C# 如何解析PDF文件？PDFBox？还是...

[天罡gg]

最近需要用C#解析一些PDF文件
开始用iTextSharp，创建PDF没问题，解析就不容易了；
现在在用PDFBox,只能读取出文字，在网上看到的无非是这三行代码

		private static string parseUsingPDFBox(string input)
		{
           
			PDDocument doc = PDDocument.load(input);
            
			PDFTextStripper stripper = new PDFTextStripper();

			return stripper.getText(doc);
            
		}

除了文字，还有其它很多东西要解析呢！图片，表格等等，如何解析？高手指点一二！

[心灵彩虹]

	private void ParseXRef(int startxref)
		{
			int objNumber = 0;
			int generationNumber;
			int offset;
			PdfCrossReferenceEntry entry;
			int endOfXRef = pdf.IndexOf("trailer", startxref + 4);
			string xRef = pdf.Substring(startxref + 4, endOfXRef - (startxref + 4) + 1);
			MatchCollection refMatches = refRegex.Matches(xRef);

			foreach (Match refMatch in refMatches)
			{
				if (!refMatch.Groups[2].Value.Equals(""))
				{
					objNumber = Int32.Parse(refMatch.Groups[2].Value, CultureInfo.InvariantCulture);
				}

				offset = Int32.Parse(refMatch.Groups[4].Value, CultureInfo.InvariantCulture);
				generationNumber = Int32.Parse(refMatch.Groups[5].Value, CultureInfo.InvariantCulture);

				if (refMatch.Groups[6].Value.Equals("n"))
				{
					entry = new PdfCrossReferenceEntry(objNumber, generationNumber, offset,
						refMatch.Groups[6].Value.Equals("n"));

					if (!objects.Contains(objNumber))
					{
						offsets.Add(offset, entry);
						objects.Add(objNumber, entry);
					}
				}
				else if (objNumber == 0)
				{
					// special case:
					// in order to build a new xref table we need the first free object number
					nullOffset = offset;
				}

				objNumber++;
			}
		}

		/// <summary>
		/// Gets the end offset of the specified object.
		/// The offset is determined by the beginning offset of the object with next higher start offset.
		/// If the object is the last object, -1 is returned.
		/// If the document was updated, the offset may be after the xref table and trailer that
		/// follow the specified object.
		/// </summary>
		/// <param name="objNumber">The object number to find the offset for.</param>
		/// <returns>The end offset of the specified object or -1 if it is the last object</returns>
		private int GetEndOfObject(int objNumber)
		{
			PdfCrossReferenceEntry theObject = (PdfCrossReferenceEntry)objects[objNumber];
			int objectIndex = offsets.IndexOfKey(theObject.Offset);
			
			if ((objectIndex + 1) < offsets.Count)
			{
				PdfCrossReferenceEntry nextObject = (PdfCrossReferenceEntry)offsets.GetByIndex(objectIndex + 1);
				return nextObject.Offset;
			}
			else
			{
				return -1;
			}
		}

		/// <summary>
		/// Parses the Interactive Form Dictionary referenced by the AcroForm entry in
		/// the Document Catalog.
		/// </summary>
		private void ParseAcroForm()
		{
			PdfCrossReferenceEntry documentCatalog = (PdfCrossReferenceEntry)objects[rootObjectNumber];
			int documentCatalogStart = documentCatalog.Offset;
			int documentCatalogEnd = GetEndOfObject(documentCatalog.ObjectNumber);

			PdfDictionary documentCatalogObject = ParseObject(documentCatalogStart, documentCatalogEnd) as PdfDictionary;
			PdfObject acroFormObject = documentCatalogObject["/AcroForm"];

			if (acroFormObject is PdfReference)
			{
				PdfReference acroFormRef = (PdfReference)acroFormObject;

				// extract the AcroForm object
				int acroNumber = acroFormRef.ObjectNumber;
				int acroGeneration = acroFormRef.GenerationNumber;
				int acroStart = ((PdfCrossReferenceEntry)objects[acroNumber]).Offset;
				int acroEnd = GetEndOfObject(acroNumber);

				PdfDictionary formDictionary = (PdfDictionary)ParseObject(acroStart, acroEnd);
				form = new PdfForm(acroNumber, acroGeneration, formDictionary);
			}
		}

		private PdfObject ParseObject(int start, int end)
		{
			Match match;
			if (end < 0)
			{
				match = objRegex.Match(pdf, start);
			}
			else
			{
				match = objRegex.Match(pdf, start, end - start);
			}

			if (match.Success)
			{
				int objNumber = Int32.Parse(match.Groups[1].Value, CultureInfo.InvariantCulture);
				int generationNumber = Int32.Parse(match.Groups[2].Value, CultureInfo.InvariantCulture);
				int endOfMatch = match.Index + match.Length;

				string afterObj;
				if (end < 0)
				{
					afterObj = pdf.Substring(endOfMatch);
				}
				else
				{
					afterObj = pdf.Substring(endOfMatch, end - endOfMatch + 1);
				}
				return PdfObject.GetPdfObject(ref afterObj);
			}
			else
			{
				return null;
			}
		}

		/// <summary>
		/// Writes the PDF out to the specified stream including all updates made after this object was created.
		/// </summary>
		/// <param name="stream">The <see cref="System.IO.Stream"/> the PDF will be written to.</param>
		public void WritePdf(Stream stream)
		{
			string lines = pdf + GetUpdate();

			byte[] buffer = new byte[lines.Length];

			for (int i = 0; i < lines.Length; i++)
			{
				buffer[i] = (byte)lines[i];
			}

			stream.Write(buffer, 0, buffer.Length);
		}

		private string GetUpdate()
		{
			if (form != null)
			{
				int offset = pdf.Length;
				string update = "";
				string xref = "xref\n";
				xref += "0 1\n";
				xref += nullOffset.ToString("0000000000", CultureInfo.InvariantCulture) + " 65535 f \n";

				// write the AcroForm object
				string formString = form.ToString();
				update += formString;
				xref += form.ObjectNumber.ToString(CultureInfo.InvariantCulture) + " 1\n";
				xref += offset.ToString("0000000000", CultureInfo.InvariantCulture) + " " + form.GenerationNumber.ToString("00000", CultureInfo.InvariantCulture) + " n \n";
				offset += formString.Length;

				foreach (PdfField field in Fields)
				{
					if (field.HasChanged())
					{
						string fieldString = field.ToString();
						update += fieldString;
						xref += field.ObjectNumber.ToString(CultureInfo.InvariantCulture) + " 1\n";
						xref += offset.ToString("0000000000", CultureInfo.InvariantCulture) + " " + field.GenerationNumber.ToString("00000", CultureInfo.InvariantCulture) + " n \n";
						offset += fieldString.Length;
					}
				}

				string trailer = GetTrailer(offset);

				return update + xref + trailer;
			}
			else
			{
				return "";
			}
		}

		private string GetTrailer(int xrefOffset)
		{
			Hashtable trailerHash = new Hashtable();
			PdfName prevName = new PdfName("/Prev");
			PdfName rootName = new PdfName("/Root");
			PdfName sizeName = new PdfName("/Size");

			trailerHash[prevName] = new PdfNumber(previous);
			trailerHash[rootName] = previousTrailer["/Root"];
			trailerHash[sizeName] = previousTrailer["/Size"];

			PdfDictionary newTrailer = new PdfDictionary(trailerHash);

			string s = "";
			s += "trailer\n";
			s += newTrailer + "\n";			
			s += "startxref\n";
			s += xrefOffset.ToString(CultureInfo.InvariantCulture) + "\n";
			s += "%%EOF\n";

			return s;
		}

		/// <summary>
		/// Gets a collection of all form fields.
		/// </summary>
		public PdfField[] Fields
		{
			get
			{
				return fields;
			}
		}

		/// <summary>
		/// Gets a Hashtable of all form fields keyed by their name.
		/// </summary>
		public Hashtable FieldsByName
		{
			get
			{
				return fieldsByName;
			}
		}

		/// <summary>
		/// Reads the first file on the command line, parses it and writes it to the second file on the command line.
		/// </summary>
		/// <param name="args">Two filenames, the first must be a PDF file, the second will be written to.</param>
		public static void Main(string[] args)
		{
			if (args.Length == 2)
			{
				Console.Error.WriteLine("Usage: " + Environment.GetCommandLineArgs()[0] + " file1 file2");
				Console.Error.WriteLine("Reads file1, parses it as a PDF file, and writes to file2");
			}
			else
			{
				PdfReader reader = PdfReader.GetPdfReader("E:\\Life.pdf");
				
				foreach (PdfField field in reader.Fields)
				{
					Debug.WriteLine(field.FieldName);
				}
				
				FileStream fileStream = new FileStream(@"D:\Downloads\PDFReader_src\PDFReader\samples\myPDF.pdf", System.IO.FileMode.Create);
				reader.WritePdf(fileStream);
				fileStream.Close();
			}
		}
	

[心灵彩虹]

[CLSCompliant(true)]
	public class PdfReader
	{
		private string pdf;
		private PdfField[] fields = new PdfField[0];
		private Hashtable fieldsByName = new Hashtable();
		private PdfForm form;
		private static readonly Regex xrefRegex = new Regex(@"startxref\s*(\d+)\s*%%EOF", RegexOptions.Singleline);
		private static readonly Regex trailerRegex = new Regex(@"trailer\s*<<(.*?)>>", RegexOptions.Singleline);
		private static readonly Regex rootRegex = new Regex(@"/Root\s*(\d+)\s+(\d+)\s*R", RegexOptions.Singleline);
		private static readonly Regex sizeRegex = new Regex(@"/Size\s*(\d+)", RegexOptions.Singleline);
		private static readonly Regex nullRegex = new Regex(@"\sxref\s+0\s+\d+\s+(\d+)", RegexOptions.Singleline);
		private static readonly Regex refRegex = new Regex(@"\s*((\d+)\s+(\d+))?\s*(\d{10})\s+(\d{5})\s+(n|f)", RegexOptions.Singleline);
		private static readonly Regex objectRegex = new Regex(@"(\d+)\s+(\d+)\s+obj(.*?)endobj", RegexOptions.Singleline);
		private static readonly Regex fieldRegex = new Regex(@"^\s*<<(.*?/FT\s+/(Btn|Tx|Ch).*>>)", RegexOptions.Singleline);
		private static readonly Regex formRegex = new Regex(@"^\s*<<(.*?/Fields\s+\[.*>>)", RegexOptions.Singleline);
		private static readonly Regex acroFormRegex = new Regex(@"/AcroForm\s+(\d+)\s+(\d+)", RegexOptions.Singleline);
		private static readonly Regex trailRegex = new Regex(@"trailer\s+<<(.*>>)(\s+startxref\s+(\d+))?", RegexOptions.Singleline);
		private static readonly Regex objRegex = new Regex(@"(\d+)\s+(\d+)\s+obj", RegexOptions.Singleline);
		private static readonly Regex linearizedRegex = new Regex(@"/Linearized\s+1", RegexOptions.Singleline);

		private int previous = -1; // location of the previous cross-reference table
		private PdfDictionary previousTrailer;
		private int rootObjectNumber;
		private int nullOffset;
		private bool linearized = false;

		private Hashtable objects = new Hashtable();
		private SortedList offsets = new SortedList();

		/// <summary>
		/// Initializes a new instance of PdfReader with the specified file.
		/// </summary>
		/// <param name="name">The file containing the PDF data.</param>
		public static PdfReader GetPdfReader(string name)
		{
			PdfReader reader;

			using (FileStream stream = new FileStream(name, FileMode.Open))
			{
				reader = new PdfReader(stream);
			}

			return reader;
		}

		/// <summary>
		/// Initializes a new instance of PdfReader with the specified Stream.
		/// </summary>
		/// <param name="stream">The Stream containing the PDF data.</param>
		public PdfReader(Stream stream)
		{
			DOMConfigurator.ConfigureAndWatch(new FileInfo("PdfReader.exe.log4net"));

			byte[] buffer = new byte[stream.Length];
			stream.Read(buffer, 0, (int)stream.Length);

			char[] chars = new char[buffer.Length];
		
			for (int i = 0; i < buffer.Length; i++)
			{
				chars[i] = (char)buffer[i];
			}

			pdf = new String(chars);

			Parse();
		}

		/// <summary>
		/// Returns the PDF object referenced by the specified PDF reference.
		/// </summary>
		/// <param name="reference">The reference to the object.</param>
		/// <returns>The PDF object referenced.</returns>
		public PdfObject GetObjectForReference(PdfReference reference)
		{
			PdfObject PdfObject = null;

			// is the object active?
			if (objects.Contains(reference.ObjectNumber))
			{
				PdfCrossReferenceEntry entry = (PdfCrossReferenceEntry)objects[reference.ObjectNumber];

				if (entry.Active)
				{
					int start = entry.Offset;
					int end = GetEndOfObject(reference.ObjectNumber);

					PdfObject = ParseObject(start, end);
				}
			}

			return PdfObject;
		}
public void SetSelectedIndexes(params int[] indexes)
		{
			if (indexes == null || indexes.Length <= 0)
			{
				FieldDictionary.SetElement(IName, new PdfNull());
			}
			else if (indexes.Length == 1)
			{
				PdfNumber val = new PdfNumber((double)indexes[0]);

				FieldDictionary.SetElement(IName, val);
			}
			else
			{
				PdfNumber[] items = new PdfNumber[indexes.Length];
				int i = 0;
				foreach (int index in indexes)
				{
					items[i] = new PdfNumber((double)index);
					i++;
				}
				
				PdfArray val = new PdfArray(items);
				FieldDictionary.SetElement(IName, val);
			}
		}

		/// <summary>
		/// Gets the indexes of the items that should appear selected.
		/// </summary>
		/// <returns>An array of index that are selected.</returns>
		public int[] GetSelectedIndexes()
		{
			PdfObject iObject = FieldDictionary.GetElement(IName);

			if (iObject == null || iObject.GetType() == typeof (PdfNull))
			{
				return new int[] {};
			}
			else if (iObject.GetType() == typeof (PdfNumber))
			{
				return new int[] { (int)((PdfNumber)iObject).Number };
			}
			else
			{
				PdfArray array = (PdfArray)iObject;
				int[] indexes = new int[array.Elements.Count];

				for (int i = 0; i < array.Elements.Count; i++)
				{
					indexes[i] = (int)((PdfNumber)array[i]).Number;
				}

				return indexes;
			}
		}
	}

	/// <summary>
	/// Represents an AcroForm object in a PDF document. See the PDF reference 8.6.1 Interactive Form Dictionary.
	/// </summary>
	[CLSCompliant(true)]
	public class PdfForm: PdfField
	{
		private static PdfName NAName = new PdfName("/NeedAppearances");

		/// <summary>
		/// Initializes a new instance of PdfForm with the specified object number, generation number,
		/// and field dictionary.
		/// </summary>
		/// <param name="objNumber">The object number.</param>
		/// <param name="generationNumber">The generation number.</param>
		/// <param name="fieldDictionary">The field dictionary.</param>
		public PdfForm(int objNumber, int generationNumber, PdfDictionary fieldDictionary): base(objNumber,
			generationNumber, fieldDictionary)
		{
			// set NeedAppearances key so the viewer application regenerates the appearance streams
			// for the form fields.
			FieldDictionary.SetElement(NAName, new PdfBool(true));
		}
	}

	/// <summary>
	/// Represents an entry in the PDF Cross Reference table. 
	/// See the PDF Reference 3.4.3 Cross-Reference Table.
	/// </summary>
	[CLSCompliant(true)]
	public class PdfCrossReferenceEntry
	{
		private int objectNumber;
		private int generationNumber;
		private int offset;
		private bool active;

		/// <summary>
		/// The object number.
		/// </summary>
		public int ObjectNumber
		{
			get
			{
				return objectNumber;
			}

			set
			{
				objectNumber = value;
			}
		}

		/// <summary>
		/// The generation number.
		/// </summary>
		public int GenerationNumber
		{
			get
			{
				return generationNumber;
			}

			set
			{
				generationNumber = value;
			}
		}

		/// <summary>
		/// The byte offset of the object within the document.
		/// </summary>
		public int Offset
		{
			get
			{
				return offset;
			}

			set
			{
				offset = value;
			}
		}

		/// <summary>
		/// true if the object is not free, false otherwise.
		/// </summary>
		public bool Active
		{
			get
			{
				return active;
			}

			set
			{
				active = value;
			}
		}

		/// <summary>
		/// Initializes a new PdfCrossReferenceEntry object.
		/// </summary>
		/// <param name="objNumber">The object number</param>
		/// <param name="generationNumber">The generation number</param>
		/// <param name="offset">The byte offset within the PDF file</param>
		/// <param name="active">true if the object is not free, false otherwise</param>
		public PdfCrossReferenceEntry(int objNumber, int generationNumber,
			int offset, bool active)
		{
			ObjectNumber = objNumber;
			GenerationNumber = generationNumber;
			Offset = offset;
			Active = active;
		}
	}

[天罡gg]

请问gefangliang

(心灵彩虹<逍遥派掌门>)

上面代码引用的是什么DLL？

[天罡gg]

waiting...

[636f6c696e]

http://download.csdn.net/source/3209788

[心灵彩虹]

using System;
using System.IO;
using System.Text;
using System.Text.RegularExpressions;
using System.Collections;
using System.Diagnostics;
using System.Globalization;
using System.Security.Permissions;
using System.Runtime.InteropServices;
using System.Reflection;
using System.Runtime.CompilerServices;
using log4net;
using log4net.Config;

[assembly: CLSCompliant(true)]
[assembly: FileIOPermission(SecurityAction.RequestMinimum)]
[assembly: ComVisible(false)]
[assembly: AssemblyVersion("1.3.*")]
什么dll也没有，你要留下邮箱我给你发过去。

[xiaohuan_5377]

waiting...

[wangxiao2008]

mark

[天罡gg]

邮箱：[e-mail]
3Q

[天罡gg]

[Quote=引用 5 楼 colinfang2006 的回复:]
http://download.csdn.net/source/3209788
[/Quote]

哥们，我说的就是这个，只能解析文字

[天罡gg]

[Quote=引用 6 楼 gefangliang 的回复:]
using System;
using System.IO;
using System.Text;
using System.Text.RegularExpressions;
using System.Collections;
using System.Diagnostics;
using System.Globalization;
using System.Security.Per……
[/Quote]

那些对象PdfField，PdfObject，PdfReference，PdfDictionary，PdfForm等是什么？
.NET中有这两个引用？？
using log4net;
using log4net.Config;

[天罡gg]

确实有难度
http://topic.csdn.net/u/20081104/15/e4908012-d771-414c-acce-8d5a4a05f47f.html

gefangliang你不用给我发源码了，我自已找到了，那需要引用JAVA的两个DLL，不过没用，解析不正确

[scm1033189]

不能解析， MS没有支持过PDF，除非你用ADOB的插件。
而且PDF 就是类似是图片之类的文件 想把上面的文字用计算机完全读出来 ， 十分困难，尤其那些不是特别清晰PDF文件。 比如有一些书本有点旧了，或上面有标记， 那些字人虽然可以看懂 ，但是电脑识别起来是十分困难的。现在好像没有一个软件能完全将PDF全部正确的 转换为文本文件，总是存在错误机率.图像识别本来就是难题，尤其是中文，字太多，又复杂...

[上海猿猿]

正在研究，就是不知道如何调用Acrobat Core API

[lddwww]

正在看，关注。

[lddwww]

int x = Convert.ToInt32(textBox2.Text);//起始坐标 int y = Convert.ToInt32(textBox3.Text); int w = Convert.ToInt32(textBox4.Text);//宽度和高度 int h = Convert.ToInt32(textBox5.Text); PdfReader reader = new PdfReader("input.pdf"); //抽取文件 PdfReaderContentParser parser = new PdfReaderContentParser(reader); RectangleJ rect = new RectangleJ(x,y, w, h); RenderFilter filter = new RegionTextRenderFilter(rect); RenderFilter[] a = {filter}; FilteredTextRenderListener strategy = new FilteredTextRenderListener(new LocationTextExtractionStrategy(), a); textBox1.AppendText ( PdfTextExtractor.GetTextFromPage(reader, 1,strategy)+"\r\n");//输出

[lddwww]

这样可以解析指定区域里面的内容，在C#中添加对itextSharp.dll组件的引用。