O(1) \"inside\" 确定的嵌套结构

Answer 1

使用m-ary tree。

有关完整源代码，请参阅the repository。

鉴于：

• 一个Stem标记接口
• 类Token实现Stem
• 类Tree实现Stem

然后是经典 m-ary 树的变体，具有以下有效负载：

• 子树和标记的插入顺序集（标记为Stems）；
• 打开令牌；和
• 关闭令牌。

Tree 源代码类似于 isNested 为 O(1)：

class Tree<T extends Token> implements Stem {
  private final Tree<T> mParent;

  /**
   * Provides O(1) lookup time.
   */
  private final Collection<Stem> mStems = new LinkedHashSet<>( 128 );

  private T mOpening = (T) NONE;
  private T mClosing = (T) NONE;

  public Tree() {
    mParent = null;
  }

  private Tree( final Tree<T> parent, final T opening ) {
    assert parent != null;
    assert opening != null;
    assert opening != NONE;

    mParent = parent;
    mOpening = opening;
  }

  public Tree<T> opening( final T opening ) {
    assert opening != null;

    final var tree = new Tree<>( this, opening );
    mStems.add( tree );

    return tree;
  }

  public Tree<T> closing( final T closing ) {
    assert closing != NONE;
    assert mOpening.isBefore( closing );

    mClosing = closing;

    return mParent == null ? this : mParent;
  }

  public void add( final Stem stem ) {
    assert stem != null;

    mStems.add( stem );
  }

  public boolean isNested( final Stem stem ) {
    return mStems.contains( stem );
  }

  public boolean isBalanced() {
    return
      mOpening.isType( QUOTE_OPENING_DOUBLE ) &&
        mClosing.isType( QUOTE_CLOSING_DOUBLE ) ||
        mOpening.isType( QUOTE_OPENING_SINGLE ) &&
          mClosing.isType( QUOTE_CLOSING_SINGLE );
  }

  public Tree<T> parent() {
    return mParent;
  }

  public String toXml() {
    final var sb = new StringBuilder( 128 );
    final var name = parent() == null ? "root" : "tree";

    sb.append( '<' );
    sb.append( name );
    sb.append( '>' );

    if( !mOpening.isType( TokenType.NONE ) ) {
      sb.append( mOpening.toXml() );
    }

    mStems.forEach( stem -> sb.append( stem.toXml() ) );

    if( !mClosing.isType( TokenType.NONE ) ) {
      sb.append( mClosing.toXml() );
    }

    sb.append( "</" );
    sb.append( name );
    sb.append( '>' );

    return sb.toString();
  }
}

给定一组引号标记，例如：

• QUOTE_OPENING_SINGLE("开单")
• QUOTE_OPENING_DOUBLE("开双")
• QUOTE_CLOSING_SINGLE("收尾单")
• QUOTE_CLOSING_DOUBLE（“关闭双”）
• QUOTE_APO​​STROPHE("撇号")
• QUOTE_AMBIGUOUS_LEADING("开头不明确")
• QUOTE_AMBIGUOUS_LAGGING("关闭-模糊")
• 模棱两可（“模棱两可”）
• 无

使用Tree 如下所示，其中每个Token 实例都由词法分析器发出（未显示）：

  @Override
  public void accept( final Token token ) {
    // Create a new subtree when an opening quotation mark is found.
    if( token.isType( QUOTE_OPENING_SINGLE ) ||
      token.isType( QUOTE_OPENING_DOUBLE ) ) {
      mTree = mTree.opening( token );
    }
    // Close the subtree if it was open, try to close it.
    else if( token.isType( QUOTE_CLOSING_SINGLE ) ||
      token.isType( QUOTE_CLOSING_DOUBLE ) ) {
      mTree = mTree.closing( token );
    }
    // Add any ambiguous tokens to the subtree, which are resolved after
    // the in-memory AST is built.
    else if( token.isAmbiguous() ) {
      mTree.add( token );
    }
  }

这产生了一个嵌套的树结构，其中每个嵌套级别对应于一个引用中的一个引用（只要有可能识别）。

示例输出：

“她说，‘那是山姆的’，”山姆的猫说。

<?xml version="1.0" encoding="UTF-8"?>
<root>
   <tree>
      <opening-double type="QUOTE_OPENING_DOUBLE" began="0" ended="1" />
      <tree>
         <opening-single type="QUOTE_OPENING_SINGLE" began="11" ended="12" />
         <closing-ambiguous type="QUOTE_AMBIGUOUS_LAGGING" began="24" ended="25" />
         <closing-double type="QUOTE_CLOSING_DOUBLE" began="26" ended="27" />
      </tree>
      <closing-ambiguous type="QUOTE_AMBIGUOUS_LAGGING" began="41" ended="42" />
   </tree>
</root>

示例输出：

“A”、“B”和“C”是字母。

<?xml version="1.0" encoding="UTF-8"?>
<root>
   <opening-ambiguous type="QUOTE_AMBIGUOUS_LEADING" began="0" ended="1" />
   <tree>
      <opening-single type="QUOTE_OPENING_SINGLE" began="5" ended="6" />
      <closing-single type="QUOTE_CLOSING_SINGLE" began="7" ended="8" />
   </tree>
   <tree>
      <opening-single type="QUOTE_OPENING_SINGLE" began="14" ended="15" />
      <closing-single type="QUOTE_CLOSING_SINGLE" began="16" ended="17" />
   </tree>
   <closing-single type="QUOTE_CLOSING_SINGLE" began="2" ended="3" />
</root>

示例输出：

“没有多少同胞，我想，小姐；我所知道的——我的老主人，作为一个知道战争的人，曾经说过，他说，‘如果我播种小麦而没有卤汁’ ，我是个荷兰人，”他说，“那场战争就像荷兰人在打一场傻瓜或隔壁的战争一样。不，不，我不会因为荷兰人而打扰我的。还有傻瓜，一个流氓伊诺，我不会为他们看书。”

<?xml version="1.0" encoding="UTF-8"?>
<root>
   <tree>
      <opening-double type="QUOTE_OPENING_DOUBLE" began="0" ended="1" />
      <tree>
         <opening-single type="QUOTE_OPENING_SINGLE" began="115" ended="116" />
         <closing-single type="QUOTE_CLOSING_SINGLE" began="170" ended="171" />
      </tree>
      <closing-ambiguous type="QUOTE_AMBIGUOUS_LAGGING" began="362" ended="363" />
      <closing-double type="QUOTE_CLOSING_DOUBLE" began="378" ended="379" />
   </tree>
</root>