Drupal: číštění HTML – Tidy.
3 * see AUTHORS for the list of contributors
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.
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.
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/>.
18 package org.sonews.storage;
20 import java.io.BufferedReader;
21 import java.io.ByteArrayOutputStream;
22 import java.io.IOException;
23 import java.io.InputStream;
24 import java.io.InputStreamReader;
25 import java.io.PrintStream;
26 import java.io.StringReader;
27 import java.io.StringWriter;
28 import java.io.UnsupportedEncodingException;
29 import java.sql.ResultSet;
30 import java.sql.SQLException;
31 import java.util.ArrayList;
32 import java.util.Date;
33 import java.util.Enumeration;
34 import java.util.logging.Level;
35 import java.util.logging.Logger;
36 import javax.mail.Header;
37 import javax.mail.MessagingException;
38 import javax.mail.Multipart;
39 import javax.mail.Session;
40 import javax.mail.internet.InternetAddress;
41 import javax.mail.internet.MimeBodyPart;
42 import javax.mail.internet.MimeMessage;
43 import javax.mail.internet.MimeMultipart;
44 import javax.xml.transform.Transformer;
45 import javax.xml.transform.TransformerFactory;
46 import javax.xml.transform.stream.StreamResult;
47 import javax.xml.transform.stream.StreamSource;
48 import org.sonews.util.io.Resource;
51 * This is MimeMessage which enables custom Message-ID header
52 * (this header will not be overwritten by the default one like in MimeMessage).
54 * Also add header and body separate serialization.
56 * And can be deserialized from SQL ResultSet
58 * @author František Kučera (frantovo.cz)
60 public class DrupalMessage extends MimeMessage {
62 private static final Logger log = Logger.getLogger(DrupalMessage.class.getName());
63 private static final String MESSAGE_ID_HEADER = "Message-ID";
64 private static final String CRLF = "\r\n";
65 public static final String CHARSET = "UTF-8";
66 private static final String XHTML_CONTENT_TYPE = "text/html; charset=" + CHARSET;
67 private String messageID;
70 * Constructs MIME message from SQL result.
71 * @param rs ResultSet containing message data. No {@link ResultSet#next()} will be called, just values from current row will be read.
72 * @param constructBody true if whole message should be constructed | false if we need only message headers (body will be dummy).
74 public DrupalMessage(ResultSet rs, String myDomain, boolean constructBody) throws SQLException, UnsupportedEncodingException, MessagingException {
75 super(Session.getDefaultInstance(System.getProperties()));
77 addHeader("Message-id", constructMessageId(rs.getInt("id"), rs.getInt("group_id"), rs.getString("group_name"), myDomain));
78 addHeader("Newsgroups", rs.getString("group_name"));
79 setFrom(new InternetAddress(rs.getString("sender_email"), rs.getString("sender_name")));
80 setSubject(rs.getString("subject"));
81 setSentDate(new Date(rs.getLong("created")));
83 int parentID = rs.getInt("parent_id");
85 String parentMessageID = constructMessageId(parentID, rs.getInt("group_id"), rs.getString("group_name"), myDomain);
86 addHeader("In-Reply-To", parentMessageID);
87 addHeader("References", parentMessageID);
91 Multipart multipart = new MimeMultipart("alternative");
92 setContent(multipart);
94 /** Plain text part */
95 MimeBodyPart textPart = new MimeBodyPart();
96 multipart.addBodyPart(textPart);
97 textPart.setText(readPlainText(rs));
100 MimeBodyPart htmlPart = new MimeBodyPart();
101 multipart.addBodyPart(htmlPart);
102 htmlPart.setContent(readXhtmlText(rs), XHTML_CONTENT_TYPE);
108 private String readPlainText(ResultSet rs) {
110 * TODO: převést na prostý text
112 return "TODO: obyčejný text";
115 private String readXhtmlText(ResultSet rs) {
117 * TODO: znovupoužívat XSL transformér
120 String originalText = rs.getString("text");
123 * TODO: používat cache, ukládat si vygenerované články
126 * Místo markdownu jen ošetřit:
128 * - nesmyslné entity v odkazech
129 * - neuzavřené značky: br, hr, img
131 String tidyTexy = tidyXhtml("<html><body>" + originalText + "</body></html>");
135 StringReader input = new StringReader(tidyTexy);
136 StringWriter output = new StringWriter(2 * tidyTexy.length());
137 TransformerFactory tf = TransformerFactory.newInstance();
138 Transformer t = tf.newTransformer(new StreamSource(Resource.getAsStream("helpers/mimeXhtmlPart.xsl")));
139 t.setParameter("isRoot", (rs.getInt("parent_id") == 0));
140 t.setParameter("title", rs.getString("subject"));
141 t.setParameter("urlBase", rs.getString("urlBase"));
142 t.setParameter("wwwRead", rs.getString("wwwRead"));
143 t.setParameter("wwwPost", rs.getString("wwwPost"));
144 t.transform(new StreamSource(input), new StreamResult(output));
146 return output.toString();
147 } catch (Exception e) {
149 * TODO: lepší ošetření chyby
151 log.log(Level.WARNING, "Error while transforming article to XHTML", e);
152 return "<html><body><p>Při transformaci příspěvku bohužel došlo k chybě.</p></body></html>";
157 * TODO: refaktorovat, přesunout
159 private static String tidyXhtml(String inputText) throws IOException {
160 Runtime r = Runtime.getRuntime();
161 Process p = r.exec(new String[]{"tidy",
167 "--logical-emphasis", "true",
168 "--show-errors", "0"});
170 PrintStream vstupProcesu = new PrintStream(p.getOutputStream());
171 vstupProcesu.print(inputText);
172 vstupProcesu.close();
174 String outputText = streamToString(p.getInputStream());
180 * TODO: refaktorovat, přesunout
182 private static String streamToString(InputStream proud) throws IOException {
183 StringBuilder výsledek = new StringBuilder();
184 BufferedReader buf = new BufferedReader(new InputStreamReader(proud));
186 String radek = buf.readLine();
190 výsledek.append(radek);
191 výsledek.append("\n");
194 return výsledek.toString();
197 private static String constructMessageId(int articleID, int groupID, String groupName, String domainName) {
198 StringBuilder sb = new StringBuilder();
200 sb.append(articleID);
204 sb.append(groupName);
206 sb.append(domainName);
208 return sb.toString();
212 public void setHeader(String name, String value) throws MessagingException {
213 super.setHeader(name, value);
215 if (MESSAGE_ID_HEADER.equalsIgnoreCase(name)) {
221 public final void addHeader(String name, String value) throws MessagingException {
222 super.addHeader(name, value);
224 if (MESSAGE_ID_HEADER.equalsIgnoreCase(name)) {
230 public void removeHeader(String name) throws MessagingException {
231 super.removeHeader(name);
233 if (MESSAGE_ID_HEADER.equalsIgnoreCase(name)) {
238 public void setMessageID(String messageID) {
239 this.messageID = messageID;
243 protected void updateMessageID() throws MessagingException {
244 if (messageID == null) {
245 super.updateMessageID();
247 setHeader(MESSAGE_ID_HEADER, messageID);
252 * Call {@link #saveChanges()} before this method, if you want all headers including such ones like:
254 * <pre>MIME-Version: 1.0
255 *Content-Type: multipart/alternative;</pre>
257 * @return serialized headers
258 * @throws MessagingException if getAllHeaders() fails
260 public String getHeaders() throws MessagingException {
261 StringBuilder sb = new StringBuilder();
262 for (Enumeration eh = getAllHeaderLines(); eh.hasMoreElements();) {
263 sb.append(eh.nextElement());
266 return sb.toString();
269 public byte[] getBody() throws IOException, MessagingException {
272 ArrayList<String> skipHeaders = new ArrayList<String>();
273 for (Enumeration eh = getAllHeaders(); eh.hasMoreElements();) {
274 Header h = (Header) eh.nextElement();
275 skipHeaders.add(h.getName());
278 ByteArrayOutputStream baos = new ByteArrayOutputStream(1024);
279 writeTo(baos, skipHeaders.toArray(new String[skipHeaders.size()]));
280 return baos.toByteArray();