在C#中使用IFilter并从数据库而不是文件系统中检索文件

joh*_*hna 5 c# ifilter

对于C#Web应用程序,我希望索引存储在数据库中的PDF,DOC等文件的文本.

我一直在试验Code Project上的IFilter示例,该示例适用于文件系统中的文件,但我的文件存储在MS-SQL数据库中.

任何人都可以帮我找到一个样本来从存储在数据库中的文件中提取文本,或者知道如何修改Code Project代码以使用数据库而不是文件系统?

Mar*_*icz 13

最后几个小时后,我想出了如何使这项工作!我需要在存储在数据库中的PDF内容上运行IFilter,我想避免将数据保存到临时文件中.

首先,我尝试使用BindIFilterFromStream API为存储在Stream中的内容创建IFilter,但似乎它无法正常工作(至少不适用于此场景).所以不要这样做.

相反,您需要为文件扩展名创建一个IFilter(或以其他方式访问它).然后,您可以访问IPersistStream COM接口并使用它将PDF内容加载到IFilter中.其余的工作与文件相同.但请注意,每个IFilter都可能无法实现IPersistStream API.它适用于Adobe PDF IFilter.

代码应如下所示(我删除了一些返回代码检查以使代码更具可读性,但是,您应该检查所有可能的返回代码).

private string ParseIFilter(Stream s)
{
   // Get an IFilter for a file or file extension
   IFilter filter = null;
   FilterReturnCodes result = NativeMethods.LoadIFilter(".pdf", null, ref filter);
   if (result != FilterReturnCodes.S_OK)
   {
      Marshal.ThrowExceptionForHR((int)result);
   }

   // Copy the content to global memory
   byte[] buffer = new byte[s.Length];
   s.Read(buffer, 0, buffer.Length);
   IntPtr nativePtr = Marshal.AllocHGlobal(buffer.Length);
   Marshal.Copy(buffer, 0, nativePtr, buffer.Length);

   // Create a COM stream
   System.Runtime.InteropServices.ComTypes.IStream comStream;
   NativeMethods.CreateStreamOnHGlobal(nativePtr, true, out comStream);

   // Load the contents to the iFilter using IPersistStream interface
   var persistStream = (IPersistStream)filter;
   persistStream.Load(comStream);

   // Initialize iFilter
   FilterFlags filterFlags;
   FilterReturnCodes result = filter.Init(
      FilterInit.IFILTER_INIT_INDEXING_ONLY, 0, IntPtr.Zero, out filterFlags);

   return ExtractTextFromIFilter(filter);
}
Run Code Online (Sandbox Code Playgroud)

过滤器中的文本提取在我的代码中看起来像这样.网上有很多这样的例子,它可以根据你的需要以多种方式实现.

private string ExtractTextFromIFilter(IFilter filter)
{
   var sb = new StringBuilder();

   while (true)
   {
      StatChunk chunk;
      result = filter.GetChunk(out chunk);

      if (result == FilterReturnCodes.S_OK)
      {
         if (chunk.flags == ChunkState.CHUNK_TEXT)
         {
            sb.Append(ExtractTextFromChunk(filter, chunk));
         }

         continue;
      }

      if (result == FilterReturnCodes.FILTER_E_END_OF_CHUNKS)
      {
         return sb.ToString();
      }

      Marshal.ThrowExceptionForHR((int)result);
   }
}

private virtual string ExtractTextFromChunk(IFilter filter, StatChunk chunk)
{
   var sb = new StringBuilder();

   var result = FilterReturnCodes.S_OK;
   while (result == FilterReturnCodes.S_OK)
   {
      int sizeBuffer = 16384;
      var buffer = new StringBuilder(sizeBuffer);
      result = filter.GetText(ref sizeBuffer, buffer);

      if ((result == FilterReturnCodes.S_OK) || (result == FilterReturnCodes.FILTER_S_LAST_TEXT))
      {
         if((sizeBuffer > 0) && (buffer.Length > 0))
         {
            sb.Append(buffer.ToString(0, sizeBuffer));
         }
      }

      if (result == FilterReturnCodes.FILTER_E_NO_TEXT)
      {
         return string.Empty;
      }

      if ((result == FilterReturnCodes.FILTER_S_LAST_TEXT) || (result == FilterReturnCodes.FILTER_E_NO_MORE_TEXT))
      {
         return sb.ToString();
      }
   }

   return sb.ToString();
}
Run Code Online (Sandbox Code Playgroud)

以下是本机方法的定义及其使用的结构.

internal static class NativeMethods
{
    [DllImport("query.dll", SetLastError = true, CharSet = CharSet.Unicode)]
    public static extern FilterReturnCodes LoadIFilter(
        string pwcsPath,
        [MarshalAs(UnmanagedType.IUnknown)] object pUnkOuter,
        ref IFilter ppIUnk);

    [DllImport("ole32.dll")]
    public static extern int CreateStreamOnHGlobal(IntPtr hGlobal, bool fDeleteOnRelease, out IStream ppstm);
}

[ComImport, Guid("89BCB740-6119-101A-BCB7-00DD010655AF")]
[InterfaceType(ComInterfaceType.InterfaceIsIUnknown)]
public interface IFilter
{
    [PreserveSig]
    FilterReturnCodes Init(FilterInit grfFlags, int cAttributes, IntPtr aAttributes, out FilterFlags pdwFlags);

    [PreserveSig]
    FilterReturnCodes GetChunk(out StatChunk pStat);

    [PreserveSig]
    FilterReturnCodes GetText(
        ref int pcwcBuffer,
        [Out, MarshalAs(UnmanagedType.LPWStr)] StringBuilder awcBuffer);

    [PreserveSig]
    FilterReturnCodes GetValue(ref IntPtr propVal);

    [PreserveSig]
    FilterReturnCodes BindRegion(ref FilterRegion origPos, ref Guid riid, ref object ppunk);
}

[InterfaceType(ComInterfaceType.InterfaceIsIUnknown), Guid("0000010c-0000-0000-C000-000000000046")]
public interface IPersist
{
    void GetClassID(out Guid pClassID);
}

[InterfaceType(ComInterfaceType.InterfaceIsIUnknown), Guid("00000109-0000-0000-C000-000000000046")]
public interface IPersistStream : IPersist
{
    new void GetClassID(out Guid pClassID);

    [PreserveSig]
    int IsDirty();

    void Load([In] IStream pStm);

    void Save(
        [In] IStream pStm,
        [In, MarshalAs(UnmanagedType.Bool)] bool fClearDirty);

    void GetSizeMax(out long pcbSize);
}

public struct StatChunk
{
    public int idChunk;
    [MarshalAs(UnmanagedType.U4)]
    public ChunkBreaktype breakType;
    [MarshalAs(UnmanagedType.U4)]
    public ChunkState flags;
    public int locale;
    public FullPropSpec attribute;
    public int idChunkSource;
    public int cwcStartSource;
    public int cwcLenSource;
}

public enum ChunkBreaktype
{
    CHUNK_NO_BREAK = 0,
    CHUNK_EOW = 1,
    CHUNK_EOS = 2,
    CHUNK_EOP = 3,
    CHUNK_EOC = 4
}

public enum ChunkState
{
    CHUNK_TEXT = 0x1,
    CHUNK_VALUE = 0x2,
    CHUNK_FILTER_OWNED_VALUE = 0x4
}

[Flags]
public enum FilterFlags
{
    IFILTER_FLAGS_OLE_PROPERTIES = 1
}

[Flags]
public enum FilterInit
{
    IFILTER_INIT_CANON_PARAGRAPHS = 1,
    IFILTER_INIT_HARD_LINE_BREAKS = 2,
    IFILTER_INIT_CANON_HYPHENS = 4,
    IFILTER_INIT_CANON_SPACES = 8,
    IFILTER_INIT_APPLY_INDEX_ATTRIBUTES = 16,
    IFILTER_INIT_APPLY_CRAWL_ATTRIBUTES = 256,
    IFILTER_INIT_APPLY_OTHER_ATTRIBUTES = 32,
    IFILTER_INIT_INDEXING_ONLY = 64,
    IFILTER_INIT_SEARCH_LINKS = 128,
    IFILTER_INIT_FILTER_OWNED_VALUE_OK = 512
}

public struct FilterRegion
{
    public int idChunk;
    public int cwcStart;
    public int cwcExtent;
}

public enum FilterReturnCodes : uint
{
    S_OK = 0,
    E_ACCESSDENIED = 0x80070005,
    E_HANDLE = 0x80070006,
    E_INVALIDARG = 0x80070057,
    E_OUTOFMEMORY = 0x8007000E,
    E_NOTIMPL = 0x80004001,
    E_FAIL = 0x80000008,
    FILTER_E_PASSWORD = 0x8004170B,
    FILTER_E_UNKNOWNFORMAT = 0x8004170C,
    FILTER_E_NO_TEXT = 0x80041705,
    FILTER_E_NO_VALUES = 0x80041706,
    FILTER_E_END_OF_CHUNKS = 0x80041700,
    FILTER_E_NO_MORE_TEXT = 0x80041701,
    FILTER_E_NO_MORE_VALUES = 0x80041702,
    FILTER_E_ACCESS = 0x80041703,
    FILTER_W_MONIKER_CLIPPED = 0x00041704,
    FILTER_E_EMBEDDING_UNAVAILABLE = 0x80041707,
    FILTER_E_LINK_UNAVAILABLE = 0x80041708,
    FILTER_S_LAST_TEXT = 0x00041709,
    FILTER_S_LAST_VALUES = 0x0004170A
}

public struct FullPropSpec
{
    public Guid guidPropSet;
    public PropSpec psProperty;
}

[StructLayout(LayoutKind.Explicit)]
public struct PropSpec
{
    [FieldOffset(0)]
    public int ulKind;     

    [FieldOffset(4)]
    public int propid;

    [FieldOffset(4)]
    public IntPtr lpwstr;
}
Run Code Online (Sandbox Code Playgroud)

  • 哇,这就是我所说的详细答案! (2认同)