/**
* 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 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;
}
/* Save a Package as a Zip file in the outputstream provided */
public boolean save(OutputStream realOS) throws Docx4JException {
handled = new HashMap<String, String>();
try {
ZipOutputStream out = new ZipOutputStream(realOS);
// 3. Save [Content_Types].xml
ContentTypeManager ctm = p.getContentTypeManager();
out.putNextEntry(new ZipEntry("[Content_Types].xml"));
ctm.marshal(out);
out.closeEntry();
// 4. Start with _rels/.rels
// <Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">
// <Relationship Id="rId3"
// Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/extended-properties" Target="docProps/app.xml"/>
// <Relationship Id="rId2"
// Type="http://schemas.openxmlformats.org/package/2006/relationships/metadata/core-properties" Target="docProps/core.xml"/>
// <Relationship Id="rId1"
// Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument"
// Target="word/document.xml"/>
// </Relationships>
String partName = "_rels/.rels";
RelationshipsPart rp = p.getRelationshipsPart();
// deprecatedSaveRawXmlPart(out, partName, rp.getDocument() );
// 2008 06 12 - try this neater method
saveRawXmlPart(out, rp, partName);
// 5. Now recursively
// addPartsFromRelationships(out, "", rp );
addPartsFromRelationships(out, rp);
// Complete the ZIP file
// Don't forget to do this or everything will appear
// to work, but when you open the zip file you'll get an error
// "End-of-central-directory signature not found."
out.close();
realOS.close();
} catch (Exception e) {
e.printStackTrace();
if (e instanceof Docx4JException) {
throw (Docx4JException) e;
} else {
throw new Docx4JException("Failed to save package", e);
}
}
log.info("...Done!");
return true;
}
// public static Part getBinaryPart(ZipFile zf, ContentTypeManager ctm, String resolvedPartUri)
public static Part getBinaryPart(
HashMap<String, ByteArray> partByteArrays, ContentTypeManager ctm, String resolvedPartUri)
throws Docx4JException {
Part part = null;
InputStream in = null;
try {
// in = zf.getInputStream( zf.getEntry(resolvedPartUri ) );
in = partByteArrays.get(resolvedPartUri).getInputStream();
part = new BinaryPart(new PartName("/" + resolvedPartUri));
// Set content type
part.setContentType(new ContentType(ctm.getContentType(new PartName("/" + resolvedPartUri))));
((BinaryPart) part).setBinaryData(in);
log.info("Stored as BinaryData");
} catch (Exception ioe) {
ioe.printStackTrace();
} finally {
if (in != null) {
try {
in.close();
} catch (IOException exc) {
exc.printStackTrace();
}
}
}
return part;
}
/**
* Create a linked image part, and attach it as a rel of the specified source part (eg a header
* part).
*
* <p>The current behaviour is that the part is added to the package, but since the target mode of
* the rel is external, the part is redundant.
*
* @param wordMLPackage
* @param sourcePart
* @param url
* @return
* @throws Exception
*/
public static BinaryPartAbstractImage createLinkedImagePart(
OpcPackage opcPackage, Part sourcePart, URL url) throws Exception {
log.debug("Incoming url for linked image: " + url.toString());
ImageInfo info =
ensureFormatIsSupported(url, null, null, false); // final param doesn't matter in this case
ContentTypeManager ctm = opcPackage.getContentTypeManager();
String proposedRelId = sourcePart.getRelationshipsPart().getNextId();
// In order to ensure unique part name,
// idea is to use the relId, which ought to be unique
String ext = info.getMimeType().substring(info.getMimeType().indexOf("/") + 1);
BinaryPartAbstractImage imagePart =
(BinaryPartAbstractImage)
ctm.newPartForContentType(
info.getMimeType(),
createImageName(opcPackage, sourcePart, proposedRelId, ext),
null);
// NB: contents never populated
log.debug(
"created part "
+ imagePart.getClass().getName()
+ " with name "
+ imagePart.getPartName().toString());
imagePart.rels.add(
sourcePart.addTargetPart(
imagePart)); // want to create rel with suitable name; side effect is to add part
imagePart.getRelLast().setTargetMode("External");
opcPackage.getExternalResources().put(imagePart.getExternalTarget(), imagePart);
// if (!url.getProtocol().equals("file") && new File(url.toString() ).isFile()) {
// imagePart.rel.setTarget("file:///" + url);
// } else {
// imagePart.rel.setTarget(url.toString());
// }
imagePart.getRelLast().setTarget(url.toString());
imagePart.setImageInfo(info);
return imagePart;
}
/**
* 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;
}
public Part getBinaryPart(ContentTypeManager ctm, String resolvedPartUri) throws Docx4JException {
Part part = null;
InputStream is = null;
try {
is = partStore.loadPart(resolvedPartUri);
// in = partByteArrays.get(resolvedPartUri).getInputStream();
part = new BinaryPart(new PartName("/" + resolvedPartUri));
// Set content type
part.setContentType(new ContentType(ctm.getContentType(new PartName("/" + resolvedPartUri))));
((BinaryPart) part).setBinaryData(is);
log.info("Stored as BinaryData");
} catch (Exception ioe) {
ioe.printStackTrace();
} finally {
IOUtils.closeQuietly(is);
}
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 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;
}
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;
}
/**
* 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 OpcPackage get(File f) throws Docx4JException {
log.info("Filepath = " + f.getPath());
ZipFile zf = null;
try {
if (!f.exists()) {
log.info("Couldn't find " + f.getPath());
}
zf = new ZipFile(f);
} catch (IOException ioe) {
ioe.printStackTrace();
throw new Docx4JException("Couldn't get ZipFile", ioe);
}
// dumpZipFileContents(zf);
// 1. The idea is to walk the tree of relationships, getting
// everything we need from the zip file. But I'd like to know
// whether there are any orphans, so first we make
// a HashMap containing the names of all the zip file
// entries, so we can tick them off.
// unusedZipEntries = new HashMap();
Enumeration entries = zf.entries();
while (entries.hasMoreElements()) {
ZipEntry entry = (ZipEntry) entries.nextElement();
// unusedZipEntries.put(entry.getName(), new Boolean(true) );
}
// 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(zf, "[Content_Types].xml");
ctm.parseContentTypesFile(is);
} catch (IOException e) {
throw new Docx4JException("Couldn't get [Content_Types].xml from ZipFile", e);
}
OpcPackage p = ctm.createPackage();
// 3. Get [Content_Types].xml
// Once we've got this, then we can look up the content type for
// each PartName, and use it in the Part constructor.
// p.setContentTypeManager(ctm); - 20080111 - done by ctm.createPackage();
// unusedZipEntries.put("[Content_Types].xml", new Boolean(false));
// 4. Start with _rels/.rels
// <Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">
// <Relationship Id="rId3"
// Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/extended-properties" Target="docProps/app.xml"/>
// <Relationship Id="rId2"
// Type="http://schemas.openxmlformats.org/package/2006/relationships/metadata/core-properties"
// Target="docProps/core.xml"/>
// <Relationship Id="rId1"
// Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument"
// Target="word/document.xml"/>
// </Relationships>
String partName = "_rels/.rels";
RelationshipsPart rp = getRelationshipsPartFromZip(p, zf, partName);
p.setRelationships(rp);
// rp.setPackageRelationshipPart(true);
// unusedZipEntries.put(partName, new Boolean(false));
log.info("Object created for: " + partName);
// log.info( rp.toString());
// 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(zf, 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));
// }
// }
try {
zf.close();
} catch (IOException exc) {
exc.printStackTrace();
}
registerCustomXmlDataStorageParts(p);
return p;
}
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
}
}
}
/**
* 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;
}
