImage(WordprocessingMLPackage wmlPkg, org.docx4j.dml.Graphic graphic) {
if (wmlPkg == null
|| wmlPkg.getMainDocumentPart() == null
|| wmlPkg.getMainDocumentPart().getRelationshipsPart() == null) {
return null;
}
Pic pic = graphic.getGraphicData().getPic();
String rId = pic.getBlipFill().getBlip().getEmbed();
if (rId.equals("")) {
rId = pic.getBlipFill().getBlip().getLink();
}
log.debug("Image rel id: " + rId);
org.docx4j.relationships.Relationship rel =
wmlPkg.getMainDocumentPart().getRelationshipsPart().getRelationshipByID(rId);
if (rel != null) {
org.docx4j.openpackaging.parts.Part part =
wmlPkg.getMainDocumentPart().getRelationshipsPart().getPart(rel);
if (part == null) {
log.error("Couldn't get Part!");
} else if (part instanceof BinaryPart) {
log.debug("getting bytes...");
return ((BinaryPart) part).getBytes();
} else {
log.error("Part was a " + part.getClass().getName());
}
} else {
log.error("Couldn't find rel " + rId);
}
return null;
}
/**
* Get a Part (except a relationships part), but not its relationships part or related parts.
* Useful if you need quick access to just this part. 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.
*
* @see To get a Part and all its related parts, and add all to a package, use getPart.
* @param partByteArrays
* @param ctm
* @param resolvedPartUri
* @param rel
* @return
* @throws Docx4JException including if result is null
*/
public Part getRawPart(ContentTypeManager ctm, String resolvedPartUri, Relationship rel)
throws Docx4JException {
Part part = null;
InputStream is = null;
try {
try {
log.debug("resolved uri: " + resolvedPartUri);
// Get a subclass of Part appropriate for this content type
// This will throw UnrecognisedPartException in the absence of
// specific knowledge. Hence it is important to get the is
// first, as we do above.
part = ctm.getPart("/" + resolvedPartUri, rel);
log.info("ctm returned " + part.getClass().getName());
if (part instanceof org.docx4j.openpackaging.parts.ThemePart
|| part instanceof org.docx4j.openpackaging.parts.DocPropsCorePart
|| part instanceof org.docx4j.openpackaging.parts.DocPropsCustomPart
|| part instanceof org.docx4j.openpackaging.parts.DocPropsExtendedPart
|| part instanceof org.docx4j.openpackaging.parts.CustomXmlDataStoragePropertiesPart
|| part instanceof org.docx4j.openpackaging.parts.digitalsignature.XmlSignaturePart
|| part instanceof org.docx4j.openpackaging.parts.JaxbXmlPart) {
// Nothing to do here
} else if (part instanceof org.docx4j.openpackaging.parts.WordprocessingML.BinaryPart) {
log.debug("Detected BinaryPart " + part.getClass().getName());
// is = partStore.loadPart( resolvedPartUri);
// ((BinaryPart)part).setBinaryData(is);
} else if (part instanceof org.docx4j.openpackaging.parts.CustomXmlDataStoragePart) {
// ContentTypeManager initially detects them as CustomXmlDataStoragePart;
// the below changes as necessary
// Is it a part we know?
is = partStore.loadPart(resolvedPartUri);
try {
Unmarshaller u = Context.jc.createUnmarshaller();
Object o = u.unmarshal(is);
log.debug(o.getClass().getName());
PartName name = part.getPartName();
if (o instanceof CoverPageProperties) {
part = new DocPropsCoverPagePart(name);
((DocPropsCoverPagePart) part).setJaxbElement((CoverPageProperties) o);
} else if (o instanceof org.opendope.conditions.Conditions) {
part = new ConditionsPart(name);
((ConditionsPart) part).setJaxbElement((org.opendope.conditions.Conditions) o);
} else if (o instanceof org.opendope.xpaths.Xpaths) {
part = new XPathsPart(name);
((XPathsPart) part).setJaxbElement((org.opendope.xpaths.Xpaths) o);
} else if (o instanceof org.opendope.questions.Questionnaire) {
part = new QuestionsPart(name);
((QuestionsPart) part).setJaxbElement((org.opendope.questions.Questionnaire) o);
} else if (o instanceof org.opendope.answers.Answers) {
part = new StandardisedAnswersPart(name);
((StandardisedAnswersPart) part).setJaxbElement((org.opendope.answers.Answers) o);
} else if (o instanceof org.opendope.components.Components) {
part = new ComponentsPart(name);
((ComponentsPart) part).setJaxbElement((org.opendope.components.Components) o);
} else if (o instanceof JAXBElement<?>
&& XmlUtils.unwrap(o) instanceof org.docx4j.bibliography.CTSources) {
part = new BibliographyPart(name);
((BibliographyPart) part)
.setJaxbElement((JAXBElement<org.docx4j.bibliography.CTSources>) o);
} else {
log.error("TODO: handle known CustomXmlPart part " + o.getClass().getName());
CustomXmlDataStorage data = getCustomXmlDataStorageClass().factory();
is.reset();
data.setDocument(is); // Not necessarily JAXB, that's just our method name
((org.docx4j.openpackaging.parts.CustomXmlDataStoragePart) part).setData(data);
}
} catch (javax.xml.bind.UnmarshalException ue) {
log.warn("No JAXB model for this CustomXmlDataStorage part; " + ue.getMessage());
CustomXmlDataStorage data = getCustomXmlDataStorageClass().factory();
is.reset();
data.setDocument(is); // Not necessarily JAXB, that's just our method name
((org.docx4j.openpackaging.parts.CustomXmlDataStoragePart) part).setData(data);
}
} else if (part instanceof org.docx4j.openpackaging.parts.XmlPart) {
is = partStore.loadPart(resolvedPartUri);
// try {
((XmlPart) part).setDocument(is);
// Experimental 22/6/2011; don't fall back to binary (which we used to)
// } catch (Docx4JException d) {
// // This isn't an XML part after all,
// // even though ContentTypeManager detected it as such
// // So get it as a binary part
// part = getBinaryPart(partByteArrays, ctm, resolvedPartUri);
// log.warn("Could not parse as XML, so using BinaryPart for "
// + resolvedPartUri);
// ((BinaryPart)part).setBinaryData(is);
// }
} else {
// Shouldn't happen, since ContentTypeManagerImpl should
// return an instance of one of the above, or throw an
// Exception.
log.error("No suitable part found for: " + resolvedPartUri);
part = null;
}
} catch (PartUnrecognisedException e) {
log.error("PartUnrecognisedException shouldn't happen anymore!", e);
// Try to get it as a binary part
part = getBinaryPart(ctm, resolvedPartUri);
log.warn("Using BinaryPart for " + resolvedPartUri);
// is = partStore.loadPart( resolvedPartUri);
// ((BinaryPart)part).setBinaryData(is);
}
} catch (Exception ex) {
// IOException, URISyntaxException
ex.printStackTrace();
throw new Docx4JException("Failed to getPart", ex);
} finally {
IOUtils.closeQuietly(is);
}
if (part == null) {
throw new Docx4JException(
"cannot find part "
+ resolvedPartUri
+ " from rel "
+ rel.getId()
+ "="
+ rel.getTarget());
}
return part;
}
/**
* Get a Part (except a relationships part), but not its relationships part or related parts.
* Useful if you need quick access to just this part. 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.
*
* @see To get a Part and all its related parts, and add all to a package, use getPart.
* @param zf
* @param resolvedPartUri
* @return
* @throws URISyntaxException
* @throws InvalidFormatException
*/
public static Part getRawPart(
ZipFile zf, ContentTypeManager ctm, String resolvedPartUri, Relationship rel)
throws Docx4JException {
Part part = null;
InputStream is = null;
try {
try {
log.debug("resolved uri: " + resolvedPartUri);
is = getInputStreamFromZippedPart(zf, resolvedPartUri);
// Get a subclass of Part appropriate for this content type
// This will throw UnrecognisedPartException in the absence of
// specific knowledge. Hence it is important to get the is
// first, as we do above.
part = ctm.getPart("/" + resolvedPartUri, rel);
if (part instanceof org.docx4j.openpackaging.parts.ThemePart) {
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).setJAXBContext(Context.jcThemePart);
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).unmarshal(is);
} else if (part instanceof org.docx4j.openpackaging.parts.DocPropsCorePart) {
((org.docx4j.openpackaging.parts.JaxbXmlPart) part)
.setJAXBContext(Context.jcDocPropsCore);
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).unmarshal(is);
} else if (part instanceof org.docx4j.openpackaging.parts.DocPropsCustomPart) {
((org.docx4j.openpackaging.parts.JaxbXmlPart) part)
.setJAXBContext(Context.jcDocPropsCustom);
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).unmarshal(is);
} else if (part instanceof org.docx4j.openpackaging.parts.DocPropsExtendedPart) {
((org.docx4j.openpackaging.parts.JaxbXmlPart) part)
.setJAXBContext(Context.jcDocPropsExtended);
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).unmarshal(is);
} else if (part
instanceof org.docx4j.openpackaging.parts.CustomXmlDataStoragePropertiesPart) {
((org.docx4j.openpackaging.parts.JaxbXmlPart) part)
.setJAXBContext(Context.jcCustomXmlProperties);
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).unmarshal(is);
} else if (part
instanceof org.docx4j.openpackaging.parts.digitalsignature.XmlSignaturePart) {
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).setJAXBContext(Context.jcXmlDSig);
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).unmarshal(is);
} else if (part instanceof org.docx4j.openpackaging.parts.JaxbXmlPart) {
// MainDocument part, Styles part, Font part etc
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).setJAXBContext(Context.jc);
((org.docx4j.openpackaging.parts.JaxbXmlPart) part).unmarshal(is);
} else if (part instanceof org.docx4j.openpackaging.parts.WordprocessingML.BinaryPart) {
log.debug("Detected BinaryPart " + part.getClass().getName());
if (conserveMemory) {
((BinaryPart) part).setBinaryDataRef(zf.getName(), resolvedPartUri);
} else {
((BinaryPart) part).setBinaryData(is);
}
} else if (part instanceof org.docx4j.openpackaging.parts.CustomXmlDataStoragePart) {
// Is it a part we know?
try {
Unmarshaller u = Context.jc.createUnmarshaller();
Object o = u.unmarshal(is);
log.debug(o.getClass().getName());
PartName name = part.getPartName();
if (o instanceof org.opendope.conditions.Conditions) {
part = new ConditionsPart(name);
((ConditionsPart) part).setJaxbElement((org.opendope.conditions.Conditions) o);
} else if (o instanceof org.opendope.xpaths.Xpaths) {
part = new XPathsPart(name);
((XPathsPart) part).setJaxbElement((org.opendope.xpaths.Xpaths) o);
} else if (o instanceof org.opendope.questions.Questionnaire) {
part = new QuestionsPart(name);
((QuestionsPart) part).setJaxbElement((org.opendope.questions.Questionnaire) o);
} else if (o instanceof org.opendope.answers.Answers) {
part = new StandardisedAnswersPart(name);
((StandardisedAnswersPart) part).setJaxbElement((org.opendope.answers.Answers) o);
} else if (o instanceof org.opendope.components.Components) {
part = new ComponentsPart(name);
((ComponentsPart) part).setJaxbElement((org.opendope.components.Components) o);
} else if (o instanceof JAXBElement<?>
&& XmlUtils.unwrap(o) instanceof org.docx4j.bibliography.CTSources) {
part = new BibliographyPart(name);
((BibliographyPart) part)
.setJaxbElement((JAXBElement<org.docx4j.bibliography.CTSources>) o);
} else {
log.warn("No known part after all for CustomXmlPart " + o.getClass().getName());
CustomXmlDataStorage data = getCustomXmlDataStorageClass().factory();
is.reset();
data.setDocument(is); // Not necessarily JAXB, that's just our method name
((org.docx4j.openpackaging.parts.CustomXmlDataStoragePart) part).setData(data);
}
} catch (javax.xml.bind.UnmarshalException ue) {
// No ...
CustomXmlDataStorage data = getCustomXmlDataStorageClass().factory();
is.reset();
data.setDocument(is); // Not necessarily JAXB, that's just our method name
((org.docx4j.openpackaging.parts.CustomXmlDataStoragePart) part).setData(data);
}
} else if (part instanceof org.docx4j.openpackaging.parts.XmlPart) {
try {
((XmlPart) part).setDocument(is);
} catch (Docx4JException d) {
// This isn't an XML part after all,
// even though ContentTypeManager detected it as such
// So get it as a binary part
part = getBinaryPart(zf, ctm, resolvedPartUri);
if (conserveMemory) {
((BinaryPart) part).setBinaryDataRef(zf.getName(), resolvedPartUri);
} else {
((BinaryPart) part).setBinaryData(is);
}
}
} else {
// Shouldn't happen, since ContentTypeManagerImpl should
// return an instance of one of the above, or throw an
// Exception.
log.error("No suitable part found for: " + resolvedPartUri);
part = null;
}
} catch (PartUnrecognisedException e) {
log.warn("PartUnrecognisedException shouldn't happen anymore!");
// Try to get it as a binary part
part = getBinaryPart(zf, ctm, resolvedPartUri);
if (conserveMemory) {
((BinaryPart) part).setBinaryDataRef(zf.getName(), resolvedPartUri);
} else {
((BinaryPart) part).setBinaryData(is);
}
}
} catch (Exception ex) {
// IOException, URISyntaxException
ex.printStackTrace();
throw new Docx4JException("Failed to getPart", ex);
} finally {
if (is != null) {
try {
is.close();
} catch (IOException exc) {
exc.printStackTrace();
}
}
}
return part;
}
public static void apply(WordprocessingMLPackage otherPackage, Alterations alterations)
throws Docx4JException, JAXBException {
if (alterations.getContentTypes() != null) {
log.info("replacing [Content_Types].xml");
ContentTypeManager newCTM = new ContentTypeManager();
newCTM.parseContentTypesFile(new ByteArrayInputStream(alterations.getContentTypes()));
otherPackage.setContentTypeManager(newCTM);
}
if (alterations.isEmpty()) return;
// -- Deletions
List<PartName> removedParts = new ArrayList<PartName>();
for (Alteration a : alterations.getPartsDeleted()) {
org.docx4j.xmlPackage.Part xmlpart = a.getPart();
// These deleted parts are likely to be attached to
// to otherPackage, but there is no requirement for
// that to be so. In other words, you could try
// to apply the alterations to some third docx.
// It might have already been deleted as a consequence
// of recursive deletion...
if (removedParts.contains(xmlpart.getName())) continue;
// Design decision: how to find owning rels part?
// Since part might not be from this package, can't do:
// part.getOwningRelationshipPart().removePart(p.getPartName())
// We could store the info when AlteredParts is run,
// but lets try to get away without that...
// If a part has been deleted, we know its owning rels will
// have been modified or deleted. So look for that.
// BUT IT WON'T BE THERE IF ITS BEEN DELETED! DOH!
// So we have to store the info when AlteredParts is run
// after all.
Part parentPart = otherPackage.getParts().get(a.getSourcePartName());
if (a.getPart().getContentType().equals(ContentTypes.RELATIONSHIPS_PART)) {
parentPart.setRelationships(null);
} else {
removedParts.addAll(
parentPart.getRelationshipsPart().removePart(new PartName(xmlpart.getName())));
}
}
// -- Modifications
for (Alteration a : alterations.getPartsModified()) {
log.info("Applying modifications to part: " + a.getPart().getName());
if (a.getPart().getContentType().equals(ContentTypes.RELATIONSHIPS_PART)) {
RelationshipsPart newRP = null; // FlatOpcXmlImporter.createRelationshipsPart(a.getPart());
if (a.getSourcePartName().equals("/")) {
newRP = otherPackage.getRelationshipsPart(true);
// otherPackage.setRelationships(newRP);
// newRP.setSourceP(otherPackage);
} else {
Part parentPart = otherPackage.getParts().get(a.getSourcePartName());
newRP = parentPart.getRelationshipsPart(true);
// parentPart.setRelationships(newRP);
// newRP.setSourceP(parentPart);
}
FlatOpcXmlImporter.populateRelationshipsPart(newRP, a.getPart().getXmlData().getAny());
} else {
Part targetPart = otherPackage.getParts().get(new PartName(a.getPart().getName()));
if (targetPart == null) {
log.error(
"Couldn't find " + a.getPart().getName() + " @ " + a.getSourcePartName().getName());
continue;
}
Part tmpPart =
FlatOpcXmlImporter.getRawPart(otherPackage.getContentTypeManager(), a.getPart(), null);
if (targetPart instanceof JaxbXmlPart) {
((JaxbXmlPart) targetPart).setJaxbElement(((JaxbXmlPart) tmpPart).getJaxbElement());
} else if (targetPart instanceof XmlPart) {
((XmlPart) targetPart).setDocument(((XmlPart) tmpPart).getDocument());
} else if (targetPart instanceof CustomXmlDataStoragePart) {
((CustomXmlDataStoragePart) targetPart)
.setData(((CustomXmlDataStoragePart) tmpPart).getData());
// TODO: check that
} else if (targetPart instanceof BinaryPart) {
((BinaryPart) targetPart).setBinaryData(((BinaryPart) tmpPart).getBuffer());
} else {
log.error("TODO: handle " + targetPart.getClass().getName());
}
}
}
// -- Additions
for (Alteration a : alterations.getPartsAdded()) {
log.info("Adding part: " + a.getPart().getName());
if (a.getPart().getContentType().equals(ContentTypes.RELATIONSHIPS_PART)) {
RelationshipsPart newRP = null; // FlatOpcXmlImporter.createRelationshipsPart(a.getPart());
if (a.getSourcePartName().equals("/")) {
newRP = otherPackage.getRelationshipsPart(true);
// otherPackage.setRelationships(newRP);
// newRP.setSourceP(otherPackage);
} else {
Part parentPart = otherPackage.getParts().get(a.getSourcePartName());
newRP = parentPart.getRelationshipsPart(true);
// parentPart.setRelationships(newRP);
// newRP.setSourceP(parentPart);
}
FlatOpcXmlImporter.populateRelationshipsPart(newRP, a.getPart().getXmlData().getAny());
} else {
Part parentPart = otherPackage.getParts().get(a.getSourcePartName());
Part newPart =
FlatOpcXmlImporter.getRawPart(otherPackage.getContentTypeManager(), a.getPart(), null);
// There will already be a rel for the new part,
// since we will already have modified or added the rels part
// so don't do AddTargetPart (which will probably create a new rel id,
// which will cause problems)
newPart.setOwningRelationshipPart(parentPart.getRelationshipsPart());
newPart
.getSourceRelationships()
.add(parentPart.getRelationshipsPart().getRel(new PartName(a.getPart().getName())));
otherPackage.getParts().put(newPart);
newPart.setPackage(otherPackage);
// TODO: add content type if necessary
}
}
}
/* 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);
}
}
}