src/org/sonews/storage/DrupalMessage.java
author František Kučera <franta-hg@frantovo.cz>
Thu Oct 20 10:50:58 2011 +0200 (2011-10-20)
changeset 103 a788bf0e1080
parent 102 d843b4fee5dc
child 104 b4c8a2760d6f
permissions -rw-r--r--
Drupal: zpráva od uživatele se před uložením prožene přes XSLT případně Tidy.
     1 /*
     2  *   SONEWS News Server
     3  *   see AUTHORS for the list of contributors
     4  *
     5  *   This program is free software: you can redistribute it and/or modify
     6  *   it under the terms of the GNU General Public License as published by
     7  *   the Free Software Foundation, either version 3 of the License, or
     8  *   (at your option) any later version.
     9  *
    10  *   This program is distributed in the hope that it will be useful,
    11  *   but WITHOUT ANY WARRANTY; without even the implied warranty of
    12  *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    13  *   GNU General Public License for more details.
    14  *
    15  *   You should have received a copy of the GNU General Public License
    16  *   along with this program.  If not, see <http://www.gnu.org/licenses/>.
    17  */
    18 package org.sonews.storage;
    19 
    20 import java.io.BufferedReader;
    21 import java.io.ByteArrayInputStream;
    22 import java.io.ByteArrayOutputStream;
    23 import java.io.IOException;
    24 import java.io.InputStream;
    25 import java.io.InputStreamReader;
    26 import java.io.PrintStream;
    27 import java.io.StringReader;
    28 import java.io.StringWriter;
    29 import java.io.UnsupportedEncodingException;
    30 import java.sql.ResultSet;
    31 import java.sql.SQLException;
    32 import java.util.ArrayList;
    33 import java.util.Date;
    34 import java.util.Enumeration;
    35 import java.util.logging.Level;
    36 import java.util.logging.Logger;
    37 import javax.mail.Header;
    38 import javax.mail.MessagingException;
    39 import javax.mail.Multipart;
    40 import javax.mail.Session;
    41 import javax.mail.internet.InternetAddress;
    42 import javax.mail.internet.MimeBodyPart;
    43 import javax.mail.internet.MimeMessage;
    44 import javax.mail.internet.MimeMultipart;
    45 import javax.xml.transform.Transformer;
    46 import javax.xml.transform.TransformerException;
    47 import javax.xml.transform.TransformerFactory;
    48 import javax.xml.transform.stream.StreamResult;
    49 import javax.xml.transform.stream.StreamSource;
    50 import org.sonews.daemon.NNTPConnection;
    51 import org.sonews.util.io.Resource;
    52 
    53 /**
    54  * This is MimeMessage which enables custom Message-ID header
    55  * (this header will not be overwritten by the default one like in MimeMessage).
    56  * 
    57  * Also add header and body separate serialization.
    58  * 
    59  * And can be deserialized from SQL ResultSet
    60  * 
    61  * @author František Kučera (frantovo.cz)
    62  */
    63 public class DrupalMessage extends MimeMessage {
    64 
    65 	private static final Logger log = Logger.getLogger(DrupalMessage.class.getName());
    66 	private static final String MESSAGE_ID_HEADER = "Message-ID";
    67 	private static final String CRLF = "\r\n";
    68 	public static final String CHARSET = "UTF-8";
    69 	private static final String XHTML_CONTENT_TYPE = "text/html; charset=" + CHARSET;
    70 	private static final String ZNAKČKA_KONCE_ŘÁDKU = "◆";
    71 	private String messageID;
    72 	private Long parentID;
    73 	private Long groupID;
    74 
    75 	/**
    76 	 * Constructs MIME message from SQL result.
    77 	 * @param rs ResultSet containing message data. No {@link ResultSet#next()} will be called, just values from current row will be read.
    78 	 * @param constructBody true if whole message should be constructed | false if we need only message headers (body will be dummy).
    79 	 */
    80 	public DrupalMessage(ResultSet rs, String myDomain, boolean constructBody) throws SQLException, UnsupportedEncodingException, MessagingException, TransformerException, IOException {
    81 		super(Session.getDefaultInstance(System.getProperties()));
    82 
    83 		groupID = rs.getLong("group_id");
    84 		addHeader("Message-id", constructMessageId(rs.getInt("id"), groupID, rs.getString("group_name"), myDomain));
    85 		addHeader("Newsgroups", rs.getString("group_name"));
    86 		setFrom(new InternetAddress(rs.getString("sender_email"), rs.getString("sender_name")));
    87 		setSubject(rs.getString("subject"));
    88 		setSentDate(new Date(rs.getLong("created")));
    89 
    90 		parentID = rs.getLong("parent_id");
    91 		if (parentID > 0) {
    92 			String parentMessageID = constructMessageId(parentID, rs.getInt("group_id"), rs.getString("group_name"), myDomain);
    93 			addHeader("In-Reply-To", parentMessageID);
    94 			addHeader("References", parentMessageID);
    95 		}
    96 
    97 		if (constructBody) {
    98 			Multipart multipart = new MimeMultipart("alternative");
    99 			setContent(multipart);
   100 
   101 			/** XHTML part */
   102 			MimeBodyPart htmlPart = new MimeBodyPart();
   103 			String xhtmlText = readXhtmlText(
   104 					rs.getString("text"),
   105 					rs.getString("subject"),
   106 					rs.getInt("parent_id"),
   107 					rs.getString("urlBase"),
   108 					rs.getString("wwwRead"),
   109 					rs.getString("wwwPost"));
   110 			htmlPart.setContent(xhtmlText, XHTML_CONTENT_TYPE);
   111 
   112 			/** Plain text part */
   113 			MimeBodyPart textPart = new MimeBodyPart();
   114 			String plainText = readPlainText(rs, xhtmlText);
   115 			textPart.setText(plainText);
   116 			//addHeader("Lines", String.valueOf(plainText.split("\n").length));
   117 
   118 			/**
   119 			 * Thunderbirdu záleží, v jakém pořadí části jsou 
   120 			 * (když je prostý text druhý, html se nezobrazí),
   121 			 * KNode zobrazuje HTML správně, i když je na prvním místě.
   122 			 */
   123 			multipart.addBodyPart(textPart);
   124 			multipart.addBodyPart(htmlPart);
   125 		} else {
   126 			/** empty body, just headers */
   127 			setText("");
   128 		}
   129 	}
   130 
   131 	/**
   132 	 * Constructs MIME message from article posted by user.
   133 	 * @param article article that came through NNTP.
   134 	 * @throws MessagingException 
   135 	 */
   136 	public DrupalMessage(Article article) throws MessagingException {
   137 		super(Session.getDefaultInstance(System.getProperties()), serializeArticle(article));
   138 
   139 		String[] parentHeaders = getHeader("In-Reply-To");
   140 		if (parentHeaders.length == 1) {
   141 			String parentMessageID = parentHeaders[0];
   142 			parentID = parseArticleID(parentMessageID);
   143 			groupID = parseGroupID(parentMessageID);
   144 		} else {
   145 			throw new MessagingException("Message posted by user must have exactly one In-Reply-To header.");
   146 		}
   147 	}
   148 
   149 	private static InputStream serializeArticle(Article a) {
   150 		byte articleHeaders[] = a.getHeaderSource().getBytes();
   151 		byte delimiter[] = (NNTPConnection.NEWLINE + NNTPConnection.NEWLINE).getBytes();
   152 		byte body[] = a.getBody();
   153 
   154 		byte message[] = new byte[articleHeaders.length + delimiter.length + body.length];
   155 
   156 		System.arraycopy(articleHeaders, 0, message, 0, articleHeaders.length);
   157 		System.arraycopy(delimiter, 0, message, articleHeaders.length, delimiter.length);
   158 		System.arraycopy(body, 0, message, articleHeaders.length + delimiter.length, body.length);
   159 
   160 		return new ByteArrayInputStream(message);
   161 	}
   162 
   163 	private String readPlainText(ResultSet rs, String xhtmlText) {
   164 		try {
   165 			TransformerFactory tf = TransformerFactory.newInstance();
   166 			Transformer textTransformer = tf.newTransformer(new StreamSource(Resource.getAsStream("helpers/mimeTextPart.xsl")));
   167 
   168 			StringReader input = new StringReader(xhtmlText);
   169 			StringWriter output = new StringWriter(xhtmlText.length());
   170 			textTransformer.transform(new StreamSource(input), new StreamResult(output));
   171 
   172 			return output.toString();
   173 		} catch (Exception e) {
   174 			/**
   175 			 * TODO: lepší ošetření chyby
   176 			 */
   177 			log.log(Level.WARNING, "Error while transforming article to plain text", e);
   178 			return makeSimpleXHTML("Při transformaci příspěvku bohužel došlo k chybě.");
   179 		}
   180 	}
   181 
   182 	private String readXhtmlText(String text, String subject, long parentId, String urlBase, String wwwRead, String wwwPost) throws TransformerException, IOException {
   183 		/**
   184 		 * TODO: 
   185 		 *		- znovupoužívat XSL transformér
   186 		 *		- používat cache, ukládat si vygenerované články
   187 		 */
   188 		String inputText = makeSimpleXHTML(text);
   189 
   190 		TransformerFactory tf = TransformerFactory.newInstance();
   191 		Transformer paragraphTransformer = tf.newTransformer(new StreamSource(Resource.getAsStream("helpers/mimeXhtmlPart-make-paragraphs.xsl")));
   192 
   193 		String paragraphedText;
   194 		boolean tidyWasUsed = false;
   195 		try {
   196 			StringReader input = new StringReader(inputText);
   197 			StringWriter output = new StringWriter(2 * inputText.length());
   198 			paragraphTransformer.transform(new StreamSource(input), new StreamResult(output));
   199 			paragraphedText = output.toString();
   200 		} catch (Exception e) {
   201 			log.log(Level.FINER, "HTML input was shitty – Tidy had to be called.", e);
   202 			StringReader input = new StringReader(tidyXhtml(inputText));
   203 			StringWriter output = new StringWriter(2 * inputText.length());
   204 			paragraphTransformer.transform(new StreamSource(input), new StreamResult(output));
   205 			paragraphedText = output.toString();
   206 			tidyWasUsed = true;
   207 		}
   208 
   209 		Transformer xhtmlTransformer = tf.newTransformer(new StreamSource(Resource.getAsStream("helpers/mimeXhtmlPart.xsl")));
   210 		xhtmlTransformer.setParameter("isRoot", (parentId == 0));
   211 		xhtmlTransformer.setParameter("title", subject);
   212 		xhtmlTransformer.setParameter("urlBase", urlBase);
   213 		xhtmlTransformer.setParameter("wwwRead", wwwRead);
   214 		xhtmlTransformer.setParameter("wwwPost", wwwPost);
   215 		xhtmlTransformer.setParameter("headComment", String.format("Drupal-NNTP bridge. Transformed: %1$tc. Tidy had to be used: %2$b", new Date(), tidyWasUsed));
   216 		StringReader input = new StringReader(paragraphedText);
   217 		StringWriter output = new StringWriter(2 * paragraphedText.length());
   218 		xhtmlTransformer.transform(new StreamSource(input), new StreamResult(output));
   219 
   220 		return output.toString();
   221 	}
   222 
   223 	/**
   224 	 * Does not parse XML works just with text.
   225 	 * @param body XHTML fragment that should be put between &lt;body&gt; and &lt;/body&gt;
   226 	 * @return simple XHTML document (body wrapped in html and body tags)
   227 	 */
   228 	private static String makeSimpleXHTML(String body) {
   229 		return "<html xmlns=\"http://www.w3.org/1999/xhtml\"><body>" + body + "</body></html>";
   230 	}
   231 
   232 	/**
   233 	 * Does not parse XML works just with text.
   234 	 * @param xhtml whole XHTML page
   235 	 * @return content between &lt;body&gt; and &lt;/body&gt; tags.
   236 	 */
   237 	private static String makeFragmentXHTML(String xhtml) {
   238 		final String startTag = "<body>";
   239 		final String endTag = "</body>";
   240 
   241 		int start = xhtml.indexOf(startTag) + startTag.length();
   242 		int end = xhtml.lastIndexOf(endTag);
   243 
   244 		return xhtml.substring(start, end);
   245 	}
   246 
   247 	/**
   248 	 * TODO: refaktorovat, přesunout
   249 	 */
   250 	private static String tidyXhtml(String inputText) throws IOException {
   251 		/*
   252 		 * Viz https://sourceforge.net/tracker/index.php?func=detail&aid=3424437&group_id=27659&atid=390966
   253 		 *
   254 		 * TODO:
   255 		 *		- použít delší zástupný řetězec, ne jen jeden znak
   256 		 *		- umísťovat ho jen tam, kde už nějaký text je (ne mezi >\s*<)
   257 		 */
   258 		inputText = označKonceŘádků(inputText);
   259 
   260 		Runtime r = Runtime.getRuntime();
   261 		Process p = r.exec(new String[]{"tidy", // http://tidy.sourceforge.net
   262 					"-asxml", // well formed XHTML
   263 					"-numeric", // číselné entity
   264 					"-utf8", // kódování
   265 					"--show-warnings", "false", // žádná varování nás nezajímají
   266 					"--show-errors", "0", // ani chyby
   267 					"--doctype", "omit", // doctype nepotřebujeme (doplníme si případně vlastní v XSLT)
   268 					"--logical-emphasis", "true", // em a strong místo i a b
   269 					"--literal-attributes", "true", // zachovat mezery a konce řádků v atributech
   270 					"--force-output", "true" // neznámé značky zahodíme, vložíme jen jejich obsah
   271 				});
   272 
   273 		PrintStream vstupProcesu = new PrintStream(p.getOutputStream());
   274 		vstupProcesu.print(inputText);
   275 		vstupProcesu.close();
   276 
   277 		String outputText = streamToString(p.getInputStream());
   278 
   279 		outputText = vraťKonceŘádků(outputText);
   280 
   281 		return outputText;
   282 	}
   283 
   284 	private static String označKonceŘádků(String text) {
   285 		text = text.replaceAll(">\\s+<", "> <");
   286 		text = text.replaceAll("\\n", ZNAKČKA_KONCE_ŘÁDKU + "\n");
   287 		return text;
   288 	}
   289 
   290 	private static String vraťKonceŘádků(String text) {
   291 		text = text.replaceAll(ZNAKČKA_KONCE_ŘÁDKU + "\\n", "\n");
   292 		text = text.replaceAll(ZNAKČKA_KONCE_ŘÁDKU, "\n");
   293 		return text;
   294 	}
   295 
   296 	/**
   297 	 * TODO: refaktorovat, přesunout
   298 	 */
   299 	private static String streamToString(InputStream proud) throws IOException {
   300 		StringBuilder výsledek = new StringBuilder();
   301 		BufferedReader buf = new BufferedReader(new InputStreamReader(proud));
   302 		while (true) {
   303 			String radek = buf.readLine();
   304 			if (radek == null) {
   305 				break;
   306 			} else {
   307 				výsledek.append(radek);
   308 				výsledek.append("\n");
   309 			}
   310 		}
   311 		return výsledek.toString();
   312 	}
   313 
   314 	public static String constructMessageId(long articleID, long groupID, String groupName, String domainName) {
   315 		StringBuilder sb = new StringBuilder();
   316 		sb.append("<");
   317 		sb.append(articleID);
   318 		sb.append("-");
   319 		sb.append(groupID);
   320 		sb.append("-");
   321 		sb.append(groupName);
   322 		sb.append("@");
   323 		sb.append(domainName);
   324 		sb.append(">");
   325 		return sb.toString();
   326 	}
   327 
   328 	/**
   329 	 * @return article ID of parent of this message | or null, if this is root article and not reply to another one
   330 	 */
   331 	public Long getParentID() {
   332 		return parentID;
   333 	}
   334 
   335 	/**
   336 	 * @return group ID of this message | or null, if this message is not reply to any other one – which is wrong because we have to know the group
   337 	 */
   338 	public Long getGroupID() {
   339 		return groupID;
   340 	}
   341 
   342 	/**
   343 	 * 
   344 	 * @param messageID &lt;{0}-{1}-{2}@domain.tld&gt; where {0} is nntp_id and {1} is group_id and {2} is group_name
   345 	 * @return array where [0] = nntp_id and [1] = group_id and [2] = group_name or returns null if messageID is invalid
   346 	 */
   347 	private static String[] parseMessageID(String messageID) {
   348 		if (messageID.matches("<[0-9]+\\-[0-9]+\\-[a-z0-9\\.]+@.+>")) {
   349 			return messageID.substring(1).split("@")[0].split("\\-");
   350 		} else {
   351 			return null;
   352 		}
   353 	}
   354 
   355 	public static Long parseArticleID(String messageID) {
   356 		String[] localPart = parseMessageID(messageID);
   357 		if (localPart == null) {
   358 			return null;
   359 		} else {
   360 			return Long.parseLong(localPart[0]);
   361 		}
   362 	}
   363 
   364 	public static Long parseGroupID(String messageID) {
   365 		String[] localPart = parseMessageID(messageID);
   366 		if (localPart == null) {
   367 			return null;
   368 		} else {
   369 			return Long.parseLong(localPart[1]);
   370 			// If needed:
   371 			// parseGroupName() will be same as this method, just with:
   372 			// return localPart[2];
   373 		}
   374 	}
   375 
   376 	@Override
   377 	public void setHeader(String name, String value) throws MessagingException {
   378 		super.setHeader(name, value);
   379 
   380 		if (MESSAGE_ID_HEADER.equalsIgnoreCase(name)) {
   381 			messageID = value;
   382 		}
   383 	}
   384 
   385 	@Override
   386 	public final void addHeader(String name, String value) throws MessagingException {
   387 		super.addHeader(name, value);
   388 
   389 		if (MESSAGE_ID_HEADER.equalsIgnoreCase(name)) {
   390 			messageID = value;
   391 		}
   392 	}
   393 
   394 	@Override
   395 	public void removeHeader(String name) throws MessagingException {
   396 		super.removeHeader(name);
   397 
   398 		if (MESSAGE_ID_HEADER.equalsIgnoreCase(name)) {
   399 			messageID = null;
   400 		}
   401 	}
   402 
   403 	public void setMessageID(String messageID) {
   404 		this.messageID = messageID;
   405 	}
   406 
   407 	@Override
   408 	protected void updateMessageID() throws MessagingException {
   409 		if (messageID == null) {
   410 			super.updateMessageID();
   411 		} else {
   412 			setHeader(MESSAGE_ID_HEADER, messageID);
   413 		}
   414 	}
   415 
   416 	/**
   417 	 * Call {@link #saveChanges()} before this method, if you want all headers including such ones like:
   418 	 * 
   419 	 * <pre>MIME-Version: 1.0
   420 	 *Content-Type: multipart/alternative;</pre>
   421 	 * 
   422 	 * @return serialized headers
   423 	 * @throws MessagingException if getAllHeaders() fails
   424 	 */
   425 	public String getHeaders() throws MessagingException {
   426 		StringBuilder sb = new StringBuilder();
   427 		for (Enumeration eh = getAllHeaderLines(); eh.hasMoreElements();) {
   428 			sb.append(eh.nextElement());
   429 			sb.append(CRLF);
   430 		}
   431 		return sb.toString();
   432 	}
   433 
   434 	public byte[] getBody() throws IOException, MessagingException {
   435 		saveChanges();
   436 
   437 		ArrayList<String> skipHeaders = new ArrayList<String>();
   438 		for (Enumeration eh = getAllHeaders(); eh.hasMoreElements();) {
   439 			Header h = (Header) eh.nextElement();
   440 			skipHeaders.add(h.getName());
   441 		}
   442 
   443 		ByteArrayOutputStream baos = new ByteArrayOutputStream(1024);
   444 		writeTo(baos, skipHeaders.toArray(new String[skipHeaders.size()]));
   445 		return baos.toByteArray();
   446 	}
   447 
   448 	/**
   449 	 * Transforms message content to valid XHTML and strips html and body tags.
   450 	 * When receiving message from user through NNTP 
   451 	 * this method is used to get text that should be saved into databse.
   452 	 * @return XHTML fragment – content between &lt;body&gt; and &lt;/body&gt; tags.
   453 	 */
   454 	public String getBodyXhtmlFragment() throws StorageBackendException {
   455 		/**
   456 		 * TODO: podporovat i zprávy přímo v HTML a multipart.
   457 		 */
   458 		try {
   459 			Object c = getContent();
   460 			if (isMimeType("text/plain") && c instanceof String) {
   461 				String xhtml = readXhtmlText(
   462 						(String) c,
   463 						getSubject(),
   464 						getParentID(),
   465 						null,
   466 						null,
   467 						null);
   468 				return makeFragmentXHTML(xhtml);
   469 			} else {
   470 				throw new StorageBackendException("Only text/plain messages are supported for now – post it as plain text please.");
   471 			}
   472 		} catch (Exception e) {
   473 			throw new StorageBackendException(e);
   474 		}
   475 	}
   476 }