From e8c6469cb924440df13d276bf6714c484db6b124 Mon Sep 17 00:00:00 2001
From: Thorsten Vitt <thorsten.vitt@uni-wuerzburg.de>
Date: Tue, 3 Sep 2013 18:35:25 +0200
Subject: [PATCH] Refactored towards HTMLWriter. Need tests and cleanup.
---
.../services/aggregator/html/HTML.java | 103 ++-----
.../services/aggregator/html/HTMLWriter.java | 265 ++++++++++++++++++
2 files changed, 281 insertions(+), 87 deletions(-)
create mode 100644 src/main/java/info/textgrid/services/aggregator/html/HTMLWriter.java
diff --git a/src/main/java/info/textgrid/services/aggregator/html/HTML.java b/src/main/java/info/textgrid/services/aggregator/html/HTML.java
index 881204f..ad7f358 100644
--- a/src/main/java/info/textgrid/services/aggregator/html/HTML.java
+++ b/src/main/java/info/textgrid/services/aggregator/html/HTML.java
@@ -1,22 +1,17 @@
package info.textgrid.services.aggregator.html;
-import info.textgrid.namespaces.metadata.core._2010.MetadataContainerType;
import info.textgrid.namespaces.metadata.core._2010.ObjectType;
import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.AuthFault;
import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.IoFault;
import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.MetadataParseFault;
import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.ObjectNotFoundFault;
import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.ProtocolNotImplementedFault;
-import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.TGCrudService;
-import info.textgrid.services.aggregator.GenericExceptionMapper;
import info.textgrid.services.aggregator.ITextGridRep;
import info.textgrid.services.aggregator.ITextGridRep.TGOSupplier;
import info.textgrid.services.aggregator.TextGridRepProvider;
-import info.textgrid.services.aggregator.teicorpus.TEICorpusSerializer;
import java.io.IOException;
import java.io.InputStream;
-import java.io.OutputStream;
import java.net.MalformedURLException;
import java.net.URI;
import java.net.URL;
@@ -33,27 +28,22 @@
import javax.ws.rs.QueryParam;
import javax.ws.rs.WebApplicationException;
import javax.ws.rs.core.Context;
-import javax.ws.rs.core.Response.Status;
import javax.ws.rs.core.StreamingOutput;
import javax.xml.transform.stream.StreamSource;
import net.sf.saxon.s9api.Processor;
-import net.sf.saxon.s9api.QName;
import net.sf.saxon.s9api.SaxonApiException;
-import net.sf.saxon.s9api.XdmAtomicValue;
import net.sf.saxon.s9api.XsltCompiler;
import net.sf.saxon.s9api.XsltExecutable;
-import net.sf.saxon.s9api.XsltTransformer;
import org.apache.cxf.jaxrs.model.wadl.Description;
-import com.google.common.base.Stopwatch;
+import com.google.common.base.Optional;
import com.google.common.cache.Cache;
import com.google.common.cache.CacheBuilder;
import com.google.common.cache.CacheLoader;
import com.google.common.cache.RemovalListener;
import com.google.common.cache.RemovalNotification;
-import com.google.common.io.FileBackedOutputStream;
@Path("/html")
@Description("Creates an HTML representation of the given TEI document, or aggregation of TEI documents. This is currently extremely experimental and probably broken.")
@@ -62,20 +52,20 @@ public class HTML {
private static final String TO_HTML_XSL = "/WEB-INF/stylesheets/db2xhtml.xsl";
- private ITextGridRep repository = TextGridRepProvider.getInstance();
+ ITextGridRep repository = TextGridRepProvider.getInstance();
final Logger logger = Logger
.getLogger("info.textgrid.services.aggregator.html.HTML");
private XsltExecutable toHtml;
- private final Processor xsltProcessor;
+ final Processor xsltProcessor;
@Context
private ServletContext servlet;
private Cache<URI, XsltExecutable> stylesheets;
- private XsltExecutable getToHtml() {
+ XsltExecutable getToHtml() {
if (toHtml == null) {
try {
final URL stylesheet = servlet.getResource(TO_HTML_XSL);
@@ -148,15 +138,16 @@ public XsltExecutable load(final URI url) throws Exception {
* @param uri
* the URI of the stylesheet to load
* @param sid
- * the session ID to use, or null.
+ * the session ID to use, if present
* @param forceLoad
- * TODO
+ * do not use a cached version even if present.
* @throws IOException
* if an error occurs reading the stylesheet.
* @throws SaxonApiException
* if saxon fails to compile the stylesheet.
*/
- protected XsltExecutable getStylesheet(final URI uri, final String sid,
+ protected XsltExecutable getStylesheet(final URI uri,
+ final Optional<String> sid,
final boolean forceLoad) throws SaxonApiException, IOException {
XsltExecutable executable = null;
@@ -171,7 +162,8 @@ protected XsltExecutable getStylesheet(final URI uri, final String sid,
if (TGUriResolver.isResolveable(uri)) {
// (2/3) it's a TextGrid object, load it from TG-crud.
- final TGOSupplier<InputStream> xsltSupplier = repository.read(uri, sid);
+ final TGOSupplier<InputStream> xsltSupplier = repository.read(
+ uri, sid.orNull());
executable = compiler.compile(new StreamSource(xsltSupplier
.getInput(), uri.toString()));
@@ -222,74 +214,11 @@ public StreamingOutput get(
SaxonApiException, ExecutionException {
logger.fine("HTML called for root object: " + uri);
- final Stopwatch stopwatch = new Stopwatch();
- stopwatch.start();
-
- final TGCrudService crud = repository.getCRUDService();
- final MetadataContainerType container = crud.readMetadata(sid, null,
- uri.toString());
- final ObjectType rootObject = container.getObject();
- final String mimeType = rootObject.getGeneric().getProvided().getFormat();
- final boolean aggregation = mimeType.contains("aggregation");
- if (!aggregation && !mimeType.matches("^text/.*xml.*")) {
- final String errorMsg = MessageFormat.format("The HTML export can only convert aggregations or XML documents to EPUB, however, the document {0} you referred to has the MIME type {1}.", uri, mimeType);
- throw new WebApplicationException(
- GenericExceptionMapper.toResponse(
- Status.UNSUPPORTED_MEDIA_TYPE, errorMsg, ""));
- }
- final InputStream tei;
- if (aggregation) {
- final TEICorpusSerializer corpusSerializer = new TEICorpusSerializer(
- rootObject, false, sid);
- final FileBackedOutputStream corpusBuffer = new FileBackedOutputStream(
- 1024 * 1024, true);
- corpusSerializer.write(corpusBuffer);
- corpusBuffer.close();
- tei = corpusBuffer.getSupplier().getInput();
- logger.fine(" created intermediate corpus for " + uri);
- } else {
- tei = repository.getContent(uri, sid);
- }
- logger.info("we have an input document after " + stopwatch.toString());
-
- final XsltTransformer transformer;
- if (xsluri == null || "".equals(xsluri)) {
- transformer = getToHtml().load();
- } else {
- transformer = getStylesheet(xsluri, sid, refreshStylesheet).load();
- if (sid != null) {
- transformer.setURIResolver(new TGUriResolver(repository, sid));
- } // otherwise default public URI resolver
- }
-
- transformer.setSource(new StreamSource(tei));
- transformer.setParameter(new QName("graphicsURLPattern"),
- new XdmAtomicValue(repository.getCRUDRestEndpoint()
- + "/@URI@/data"
- + ((sid == null || "".equals(sid)) ? ""
- : ("?sessionId=" + sid))));
- if (css != null) {
- transformer.setParameter(new QName("cssFile"), new XdmAtomicValue(css));
- }
-
- logger.info("we're ready to transform after " + stopwatch.toString());
- return new StreamingOutput() {
-
- @Override
- public void write(final OutputStream output) throws IOException,
- WebApplicationException {
- transformer.setDestination(xsltProcessor.newSerializer(output));
- try {
- transformer.transform();
- logger.info(MessageFormat
- .format("Finished transformation to HTML for {0} after {1}",
- uri, stopwatch.toString()));
- } catch (final SaxonApiException e) {
- throw new WebApplicationException(e);
- }
- }
- };
-
-
+ final HTMLWriter writer = new HTMLWriter(this, uri).stylesheet(xsluri)
+ .sid(sid).refresh(refreshStylesheet).css(css);
+ writer.loadSource();
+ writer.loadStylesheet();
+ return writer;
}
+
}
diff --git a/src/main/java/info/textgrid/services/aggregator/html/HTMLWriter.java b/src/main/java/info/textgrid/services/aggregator/html/HTMLWriter.java
new file mode 100644
index 0000000..b52d46a
--- /dev/null
+++ b/src/main/java/info/textgrid/services/aggregator/html/HTMLWriter.java
@@ -0,0 +1,265 @@
+package info.textgrid.services.aggregator.html;
+
+import info.textgrid.namespaces.metadata.core._2010.ObjectType;
+import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.AuthFault;
+import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.IoFault;
+import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.MetadataParseFault;
+import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.ObjectNotFoundFault;
+import info.textgrid.namespaces.middleware.tgcrud.services.tgcrudservice.ProtocolNotImplementedFault;
+import info.textgrid.services.aggregator.GenericExceptionMapper;
+import info.textgrid.services.aggregator.ITextGridRep;
+import info.textgrid.services.aggregator.ITextGridRep.TGOSupplier;
+import info.textgrid.services.aggregator.teicorpus.TEICorpusSerializer;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.io.OutputStream;
+import java.net.URI;
+import java.net.URISyntaxException;
+import java.text.MessageFormat;
+
+import javax.ws.rs.WebApplicationException;
+import javax.ws.rs.core.Response.Status;
+import javax.ws.rs.core.StreamingOutput;
+import javax.xml.transform.Source;
+import javax.xml.transform.TransformerConfigurationException;
+import javax.xml.transform.TransformerFactory;
+import javax.xml.transform.TransformerFactoryConfigurationError;
+import javax.xml.transform.stream.StreamSource;
+
+import net.sf.saxon.s9api.QName;
+import net.sf.saxon.s9api.SaxonApiException;
+import net.sf.saxon.s9api.XdmAtomicValue;
+import net.sf.saxon.s9api.XsltExecutable;
+import net.sf.saxon.s9api.XsltTransformer;
+
+import com.google.common.base.Optional;
+import com.google.common.base.Throwables;
+import com.google.common.io.ByteStreams;
+import com.google.common.io.FileBackedOutputStream;
+
+/**
+ * The essential steps:
+ *
+ * <ol>
+ * <li>Construct the transformation source.
+ * <p>
+ * Depends on different factors: If it's an aggregation, we need to recursively
+ * construct a teiCorpus first. If we need to parse the
+ * {@code <?xsl-stylesheet?>} processing instruction, we need to at least
+ * partially parse the document twice, so we download it to a temporary buffer.
+ * Otherwise, the source could be the result stream of the call to TG-crud.
+ * </p>
+ * <p>
+ * Includes some, potentially longish, requests to TextGridRep or other sources
+ * that may cause exceptions.
+ * </p>
+ * </li>
+ * <li>Load and configure the stylesheet.
+ * <p>
+ * Typically, the order is argument-specified > associated (via pi) > default
+ * stylesheet, with the associated stylesheet only checked on request.
+ * </p>
+ * </li>
+ * <li>Perform the transformation and output the result.
+ * <p>
+ * This step essentially requires the output stream to be present.
+ * </p>
+ * </li>
+ * </ol>
+ *
+ * @author Thorsten Vitt <thorsten.vitt@uni-wuerzburg.de>
+ *
+ */
+public class HTMLWriter implements StreamingOutput {
+
+ private final HTML service;
+
+ // Options
+ private final URI rootURI;
+ private Optional<URI> stylesheetURI = Optional.absent();
+ private boolean refreshStylesheet = false;
+ private boolean tryEmbeddedStylesheet = false;
+ private Optional<URI> css = Optional.absent();
+
+ // detected and extracted
+ private enum SourceType {
+ UNKNOWN, XML, AGGREGATION
+ }
+
+ private SourceType sourceType = SourceType.UNKNOWN;
+
+ private final ITextGridRep repository;
+
+ private Optional<String> sid;
+
+ private ObjectType metadata;
+
+ private Optional<URI> associatedStylesheet = Optional.absent();
+
+ private Source source = null;
+
+ private XsltTransformer transformer;
+
+ // Constructor and configuration
+
+ public HTMLWriter(final HTML service, final URI rootURI) {
+ this.service = service;
+ this.rootURI = rootURI;
+ this.repository = service.repository;
+ }
+
+ public HTMLWriter sid(final String sid) {
+ if (sid == null || sid.isEmpty()) {
+ this.sid = Optional.absent();
+ } else {
+ this.sid = Optional.of(sid);
+ }
+ return this;
+ }
+
+ public HTMLWriter stylesheet(final URI uri) {
+ this.stylesheetURI = Optional.fromNullable(uri);
+ return this;
+ }
+
+ public HTMLWriter refresh(final boolean refresh) {
+ this.refreshStylesheet = refresh;
+ return this;
+ }
+
+ public HTMLWriter embedded(final boolean embedded) {
+ this.tryEmbeddedStylesheet = embedded;
+ if (embedded) {
+ this.sourceType = SourceType.XML;
+ }
+ return this;
+ }
+
+ public HTMLWriter css(final URI css) {
+ this.css = Optional.fromNullable(css);
+ return this;
+ }
+
+ protected HTMLWriter loadSource() throws ObjectNotFoundFault,
+ MetadataParseFault,
+ IoFault, ProtocolNotImplementedFault, AuthFault, IOException {
+
+ TGOSupplier<InputStream> content = null;
+ content = service.repository.read(rootURI, sid.orNull());
+ metadata = content.getMetadata();
+ final String format = metadata.getGeneric().getProvided().getFormat();
+ if (format.contains("aggregation")) {
+ sourceType = SourceType.AGGREGATION;
+ } else if (format.matches("^text/.*xml.*$")) {
+ sourceType = SourceType.XML;
+ } else {
+ final String errorMsg = MessageFormat
+ .format("The HTML export can only convert aggregations or XML documents to EPUB, however, the document {0} you referred to has the MIME type {1}.",
+ rootURI, format);
+ throw new WebApplicationException(
+ GenericExceptionMapper.toResponse(
+ Status.UNSUPPORTED_MEDIA_TYPE, errorMsg, ""));
+ }
+
+ if (sourceType == SourceType.AGGREGATION) {
+ final TEICorpusSerializer corpusSerializer = new TEICorpusSerializer(
+ metadata, false, sid.orNull());
+ final FileBackedOutputStream corpusBuffer = new FileBackedOutputStream(
+ 1024 * 1024, true);
+ corpusSerializer.write(corpusBuffer);
+ corpusBuffer.close();
+ this.source = new StreamSource(corpusBuffer.getSupplier()
+ .getInput());
+ } else if (sourceType == SourceType.XML && tryEmbeddedStylesheet) {
+ final FileBackedOutputStream xmlBuffer = new FileBackedOutputStream(
+ 1024 * 1024, true);
+ ByteStreams.copy(content, xmlBuffer);
+ detectEmbeddedStylesheet(xmlBuffer.getSupplier().getInput());
+ this.source = new StreamSource(xmlBuffer.getSupplier().getInput());
+ } else {
+ this.source = new StreamSource(content.getInput(),
+ rootURI.toString());
+ }
+ return this;
+ }
+
+ private void detectEmbeddedStylesheet(final InputStream input) {
+ try {
+ final Source associatedStylesheet = TransformerFactory
+ .newInstance().getAssociatedStylesheet(
+ new StreamSource(input, rootURI.toString()), null,
+ null, null);
+ this.associatedStylesheet = Optional.of(new URI(
+ associatedStylesheet.getSystemId()));
+
+ } catch (final TransformerConfigurationException e) {
+ // TODO Auto-generated catch block
+ e.printStackTrace();
+ } catch (final TransformerFactoryConfigurationError e) {
+ // TODO Auto-generated catch block
+ e.printStackTrace();
+ } catch (final URISyntaxException e) {
+ // TODO Auto-generated catch block
+ e.printStackTrace();
+ }
+ }
+
+ private XsltExecutable getStylesheet() throws SaxonApiException,
+ IOException {
+ if (stylesheetURI.isPresent())
+ return service.getStylesheet(stylesheetURI.get(), sid,
+ refreshStylesheet);
+ else if (associatedStylesheet.isPresent())
+ return service.getStylesheet(associatedStylesheet.get(), sid,
+ refreshStylesheet);
+ else
+ return service.getToHtml();
+ }
+
+ protected HTMLWriter loadStylesheet() throws SaxonApiException,
+ IOException, ObjectNotFoundFault, MetadataParseFault, IoFault,
+ ProtocolNotImplementedFault, AuthFault {
+ if (source == null) {
+ loadSource();
+ }
+
+ final XsltTransformer transformer = getStylesheet().load();
+ if (sid.isPresent()) {
+ transformer.setURIResolver(new TGUriResolver(repository, sid));
+ }
+ transformer.setParameter(new QName("graphicsURLPattern"),
+ new XdmAtomicValue(repository.getCRUDRestEndpoint()
+ + "/@URI@/data"
+ + ((sid == null || "".equals(sid)) ? ""
+ : ("?sessionId=" + sid))));
+ if (css.isPresent()) {
+ transformer.setParameter(new QName("cssFile"), new XdmAtomicValue(
+ css.get()));
+ }
+ this.transformer = transformer;
+ return this;
+ }
+
+ @Override
+ public void write(final OutputStream out) throws IOException,
+ WebApplicationException {
+ try {
+ if (source == null) {
+ loadSource();
+ }
+ if (transformer == null) {
+ loadStylesheet();
+ }
+ transformer
+ .setDestination(service.xsltProcessor.newSerializer(out));
+ transformer.transform();
+
+ } catch (final Exception e) {
+ Throwables.propagateIfPossible(e, IOException.class,
+ WebApplicationException.class);
+ }
+
+ }
+
+}
--
GitLab