htmlanalysis.cs

一套很不错的C#代码,和大家一起共同分享

using System;
using System.Collections;

namespace spider
{
	/// <summary>
	/// Html解析器
	/// </summary>
	public class Analysis:AttributeList
	{
		private string HtmlSourse;//html网页
		private int HtmlId;
		public char delim;

		//<a href="http://www.qq.com"></a>
		public string AnaName;//属性名 href
		public string AnaValue;//属性值 http://www.qq.com
		public string HtmlTag;//html 标记 <a></a>
	
		#region 方法
		public bool IsEnd()
		{return (HtmlId>=HtmlSourse.Length);}
		public bool IsBlank(char ch)
		{return ("\t\n\r ".IndexOf(ch)!=-1);}
		public void EatBlank()
		{while(!IsEnd()) 
			{	if(!IsBlank(GetCurrentChar()))
					return;
				HtmlId++;
			}
		}

		//获取当前字符
		public char GetCurrentChar()
		{return GetCurrentChar(0);
		}
		//获取下一段字符
		public char GetCurrentChar(int peek)
		{	if((HtmlId+peek)<HtmlSourse.Length)
				return HtmlSourse[HtmlId+peek];
			else
				return (char)0;
		}
		//获取下一字符
		public char AdvanceCurrentChar()
		{	return HtmlSourse[HtmlId++];
		}
		public void Advance()
		{	HtmlId++;
		}


		public void GetAnaName()
		{
			EatBlank();
			while (!IsEnd()) 
			{	if (IsBlank(GetCurrentChar())||(GetCurrentChar()=='=')||(GetCurrentChar()=='>') )
					break;
				AnaName+=GetCurrentChar();
				HtmlId++;
			}
			EatBlank();
		}
		public void GetAnaValue()
		{
			if ( delim!=0 )
				return;
			if (GetCurrentChar()=='=') 
			{
				HtmlId++;
				EatBlank();
				if ((GetCurrentChar()=='\'')||(GetCurrentChar()=='\"')) 
				{	delim=GetCurrentChar();
					HtmlId++;
					while(GetCurrentChar()!=delim) 
					{
						AnaValue+=GetCurrentChar();
						HtmlId++;
					}
					HtmlId++;
				} 
				else 
				{
					while ( !IsEnd()&&!IsBlank(GetCurrentChar())&&(GetCurrentChar()!='>')) 
					{
						AnaValue+=GetCurrentChar();
						HtmlId++;
					}
				}
				EatBlank();
			}
		}
		public void AddAttribute()
		{
			Attribute a = new Attribute(AnaName,AnaValue,delim);
			Add(a);
		}
		#endregion
		#region 构造函数
		public Analysis()
		{
		
		}
		#endregion
		#region 属性
		public string ParseName
		{
			get 
			{
				return AnaName;
			}

			set 
			{
				AnaName = value;
			}
		}


		public string ParseValue
		{
			get 
			{
				return AnaValue;
			}

			set 
			{
				AnaValue = value;
			}
		}

		public char ParseDelim
		{
			get 
			{
				return delim;
			}

			set 
			{
				delim = value;
			}
		}

		public string Source
		{
			get 
			{
				return HtmlSourse;
			}

			set 
			{
				HtmlSourse = value;
			}
		}
		#endregion
	}



	public class HtmlAnalysis:Analysis
	{	
		#region 方法
		public AttributeList GetTag()
		{
			AttributeList tag = new AttributeList();
			tag.Name = HtmlTag;

			foreach(Attribute x in List)
			{
				tag.Add((Attribute)x.Clone());
			}

			return tag;
		}
		protected void ParseTag()
	 {
		 HtmlTag="";
		 Clear();

		
		 if ((GetCurrentChar()=='!')&&(GetCurrentChar(1)=='-')&&(GetCurrentChar(2)=='-')) 
		 {
			 while ( !IsEnd() ) 
			 {
				 if ((GetCurrentChar()=='-')&&(GetCurrentChar(1)=='-')&&(GetCurrentChar(2)=='>'))
					 break;
				 if (GetCurrentChar()!='\r')
					 HtmlTag+=GetCurrentChar();
				 Advance();
			 }
			 HtmlTag+="--";
			 Advance();
			 Advance();
			 Advance();
			 delim=(char)0;
			 return;
		 }

		
		 while ( !IsEnd() ) 
		 {
			 if ( IsBlank(GetCurrentChar())||(GetCurrentChar()=='>'))
				 break;
			 HtmlTag+=GetCurrentChar();
			 Advance();
		 }

		 EatBlank();

		
		 while (GetCurrentChar()!='>') 
		 {
			 ParseName = "";
			 ParseValue = "";
			 ParseDelim = (char)0;

			 GetAnaName();

			 if ( GetCurrentChar()=='>') 
			 {
				 AddAttribute();
				 break;
			 }
			 
			 GetAnaValue();
			 AddAttribute();
		 }
		 Advance();
	 }


		public char Parse()
		{
			if( GetCurrentChar()=='<' ) 
			{
				Advance();

				char ch=char.ToUpper(GetCurrentChar());
				if ((ch>='A')&&(ch<='Z')||(ch=='!')||(ch=='/')) 
				{
					ParseTag();
					return (char)0;
				} 
				else return(AdvanceCurrentChar());
			} 
			else return(AdvanceCurrentChar());
		}
		#endregion
	
		}
}