parsetext.java

HTML解析器是一个Java库

	 * and breaking the search at the index specified by <code>breakAtIndex</code>.
	 * <p>
	 * The position specified by <code>breakAtIndex</code> is not included in the search.
	 * <p>
	 * If the search is to continue to the end of the text,
	 * the value {@link #NO_BREAK ParseText.NO_BREAK} should be specified as the <code>breakAtIndex</code>.
	 * <p>
	 * If the specified string is not found then -1 is returned.
	 *
	 * @param searchString  a string.
	 * @param fromIndex  the index to start the search from.
	 * @param breakAtIndex  the index at which to break off the search, or {@link #NO_BREAK} if the search is to continue to the end of the text.
	 * @return the index within this parse text of the first occurrence of the specified string within the specified range, or -1 if the string is not found.
	 */
	public int indexOf(final String searchString, final int fromIndex, final int breakAtIndex) {
		return (searchString.length()==1)
			? indexOf(searchString.charAt(0),fromIndex,breakAtIndex)
			: indexOf(searchString.toCharArray(),fromIndex,breakAtIndex);
	}

	/**
	 * Returns the index within this parse text of the first occurrence of the specified character array,
	 * starting the search at the position specified by <code>fromIndex</code>,
	 * and breaking the search at the index specified by <code>breakAtIndex</code>.
	 * <p>
	 * The position specified by <code>breakAtIndex</code> is not included in the search.
	 * <p>
	 * If the search is to continue to the end of the text,
	 * the value {@link #NO_BREAK ParseText.NO_BREAK} should be specified as the <code>breakAtIndex</code>.
	 * <p>
	 * If the specified character array is not found then -1 is returned.
	 *
	 * @param searchCharArray  a character array.
	 * @param fromIndex  the index to start the search from.
	 * @param breakAtIndex  the index at which to break off the search, or {@link #NO_BREAK} if the search is to continue to the end of the text.
	 * @return the index within this parse text of the first occurrence of the specified character array within the specified range, or -1 if the character array is not found.
	 */
	public int indexOf(final char[] searchCharArray, final int fromIndex, final int breakAtIndex) {
		if (searchCharArray.length==0) return fromIndex;
		final char firstChar=searchCharArray[0];
		final int lastPossibleBreakAtIndex=text.length-searchCharArray.length+1;
		final int actualBreakAtIndex=(breakAtIndex==NO_BREAK || breakAtIndex>lastPossibleBreakAtIndex) ? lastPossibleBreakAtIndex : breakAtIndex;
		outerLoop: for (int i=(fromIndex<0 ? 0 : fromIndex); i<actualBreakAtIndex; i++) {
			if (text[i]==firstChar) {
				for (int j=1; j<searchCharArray.length; j++)
					if (searchCharArray[j]!=text[j+i]) continue outerLoop;
				return i;
			}
		}
		return -1;
	}

	/**
	 * Returns the index within this parse text of the last occurrence of the specified string,
	 * searching backwards starting at the position specified by <code>fromIndex</code>.
	 * <p>
	 * If the specified string is not found then -1 is returned.
	 *
	 * @param searchString  a string.
	 * @param fromIndex  the index to start the search from. 
	 * @return the index within this parse text of the last occurrence of the specified string within the specified range, or -1 if the string is not found.
	 */
	public int lastIndexOf(final String searchString, final int fromIndex) {
		return (searchString.length()==1)
			? lastIndexOf(searchString.charAt(0),fromIndex,NO_BREAK)
			: lastIndexOf(searchString.toCharArray(),fromIndex,NO_BREAK);
	}

	/**
	 * Returns the index within this parse text of the last occurrence of the specified character array,
	 * searching backwards starting at the position specified by <code>fromIndex</code>.
	 * <p>
	 * If the specified character array is not found then -1 is returned.
	 *
	 * @param searchCharArray  a character array.
	 * @param fromIndex  the index to start the search from. 
	 * @return the index within this parse text of the last occurrence of the specified character array within the specified range, or -1 if the character array is not found.
	 */
	public int lastIndexOf(final char[] searchCharArray, final int fromIndex) {
		return lastIndexOf(searchCharArray,fromIndex,NO_BREAK);
	}

	/**
	 * Returns the index within this parse text of the last occurrence of the specified string,
	 * searching backwards starting at the position specified by <code>fromIndex</code>,
	 * and breaking the search at the index specified by <code>breakAtIndex</code>.
	 * <p>
	 * The position specified by <code>breakAtIndex</code> is not included in the search.
	 * <p>
	 * If the search is to continue to the start of the text,
	 * the value {@link #NO_BREAK ParseText.NO_BREAK} should be specified as the <code>breakAtIndex</code>.
	 * <p>
	 * If the specified string is not found then -1 is returned.
	 *
	 * @param searchString  a string.
	 * @param fromIndex  the index to start the search from.
	 * @param breakAtIndex  the index at which to break off the search, or {@link #NO_BREAK} if the search is to continue to the start of the text.
	 * @return the index within this parse text of the last occurrence of the specified string within the specified range, or -1 if the string is not found.
	 */
	public int lastIndexOf(final String searchString, final int fromIndex, final int breakAtIndex) {
		return (searchString.length()==1)
			? lastIndexOf(searchString.charAt(0),fromIndex,breakAtIndex)
			: lastIndexOf(searchString.toCharArray(),fromIndex,breakAtIndex);
	}
	
	/**
	 * Returns the index within this parse text of the last occurrence of the specified character array,
	 * searching backwards starting at the position specified by <code>fromIndex</code>,
	 * and breaking the search at the index specified by <code>breakAtIndex</code>.
	 * <p>
	 * The position specified by <code>breakAtIndex</code> is not included in the search.
	 * <p>
	 * If the search is to continue to the start of the text,
	 * the value {@link #NO_BREAK ParseText.NO_BREAK} should be specified as the <code>breakAtIndex</code>.
	 * <p>
	 * If the specified character array is not found then -1 is returned.
	 *
	 * @param searchCharArray  a character array.
	 * @param fromIndex  the index to start the search from.
	 * @param breakAtIndex  the index at which to break off the search, or {@link #NO_BREAK} if the search is to continue to the start of the text.
	 * @return the index within this parse text of the last occurrence of the specified character array within the specified range, or -1 if the character array is not found.
	 */
	public int lastIndexOf(final char[] searchCharArray, int fromIndex, final int breakAtIndex) {
		if (searchCharArray.length==0) return fromIndex;
		final int rightIndex=text.length-searchCharArray.length;
		if (breakAtIndex>rightIndex) return -1;
		if (fromIndex>rightIndex) fromIndex=rightIndex;
		final int lastCharIndex=searchCharArray.length-1;
		final char lastChar=searchCharArray[lastCharIndex];
		final int actualBreakAtPos=breakAtIndex+lastCharIndex;
		outerLoop: for (int i=fromIndex+lastCharIndex; i>actualBreakAtPos; i--) {
			if (text[i]==lastChar) {
				final int startIndex=i-lastCharIndex;
				for (int j=lastCharIndex-1; j>=0; j--)
					if (searchCharArray[j]!=text[j+startIndex]) continue outerLoop;
				return startIndex;
			}
		}
		return -1;
	}

	/**
	 * Returns the length of the parse text.
	 * @return the length of the parse text.
	 */
	public int length() {
		return text.length;
	}

	/**
	 * Returns a new string that is a substring of this parse text.
	 * <p>
	 * The substring begins at the specified <code>beginIndex</code> and extends to the character at index <code>endIndex</code> - 1.
	 * Thus the length of the substring is <code>endIndex-beginIndex</code>. 
	 *
	 * @param beginIndex  the begin index, inclusive.
	 * @param endIndex  the end index, exclusive.
	 * @return a new string that is a substring of this parse text.
	 */
	public String substring(final int beginIndex, final int endIndex) {
		return new String(text,beginIndex,endIndex-beginIndex);
	}	

	/**
	 * Returns a new character sequence that is a subsequence of this sequence.
	 * <p>
	 * This is equivalent to {@link #substring(int,int) substring(beginIndex,endIndex)}.
	 *
	 * @param beginIndex  the begin index, inclusive.
	 * @param endIndex  the end index, exclusive.
	 * @return a new character sequence that is a subsequence of this sequence.
	 */
	public CharSequence subSequence(final int beginIndex, final int endIndex) {
		return substring(beginIndex,endIndex);
	}

	/**
	 * Returns the content of the parse text as a <code>String</code>.
	 * @return the content of the parse text as a <code>String</code>.
	 */
	public String toString() {
		return new String(text);
	}
}