import pinyin4j and sparta-xml source code

mymbrooks · Apr 22, 2013 · 83da65e · 83da65e
1 parent 0d43721
commit 83da65e
Show file tree

Hide file tree

Showing 72 changed files with 37,663 additions and 0 deletions.
diff --git a/pom.xml b/pom.xml
@@ -50,6 +50,15 @@
         <maven.compiler.target>${maven.compiler.source}</maven.compiler.target>
     </properties>
 
+    <dependencies>
+        <dependency>
+            <groupId>junit</groupId>
+            <artifactId>junit</artifactId>
+            <version>4.11</version>
+            <scope>test</scope>
+        </dependency>
+    </dependencies>
+
     <build>
         <plugins>
             <plugin>

diff --git a/src/main/java/com/hp/hpl/sparta/BuildDocument.java b/src/main/java/com/hp/hpl/sparta/BuildDocument.java
@@ -0,0 +1,151 @@
+package com.hp.hpl.sparta;
+
+/** This class returns the DOM Document created by parsing XML.
+
+ <blockquote><small> Copyright (C) 2002 Hewlett-Packard Company.
+ This file is part of Sparta, an XML Parser, DOM, and XPath library.
+ This library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public License
+ as published by the Free Software Foundation; either version 2.1 of
+ the License, or (at your option) any later version.  This library
+ is distributed in the hope that it will be useful, but WITHOUT ANY
+ WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE.</small></blockquote>
+ @see <a "href="doc-files/LGPL.txt">GNU Lesser General Public License</a>
+ @version  $Date: 2003/05/12 19:49:29 $  $Revision: 1.3 $
+ @author Sergio Marti
+ */
+
+class BuildDocument implements DocumentSource, ParseHandler {
+
+  public BuildDocument() {
+    this(null);
+  }
+
+  public BuildDocument(ParseLog log) {
+    log_ = (log == null) ? DEFAULT_LOG : log;
+  }
+
+  public void setParseSource(ParseSource ps) {
+    parseSource_ = ps;
+    doc_.setSystemId(ps.toString());
+  }
+
+  public ParseSource getParseSource() {
+    return parseSource_;
+  }
+
+  public String toString() {
+    if (parseSource_ != null)
+      return "BuildDoc: " + parseSource_.toString();
+    else
+      return null;
+  }
+
+  public String getSystemId() {
+    if (parseSource_ != null)
+      return parseSource_.getSystemId();
+    else
+      return null;
+  }
+
+  public int getLineNumber() {
+    if (parseSource_ != null)
+      return parseSource_.getLineNumber();
+    else
+      return -1;
+  }
+
+  /** The parsed document. */
+  public Document getDocument() {
+    return doc_;
+  }
+
+  public void startDocument() {}
+
+  public void endDocument() {
+  /* DEBUG
+     if (currentElement_ != null)
+     log_.warning("EndDocument: currentElement is not null",
+     getSystemId(), getLineNumber());
+   */
+  }
+
+  public void startElement(Element element) {
+    if (currentElement_ == null) {
+      doc_.setDocumentElement(element);
+    } else {
+      currentElement_.appendChild(element);
+    }
+    currentElement_ = element;
+  }
+
+  public void endElement(Element element) {
+    /* DEBUG
+       if (isCENull())
+       return;
+       if (element != currentElement_) {
+       log_.warning("EndElement (" + element.getTagName() +
+       ") does not match currentElement (" +
+       currentElement_.getTagName() + ")", getSystemId(),
+       getLineNumber());
+       return;
+       }
+     */
+
+    currentElement_ = (Element) currentElement_.getParentNode();
+  }
+
+  public void characters(char[] buf, int offset, int len) {
+    /* DEBUG
+       if (isCENull())
+       return;
+     */
+
+    Element element = currentElement_;
+    if (element.getLastChild() instanceof Text) {
+      Text text = (Text) element.getLastChild();
+      text.appendData(buf, offset, len);
+    } else {
+      Text text = new Text(new String(buf, offset, len));
+      element.appendChildNoChecking(text);
+    }
+  }
+
+
+  private final ParseLog log_;
+
+  private Element currentElement_ = null;
+  private final Document doc_ = new Document();
+  private ParseSource parseSource_ = null;
+}
+
+// $Log: BuildDocument.java,v $
+// Revision 1.3  2003/05/12 19:49:29  eobrain
+// Remove unused method.
+//
+// Revision 1.2  2002/10/30 16:40:27  eobrain
+// appendChild no longer throws DOMException
+//
+// Revision 1.1.1.1  2002/08/19 05:04:02  eobrain
+// import from HP Labs internal CVS
+//
+// Revision 1.6  2002/08/18 04:30:54  eob
+// Make class package-private so as not to clutter up the javadoc.
+//
+// Revision 1.5  2002/08/17 19:04:36  eob
+// Add copyright and other formatting and commenting in preparation for
+// release to SourceForge.
+//
+// Revision 1.4  2002/08/15 22:51:37  eob
+// Sparta node constructors no longer needs document
+//
+// Revision 1.3  2002/08/05 20:04:31  sermarti
+//
+// Revision 1.2  2002/08/01 23:29:17  sermarti
+// Much faster Sparta parsing.
+// Has debug features enabled by default. Currently toggled
+// in ParseCharStream.java and recompiled.
+//
+// Revision 1.1  2002/07/25 21:10:15  sermarti
+// Adding files that mysteriously weren't added from Sparta before.
diff --git a/src/main/java/com/hp/hpl/sparta/CharCircBuffer.java b/src/main/java/com/hp/hpl/sparta/CharCircBuffer.java
@@ -0,0 +1,96 @@
+package com.hp.hpl.sparta;
+
+/**
+ * Circular character buffer used to store parsing history for debug
+ * purposes.
+
+ <blockquote><small> Copyright (C) 2002 Hewlett-Packard Company.
+ This file is part of Sparta, an XML Parser, DOM, and XPath library.
+ This library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public License
+ as published by the Free Software Foundation; either version 2.1 of
+ the License, or (at your option) any later version.  This library
+ is distributed in the hope that it will be useful, but WITHOUT ANY
+ WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE.</small></blockquote>
+ @see <a "href="doc-files/LGPL.txt">GNU Lesser General Public License</a>
+ @version  $Date: 2003/07/17 21:36:29 $  $Revision: 1.2 $
+ @author Sergio Marti
+
+ */
+
+
+class CharCircBuffer {
+
+  CharCircBuffer(int n) {
+    buf_ = new int[n];
+  }
+
+  void enable() {
+    enabled_ = true;
+  }
+
+  void disable() {
+    enabled_ = false;
+  }
+
+  void addInt(int i) {
+    addRaw(i + 0x10000);
+  }
+
+  void addChar(char ch) {
+    addRaw(ch);
+  }
+
+  private void addRaw(int v) {
+    if (enabled_) {
+      buf_[next_] = v;
+      next_ = (next_ + 1) % buf_.length;
+      ++total_;
+    }
+  }
+
+  void addString(String s) {
+    char[] chars = s.toCharArray();
+    int slen = chars.length;
+    for (int i = 0; i < slen; ++i)
+      addChar(chars[i]);
+  }
+
+  public String toString() {
+    StringBuffer result = new StringBuffer(11 * buf_.length / 10);
+    int first_ = total_ < buf_.length ? buf_.length - total_ : 0;
+    for (int i = first_; i < buf_.length; ++i) {
+      int ii = (i + next_) % buf_.length;
+      int v = buf_[ii];
+      if (v < 0x10000)
+        result.append((char) v);
+      else
+        result.append(Integer.toString(v - 0x10000));
+    }
+    return result.toString();
+  }
+
+  private final int[] buf_; //Stores either the chars or the integers+0x10000
+  private int next_ = 0;
+  private int total_ = 0;
+  private boolean enabled_ = true;
+}
+
+
+// $Log: CharCircBuffer.java,v $
+// Revision 1.2  2003/07/17 21:36:29  eobrain
+// Use integer arithmetic instead of floating-point arithmetic which is
+// not supported in the J2ME we were using on a Nokia phone.
+//
+// Revision 1.1.1.1  2002/08/19 05:04:02  eobrain
+// import from HP Labs internal CVS
+//
+// Revision 1.2  2002/08/18 04:31:45  eob
+// Add copyright and other formatting and commenting in preparation for
+// release to SourceForge.
+//
+// Revision 1.1  2002/08/01 23:29:17  sermarti
+// Much faster Sparta parsing.
+// Has debug features enabled by default. Currently toggled
+// in ParseCharStream.java and recompiled.
diff --git a/src/main/java/com/hp/hpl/sparta/DOMException.java b/src/main/java/com/hp/hpl/sparta/DOMException.java
@@ -0,0 +1,53 @@
+package com.hp.hpl.sparta;
+
+/**
+ * Thrown when problem constructing the DOM.
+
+ <blockquote><small> Copyright (C) 2002 Hewlett-Packard Company.
+ This file is part of Sparta, an XML Parser, DOM, and XPath library.
+ This library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public License
+ as published by the Free Software Foundation; either version 2.1 of
+ the License, or (at your option) any later version.  This library
+ is distributed in the hope that it will be useful, but WITHOUT ANY
+ WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE.</small></blockquote>
+
+ @see <a "href="doc-files/LGPL.txt">GNU Lesser General Public License</a>
+ @version  $Date: 2002/08/19 05:04:01 $  $Revision: 1.1.1.1 $
+ @see org.w3c.dom.DOMException
+ @author Eamonn O'Brien-Strain
+ */
+public class DOMException extends Exception {
+  public DOMException(short code, String message) {
+    super(message);
+    this.code = code;
+  }
+
+  public short code;
+  //public static final short           INDEX_SIZE_ERR       = 1;
+  public static final short DOMSTRING_SIZE_ERR = 2;
+  public static final short HIERARCHY_REQUEST_ERR = 3;
+  //public static final short           WRONG_DOCUMENT_ERR   = 4;
+  //public static final short           INVALID_CHARACTER_ERR = 5;
+  //public static final short           NO_DATA_ALLOWED_ERR  = 6;
+  //public static final short           NO_MODIFICATION_ALLOWED_ERR = 7;
+  public static final short NOT_FOUND_ERR = 8;
+  //public static final short           NOT_SUPPORTED_ERR    = 9;
+  //public static final short           INUSE_ATTRIBUTE_ERR  = 10;
+
+}
+
+// $Log: DOMException.java,v $
+// Revision 1.1.1.1  2002/08/19 05:04:01  eobrain
+// import from HP Labs internal CVS
+//
+// Revision 1.3  2002/08/18 04:32:20  eob
+// Add copyright and other formatting and commenting in preparation for
+// release to SourceForge.
+//
+// Revision 1.2  2002/05/23 21:02:12  eob
+// Add DOMSTRING_SIZE_ERR code.
+//
+// Revision 1.1  2002/01/04 18:28:11  eob
+// initial
diff --git a/src/main/java/com/hp/hpl/sparta/DefaultParseHandler.java b/src/main/java/com/hp/hpl/sparta/DefaultParseHandler.java
@@ -0,0 +1,58 @@
+package com.hp.hpl.sparta;
+
+/** This class provides a default ParseHandler that does nothing.
+ Users should subclass it and overload the necessary methods.
+
+ <blockquote><small> Copyright (C) 2002 Hewlett-Packard Company.
+ This file is part of Sparta, an XML Parser, DOM, and XPath library.
+ This library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public License
+ as published by the Free Software Foundation; either version 2.1 of
+ the License, or (at your option) any later version.  This library
+ is distributed in the hope that it will be useful, but WITHOUT ANY
+ WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE.</small></blockquote>
+ @see <a "href="doc-files/LGPL.txt">GNU Lesser General Public License</a>
+ @version  $Date: 2002/08/19 05:04:01 $  $Revision: 1.1.1.1 $
+ @author Sergio Marti
+
+ */
+public class DefaultParseHandler implements ParseHandler {
+
+  private ParseSource parseSource_ = null;
+
+  public void setParseSource(ParseSource ps) {
+    parseSource_ = ps;
+  }
+
+  public ParseSource getParseSource() {
+    return parseSource_;
+  }
+
+  public void startDocument() throws ParseException {}
+
+
+  public void endDocument() throws ParseException {}
+
+  public void startElement(Element element) throws ParseException {}
+
+  public void endElement(Element element) throws ParseException {}
+
+  public void characters(char[] buf, int off, int len) throws ParseException {}
+
+}
+
+// $Log: DefaultParseHandler.java,v $
+// Revision 1.1.1.1  2002/08/19 05:04:01  eobrain
+// import from HP Labs internal CVS
+//
+// Revision 1.3  2002/08/18 04:32:40  eob
+// Add copyright and other formatting and commenting in preparation for
+// release to SourceForge.
+//
+// Revision 1.2  2002/08/17 00:54:14  sermarti
+//
+// Revision 1.1  2002/08/01 23:29:17  sermarti
+// Much faster Sparta parsing.
+// Has debug features enabled by default. Currently toggled
+// in ParseCharStream.java and recompiled.