C#根據html生成PDF

using System;
using System.Collections.Generic;

using System.Web;
using System.Web.UI;
using System.Web.UI.WebControls;
using iTextSharp.text;
using iTextSharp.text.pdf;
using System.IO;
using iTextSharp.tool.xml;
using System.Text;
using System.Net;

namespace PDFTools
{
    public partial class CreatePdf : System.Web.UI.Page
    {
        string INPATH = System.Web.HttpContext.Current.Server.MapPath("~/bin/temp.html");

        protected void Page_Load(object sender, EventArgs e)
        {
            string path = System.Web.HttpContext.Current.Server.MapPath("~/UploadPdf");
            if (!Directory.Exists(path))
                Directory.CreateDirectory(path);

            WebClient wc = new WebClient();
            //從網址下載Html字串
            wc.Encoding = System.Text.Encoding.UTF8;
            string htmlText = getWebContent();
            byte[] pdfFile = this.ConvertHtmlTextToPDF(htmlText);

            string fileId = "/file_" + DateTime.Now.ToString("yyyyMMddHHmmss") + ".pdf";
            System.IO.File.WriteAllBytes(path + fileId, pdfFile);
        }

        /// <summary>
        /// 判斷是否有亂碼
        /// </summary>
        /// <param name="txt"></param>
        /// <returns></returns>
        public bool isMessyCode(string txt)
        {
            var bytes = Encoding.UTF8.GetBytes(txt);            //239 191 189            
            for (var i = 0; i < bytes.Length; i++)
            {
                if (i < bytes.Length - 3)
                    if (bytes[i] == 239 && bytes[i + 1] == 191 && bytes[i + 2] == 189)
                    {
                        return true;
                    }
            }
            return false;
        }
        /// <summary>
        /// 獲取網站內容，包含了 HTML+CSS+JS
        /// </summary>
        /// <returns>String返回網頁信息</returns>
        public string getWebContent()
        {
            try
            {
                WebClient MyWebClient = new WebClient();
                MyWebClient.Credentials = CredentialCache.DefaultCredentials;
                //獲取或設置用於向Internet資源的請求進行身份驗證的網絡憑據
                Byte[] pageData = MyWebClient.DownloadData(INPATH);
                //從指定網站下載數據
                string pageHtml = Encoding.UTF8.GetString(pageData);
                //如果獲取網站頁面采用的是GB2312，則使用這句       
                bool isBool = isMessyCode(pageHtml);//判斷使用哪種編碼 讀取網頁信息
                if (!isBool)
                {
                    string pageHtml1 = Encoding.UTF8.GetString(pageData);
                    pageHtml = pageHtml1;
                }
                else
                {
                    string pageHtml2 = Encoding.Default.GetString(pageData);
                    pageHtml = pageHtml2;
                }
                return pageHtml;
            }

            catch (WebException webEx)
            {
                Console.WriteLine(webEx.Message.ToString());
                return webEx.Message;
            }
        }


        /// <summary>
        /// 將Html文字 輸出到PDF檔里
        /// </summary>
        /// <param name="htmlText"></param>
        /// <returns></returns>
        public byte[] ConvertHtmlTextToPDF(string htmlText)
        {
            if (string.IsNullOrEmpty(htmlText))
            {
                return null;
            }
            //避免當htmlText無任何html tag標簽的純文字時，轉PDF時會掛掉，所以一律加上<p>標簽
            htmlText = "<p>" + htmlText + "</p>";

            MemoryStream outputStream = new MemoryStream();//要把PDF寫到哪個串流
            byte[] data = Encoding.UTF8.GetBytes(htmlText);//字串轉成byte[]
            MemoryStream msInput = new MemoryStream(data);
            Document doc = new Document();//要寫PDF的文件，建構子沒填的話預設直式A4
            PdfWriter writer = PdfWriter.GetInstance(doc, outputStream);
            //指定文件預設開檔時的縮放為100%

            PdfDestination pdfDest = new PdfDestination(PdfDestination.XYZ, 0, doc.PageSize.Height, 1f);
            //開啟Document文件 
            doc.Open();

            //使用XMLWorkerHelper把Html parse到PDF檔里
            // XMLWorkerHelper.GetInstance().ParseXHtml(writer, doc, msInput, null, Encoding.UTF8, new UnicodeFontFactory());
            XMLWorkerHelper.GetInstance().ParseXHtml(writer, doc, msInput, null, Encoding.UTF8);

            //將pdfDest設定的資料寫到PDF檔
            PdfAction action = PdfAction.GotoLocalPage(1, pdfDest, writer);
            writer.SetOpenAction(action);
            doc.Close();
            msInput.Close();
            outputStream.Close();
            //回傳PDF檔案 
            return outputStream.ToArray();

        }

        //設置字體類
        public class UnicodeFontFactory : FontFactoryImp
        {
            private static readonly string arialFontPath = Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.Fonts),
      "arialuni.ttf");//arial unicode MS是完整的unicode字型。
            private static readonly string 標楷體Path = Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.Fonts),
            "KAIU.TTF");//標楷體


            public override Font GetFont(string fontname, string encoding, bool embedded, float size, int style, BaseColor color, bool cached)
            {
                BaseFont bfChiness = BaseFont.CreateFont(@"C:\Windows\Fonts\SIMSUN.TTC,1", BaseFont.IDENTITY_H, BaseFont.NOT_EMBEDDED);
                //可用Arial或標楷體，自己選一個
                BaseFont baseFont = BaseFont.CreateFont(標楷體Path, BaseFont.IDENTITY_H, BaseFont.EMBEDDED);
                return new Font(bfChiness, size, style, color);
            }
        }
    }
}