publish/DbkServlet.java

   1 package org.argeo.app.servlet.publish;
   2
   3 import java.io.BufferedOutputStream;
   4 import java.io.ByteArrayInputStream;
   5 import java.io.ByteArrayOutputStream;
   6 import java.io.File;
   7 import java.io.FileOutputStream;
   8 import java.io.IOException;
   9 import java.io.InputStream;
  10 import java.io.OutputStream;
  11 import java.net.URI;
  12 import java.net.URLDecoder;
  13 import java.nio.charset.StandardCharsets;
  14 import java.nio.file.Files;
  15 import java.nio.file.Paths;
  16 import java.util.Collections;
  17 import java.util.HashMap;
  18 import java.util.Map;
  19
  20 import javax.jcr.Node;
  21 import javax.jcr.Repository;
  22 import javax.jcr.RepositoryException;
  23 import javax.jcr.Session;
  24 import javax.jcr.nodetype.NodeType;
  25 import javax.servlet.ServletException;
  26 import javax.servlet.http.HttpServlet;
  27 import javax.servlet.http.HttpServletRequest;
  28 import javax.servlet.http.HttpServletResponse;
  29 import javax.xml.parsers.DocumentBuilder;
  30 import javax.xml.parsers.DocumentBuilderFactory;
  31 import javax.xml.transform.Result;
  32 import javax.xml.transform.Source;
  33 import javax.xml.transform.Templates;
  34 import javax.xml.transform.Transformer;
  35 import javax.xml.transform.TransformerConfigurationException;
  36 import javax.xml.transform.TransformerFactory;
  37 import javax.xml.transform.dom.DOMSource;
  38 import javax.xml.transform.sax.SAXResult;
  39 import javax.xml.transform.stream.StreamResult;
  40 import javax.xml.transform.stream.StreamSource;
  41
  42 import org.apache.commons.io.IOUtils;
  43 import org.apache.fop.apps.Fop;
  44 import org.apache.fop.apps.FopFactory;
  45 import org.apache.xalan.processor.TransformerFactoryImpl;
  46 import org.argeo.api.cms.CmsLog;
  47 import org.argeo.api.cms.ux.CmsTheme;
  48 import org.argeo.app.docbook.DbkType;
  49 import org.argeo.app.docbook.DbkUtils;
  50 import org.argeo.cms.auth.RemoteAuthUtils;
  51 import org.argeo.cms.servlet.ServletHttpRequest;
  52 import org.argeo.jcr.Jcr;
  53 import org.argeo.jcr.JcrException;
  54 import org.argeo.jcr.JcrUtils;
  55 import org.w3c.dom.Document;
  56
  57 /**
  58  * A servlet transforming a dbk:* JCR node into HTML, using the DocBook XSL.
  59  */
  60 public class DbkServlet extends HttpServlet {
  61         private static final long serialVersionUID = 6906020513498289335L;
  62
  63         private CmsLog log = CmsLog.getLog(DbkServlet.class);
  64
  65         private Repository repository;
  66
  67         private DocumentBuilderFactory documentBuilderFactory;
  68         private TransformerFactory transformerFactory;
  69         private Templates docBoookHtmlTemplates;
  70
  71         // FOP
  72         private Templates docBoookFoTemplates;
  73
  74         private Map<String, CmsTheme> themes = Collections.synchronizedMap(new HashMap<>());
  75
  76         private String xslBase;
  77
  78         @Override
  79         protected void doGet(HttpServletRequest req, HttpServletResponse resp) throws ServletException, IOException {
  80
  81                 String pathInfo = req.getPathInfo();
  82                 if (pathInfo.startsWith("//"))
  83                         pathInfo = pathInfo.substring(1);
  84                 String path = URLDecoder.decode(pathInfo, StandardCharsets.UTF_8);
  85
  86                 if (path.toLowerCase().endsWith(".css")) {
  87                         path = path.substring(1);
  88                         int firstSlash = path.indexOf('/');
  89                         String themeId = path.substring(0, firstSlash);
  90                         String cssPath = path.substring(firstSlash);
  91                         CmsTheme cmsTheme = themes.get(themeId);
  92                         if (cmsTheme == null)
  93                                 throw new IllegalArgumentException("Theme " + themeId + " not found.");
  94                         resp.setContentType("text/css");
  95                         IOUtils.copy(cmsTheme.getResourceAsStream(cssPath), resp.getOutputStream());
  96                         return;
  97                 }
  98
  99                 if (path.toLowerCase().endsWith("/index.html")) {
 100                         path = path.substring(0, path.length() - "/index.html".length());
 101                 }
 102
 103                 boolean pdf = false;
 104                 if (path.toLowerCase().endsWith(".pdf")) {
 105                         pdf = true;
 106                         if (path.toLowerCase().endsWith("/index.pdf")) {
 107                                 path = path.substring(0, path.length() - "/index.pdf".length());
 108                         }
 109                 }
 110
 111                 Session session = null;
 112                 try {
 113                         session = RemoteAuthUtils.doAs(() -> Jcr.login(repository, null), new ServletHttpRequest(req));
 114                         Node node = session.getNode(path);
 115
 116                         if (node.hasNode(DbkType.article.get())) {
 117                                 Node dbkNode = node.getNode(DbkType.article.get());
 118                                 if (DbkUtils.isDbk(dbkNode)) {
 119                                         CmsTheme cmsTheme = null;
 120                                         String themeId = req.getParameter("themeId");
 121                                         if (themeId != null) {
 122                                                 cmsTheme = themes.get(themeId);
 123                                                 if (cmsTheme == null)
 124                                                         throw new IllegalArgumentException("Theme " + themeId + " not found.");
 125                                         }
 126
 127                                         // TODO customise DocBook so that it outputs UTF-8
 128                                         // see http://www.sagehill.net/docbookxsl/OutputEncoding.html
 129                                         resp.setContentType("text/html; charset=ISO-8859-1");
 130
 131                                         // TODO optimise with pipes, SAX, etc. ?
 132                                         byte[] arr;
 133                                         try (ByteArrayOutputStream out = new ByteArrayOutputStream()) {
 134                                                 session.exportDocumentView(dbkNode.getPath(), out, true, false);
 135                                                 arr = out.toByteArray();
 136                                         } catch (RepositoryException e) {
 137                                                 throw new JcrException(e);
 138                                         }
 139
 140                                         try (InputStream in = new ByteArrayInputStream(arr);) {
 141
 142                                                 DocumentBuilder docBuilder = documentBuilderFactory.newDocumentBuilder();
 143                                                 Document doc = docBuilder.parse(in);
 144                                                 Source xmlInput = new DOMSource(doc);
 145
 146                                                 if (pdf) {
 147                                                         // String baseUri = req.getRequestURI();
 148                                                         FopFactory fopFactory = FopFactory.newInstance(URI.create(req.getRequestURL().toString()));
 149                                                         resp.setContentType("application/pdf");
 150
 151 //                                                      // DocBook to FO
 152 //                                                      byte[] foBytes;
 153 //                                                      try (ByteArrayOutputStream out = new ByteArrayOutputStream();) {
 154 //                                                              Result xmlOutput = new StreamResult(out);
 155 //                                                              Transformer docBookTransformer = docBoookFoTemplates.newTransformer();
 156 //                                                              docBookTransformer.transform(xmlInput, xmlOutput);
 157 //                                                              foBytes = out.toByteArray();
 158 //                                                      }
 159 //
 160 //                                                      // FO to PDF
 161 //                                                      try (InputStream foIn = new ByteArrayInputStream(foBytes)) {
 162 //                                                              Fop fop = fopFactory.newFop("application/pdf", resp.getOutputStream());
 163 //                                                              Transformer fopTransformer = transformerFactory.newTransformer(); // identity
 164 //                                                              Source src = new StreamSource(foIn);
 165 //                                                              Result fopResult = new SAXResult(fop.getDefaultHandler());
 166 //                                                              fopTransformer.transform(src, fopResult);
 167 //                                                      }
 168 //
 169
 170                                                         Fop fop = fopFactory.newFop("application/pdf", resp.getOutputStream());
 171                                                         Transformer docBookTransformer = getDocBoookFoTemplates().newTransformer();
 172                                                         Result fopResult = new SAXResult(fop.getDefaultHandler());
 173                                                         docBookTransformer.transform(xmlInput, fopResult);
 174
 175                                                 } else {
 176                                                         Result xmlOutput = new StreamResult(resp.getOutputStream());
 177                                                         resp.setContentType("text/html");
 178                                                         Transformer docBookTransformer = getDocBoookHtmlTemplates().newTransformer();
 179
 180                                                         // gather CSS
 181                                                         if (cmsTheme != null) {
 182                                                                 StringBuilder sb = new StringBuilder();
 183                                                                 for (String cssPath : cmsTheme.getWebCssPaths()) {
 184                                                                         sb.append(req.getContextPath()).append(req.getServletPath()).append('/');
 185                                                                         sb.append(themeId).append('/').append(cssPath).append(' ');
 186                                                                 }
 187                                                                 // FIXME make it more generic
 188                                                                 sb.append("https://fonts.googleapis.com/css2?family=Roboto:wght@400;700&display=swap")
 189                                                                                 .append(' ');
 190                                                                 sb.append(
 191                                                                                 "https://fonts.googleapis.com/css2?family=Montserrat:ital,wght@0,300;0,400;0,600;1,400&display=swap")
 192                                                                                 .append(' ');
 193                                                                 if (sb.length() > 0)
 194                                                                         docBookTransformer.setParameter("html.stylesheet", sb.toString());
 195                                                         }
 196                                                         docBookTransformer.transform(xmlInput, xmlOutput);
 197                                                 }
 198 //                              resp.getOutputStream().write(out.toByteArray());
 199                                         } catch (Exception e) {
 200                                                 throw new ServletException("Cannot transform " + path, e);
 201                                         }
 202                                 }
 203                         } else {
 204                                 if (node.isNodeType(NodeType.NT_FILE)) {// media download etc.
 205                                         String fileNameLowerCase = node.getName().toLowerCase();
 206                                         if (fileNameLowerCase.endsWith(".jpg") || fileNameLowerCase.endsWith(".jpeg")) {
 207                                                 resp.setContentType("image/jpeg");
 208                                         } else if (fileNameLowerCase.endsWith(".png")) {
 209                                                 resp.setContentType("image/png");
 210                                         } else if (fileNameLowerCase.endsWith(".gif")) {
 211                                                 resp.setContentType("image/gif");
 212                                         } else if (fileNameLowerCase.endsWith(".svg")) {
 213                                                 resp.setContentType("image/svg+xml");
 214                                         } else {
 215                                                 // TODO know more content types...
 216                                                 resp.setHeader("Content-Disposition", "attachment; filename=\"" + node.getName() + "\"");
 217                                         }
 218                                         IOUtils.copy(JcrUtils.getFileAsStream(node), resp.getOutputStream());
 219                                 } else {
 220                                         throw new IllegalArgumentException("Unsupported node " + node);
 221                                 }
 222                         }
 223                 } catch (RepositoryException e1) {
 224                         throw new JcrException(e1);
 225                 } finally {
 226                         Jcr.logout(session);
 227                 }
 228         }
 229
 230         @Override
 231         public void init() throws ServletException {
 232
 233                 // TODO improve configuration and provisioning of DocBook XSL
 234                 xslBase = System.getProperty("argeo.docbook.xsl");
 235                 if (xslBase == null) {
 236                         // We need namespace aware XSL!
 237                         // Fedora (sudo dnf install docbook5-style-xsl)
 238                         String defaultXslBase = "/usr/share/xml/docbook/stylesheet/docbook-xsl-ns/";
 239                         if (!Files.exists(Paths.get(defaultXslBase))) {
 240                                 defaultXslBase = "/opt/docbook-xsl";
 241                                 if (!Files.exists(Paths.get(defaultXslBase))) {
 242                                         log.error("System property argeo.docbook.xsl is not set and default location " + defaultXslBase
 243                                                         + " does not exist.");
 244                                 }
 245                         }
 246                         xslBase = defaultXslBase;
 247
 248                 }
 249
 250                 documentBuilderFactory = DocumentBuilderFactory.newInstance();
 251                 documentBuilderFactory.setXIncludeAware(true);
 252                 documentBuilderFactory.setNamespaceAware(true);
 253
 254         }
 255
 256         protected Templates createDocBookTemplates(String xsl) {
 257                 try {
 258                         if (transformerFactory == null) {
 259                                 // We must explicitly use the non-XSLTC transformer, as XSLTC is not working
 260                                 // with DocBook stylesheets
 261                                 transformerFactory = new TransformerFactoryImpl();
 262                         }
 263                         Source xslSource = new StreamSource(xsl);
 264                         Templates templates = transformerFactory.newTemplates(xslSource);
 265                         if (templates == null)
 266                                 throw new IllegalStateException("Could not instantiate XSL " + xsl);
 267                         return templates;
 268                 } catch (TransformerConfigurationException e) {
 269                         throw new IllegalStateException("Cannot instantiate XSL " + xsl, e);
 270                 }
 271
 272         }
 273
 274         protected Templates getDocBoookHtmlTemplates() {
 275                 if (docBoookHtmlTemplates == null) {
 276                         String htmlXsl = xslBase + "/html/docbook.xsl";
 277                         docBoookHtmlTemplates = createDocBookTemplates(htmlXsl);
 278                 }
 279                 return docBoookHtmlTemplates;
 280         }
 281
 282         protected Templates getDocBoookFoTemplates() {
 283                 if (docBoookFoTemplates == null) {
 284                         String foXsl = xslBase + "/fo/docbook.xsl";
 285                         docBoookFoTemplates = createDocBookTemplates(foXsl);
 286                 }
 287                 return docBoookFoTemplates;
 288         }
 289
 290         public void setRepository(Repository repository) {
 291                 this.repository = repository;
 292         }
 293
 294         public void addTheme(CmsTheme theme, Map<String, String> properties) {
 295                 themes.put(theme.getThemeId(), theme);
 296         }
 297
 298         public void removeTheme(CmsTheme theme, Map<String, String> properties) {
 299                 themes.remove(theme.getThemeId());
 300         }
 301
 302         public static void main(String[] args) throws Exception {
 303                 // Step 1: Construct a FopFactory by specifying a reference to the configuration
 304                 // file
 305                 // (reuse if you plan to render multiple documents!)
 306                 FopFactory fopFactory = FopFactory.newInstance(new File(".").toURI());
 307
 308                 // Step 2: Set up output stream.
 309                 // Note: Using BufferedOutputStream for performance reasons (helpful with
 310                 // FileOutputStreams).
 311                 OutputStream out = new BufferedOutputStream(new FileOutputStream(new File(
 312                                 System.getProperty("user.home") + "/dev/git/unstable/argeo-qa/doc/platform/argeo-platform-test.pdf")));
 313
 314                 try {
 315                         // Step 3: Construct fop with desired output format
 316                         Fop fop = fopFactory.newFop("application/pdf", out);
 317
 318                         // Step 4: Setup JAXP using identity transformer
 319                         TransformerFactory fopTransformerFactory = TransformerFactory.newInstance();
 320                         Transformer fopTransformer = fopTransformerFactory.newTransformer(); // identity transformer
 321
 322                         // Step 5: Setup input and output for XSLT transformation
 323                         // Setup input stream
 324                         Source src = new StreamSource(new File(
 325                                         System.getProperty("user.home") + "/dev/git/unstable/argeo-qa/doc/platform/argeo-platform.fo"));
 326
 327                         // Resulting SAX events (the generated FO) must be piped through to FOP
 328                         Result res = new SAXResult(fop.getDefaultHandler());
 329
 330                         // Step 6: Start XSLT transformation and FOP processing
 331                         fopTransformer.transform(src, res);
 332
 333                 } finally {
 334                         // Clean-up
 335                         out.close();
 336                 }
 337         }
 338 }