Click here to Skip to main content
Click here to Skip to main content
Add your own
alternative version

Statistical parsing of English sentences

, 13 Dec 2006
Shows how to generate parse trees for English language sentences, using a C# port of OpenNLP, a statistical natural language parsing library.
englishparsing_bin.zip
Lithium.dll
ModelConverter.exe
OpenNLP Tools.chm
OpenNLP.dll
ParseTree.exe
SharpEntropy.dll
ToolsExample.exe
englishparsing_net2_0_bin.zip
ToolsExample.exe
Lithium.dll
ModelConverter.exe
OpenNLP Tools.chm
OpenNLP.dll
ParseTree.exe
SharpEntropy.dll
englishparsing_net2_0_src.zip
Lithium
Collections
Delegates
Enums
Interfaces
IO
Lithium.csproj.vspscc
LithiumControl.bmp
Shapes
UI
Visitors
ModelConverter
App.ico
ModelConverter.csproj.vspscc
ParseTree
App.ico
ParseTree.csproj.vspscc
ToolsExample
App.ico
ToolsExample.csproj.vspscc
OpenNLP
OpenNLP.csproj.vspscc
SharpEntropy.dll
Tools
Chunker
NameFind
Parser
PosTagger
SentenceDetect
Tokenize
Util
englishparsing_src.zip
Lithium.csproj.user
LithiumControl.bmp
App.ico
ModelConverter.csproj.user
OpenNLP.csproj.user
SharpEntropy.dll
vssver.scc
vssver.scc
vssver.scc
vssver.scc
vssver.scc
vssver.scc
vssver.scc
App.ico
ParseTree.csproj.user
App.ico
ToolsExample.csproj.user
//Copyright (C) 2005 Richard J. Northedge
//
// This library is free software; you can redistribute it and/or
// modify it under the terms of the GNU Lesser General Public
// License as published by the Free Software Foundation; either
// version 2.1 of the License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
// GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this program; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.

//This file is based on the DefaultPOSContextGenerator.java source file found in the
//original java implementation of OpenNLP.  That source file contains the following header:

// Copyright (C) 2002 Jason Baldridge and Gann Bierner
// 
// This library is free software; you can redistribute it and/or
// modify it under the terms of the GNU Lesser General Public
// License as published by the Free Software Foundation; either
// version 2.1 of the License, or (at your option) any later version.
// 
// This library is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
// GNU Lesser General Public License for more details.
// 
// You should have received a copy of the GNU Lesser General Public
// License along with this program; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.

using System;
using System.Collections;
using System.Text.RegularExpressions;

namespace OpenNLP.Tools.PosTagger
{
	/// <summary> 
	/// A context generator for the POS Tagger.
	/// </summary>	
	public class DefaultPosContextGenerator : IPosContextGenerator
	{
		protected internal const string SentenceEnd = "*SE*";
		protected internal const string SentenceBeginning = "*SB*";

		private const int mPrefixLength = 4;
		private const int mSuffixLength = 4;
		
		private static Regex mHasCapitalRegex = new Regex("[A-Z]");
		private static Regex mHasNumericRegex = new Regex("[0-9]");
		
		private Util.Cache mContextsCache;
		private object mWordsKey;

		public DefaultPosContextGenerator() : this(0)
		{
		}
		
		public DefaultPosContextGenerator(int cacheSize) 
		{
			if (cacheSize > 0) 
			{
				mContextsCache = new Util.Cache(cacheSize);
			}
		}

		public virtual string[] GetContext(object input)
		{
			object[] data = (object[]) input;
			return GetContext(((int) data[0]), (object[]) data[1], (string[]) data[2], null);
		}
		
		protected internal static string[] GetPrefixes(string lex)
		{
			string[] prefixes = new string[mPrefixLength];
			for (int currentPrefix = 0; currentPrefix < mPrefixLength; currentPrefix++)
			{
				prefixes[currentPrefix] = lex.Substring(0, (System.Math.Min(currentPrefix + 1, lex.Length)) - (0));
			}
			return prefixes;
		}
		
		protected internal static string[] GetSuffixes(string lex)
		{
			string[] suffixes = new string[mSuffixLength];
			for (int currentSuffix = 0; currentSuffix < mSuffixLength; currentSuffix++)
			{
				suffixes[currentSuffix] = lex.Substring(System.Math.Max(lex.Length - currentSuffix - 1, 0));
			}
			return suffixes;
		}
		
		public virtual string[] GetContext(int index, object[] sequence, string[] priorDecisions, object[] additionalContext) 
		{
			return GetContext(index, sequence, priorDecisions);
		}

		/// <summary>
		/// Returns the context for making a pos tag decision at the specified token index given the specified tokens and previous tags.
		/// </summary>
		/// <param name="index">
		/// The index of the token for which the context is provided.
		/// </param>
		/// <param name="tokens">
		/// The tokens in the sentence.
		/// </param>
		/// <param name="tags">
		/// The tags assigned to the previous words in the sentence.
		/// </param>
		/// <returns>
		/// The context for making a pos tag decision at the specified token index given the specified tokens and previous tags.
		/// </returns>
		public virtual string[] GetContext(int index, object[] tokens, string[] tags) 
		{
			string next, nextNext, lex, previous, previousPrevious;
			string tagPrevious, tagPreviousPrevious;
			tagPrevious = tagPreviousPrevious = null;
			next = nextNext = lex = previous = previousPrevious = null;
			
			lex = tokens[index].ToString();
			if (tokens.Length > index + 1) 
			{
				next = tokens[index + 1].ToString();
				if (tokens.Length > index + 2)
				{
					nextNext = tokens[index + 2].ToString();
				}
				else
				{
					nextNext = SentenceEnd; 
				}
			}
			else
			{
				next = SentenceEnd; 
			}
			
			if (index - 1 >= 0) 
			{
				previous = tokens[index - 1].ToString();
				tagPrevious = tags[index - 1].ToString();

				if (index - 2 >= 0) 
				{
					previousPrevious = tokens[index - 2].ToString();
					tagPreviousPrevious = tags[index - 2].ToString();
				}
				else
				{
					previousPrevious = SentenceBeginning; 
				}
			}
			else
			{
				previous = SentenceBeginning; 
			}
			
			String cacheKey = index.ToString(System.Globalization.CultureInfo.InvariantCulture) + tagPrevious + tagPreviousPrevious;
			if (!(mContextsCache == null)) 
			{
				if (mWordsKey == tokens)
				{
					string[] cachedContexts = (string[]) mContextsCache[cacheKey];    
					if (cachedContexts != null) 
					{
						return cachedContexts;
					}
				}
				else 
				{
					mContextsCache.Clear();
					mWordsKey = tokens;
				}

			}

			ArrayList eventList = new ArrayList();
			
			// add the word itself
			eventList.Add("w=" + lex);
			
			// do some basic suffix analysis
			string[] suffixes = GetSuffixes(lex);
			for (int currentSuffix = 0; currentSuffix < suffixes.Length; currentSuffix++)
			{
				eventList.Add("suf=" + suffixes[currentSuffix]);
			}
			
			string[] prefixes = GetPrefixes(lex);
			for (int currentPrefix = 0; currentPrefix < prefixes.Length; currentPrefix++)
			{
				eventList.Add("pre=" + prefixes[currentPrefix]);
			}
			// see if the word has any special characters
			if (lex.IndexOf((char) '-') != - 1)
			{
				eventList.Add("h");
			}
			
			if (mHasCapitalRegex.IsMatch(lex)) 
			{
				eventList.Add("c");
			}
			
			if (mHasNumericRegex.IsMatch(lex))
			{
				eventList.Add("d");
			}
			
			// add the words and positions of the surrounding context
			if ((object) previous != null)
			{
				eventList.Add("p=" + previous);
				if ((object) tagPrevious != null)
				{
					eventList.Add("t=" + tagPrevious);
				}
				if ((object) previousPrevious != null)
				{
					eventList.Add("pp=" + previousPrevious);
					if ((object) tagPreviousPrevious != null)
					{
						eventList.Add("tt=" + tagPreviousPrevious);
					}
				}
			}
			
			if ((object) next != null)
			{
				eventList.Add("n=" + next);
				if ((object) nextNext != null)
				{
					eventList.Add("nn=" + nextNext);
				}
			}

			string[] contexts = (string[]) eventList.ToArray(typeof(string));
			if (mContextsCache != null) 
			{
				mContextsCache[cacheKey] = contexts;
			}
			return (contexts);
		}
		
	}
}

By viewing downloads associated with this article you agree to the Terms of Service and the article's licence.

If a file you wish to view isn't highlighted, and is a text file (not binary), please let us know and we'll add colourisation support for it.

License

This article has no explicit license attached to it but may contain usage terms in the article text or the download files themselves. If in doubt please contact the author via the discussion board below.

A list of licenses authors might use can be found here

Share

About the Author

Richard Northedge
Web Developer
United Kingdom United Kingdom
Richard Northedge is a senior developer with a UK Microsoft Gold Partner company. He has a postgraduate degree in English Literature, has been programming professionally since 1998 and has been an MCSD since 2000.

| Advertise | Privacy | Mobile
Web04 | 2.8.141022.2 | Last Updated 13 Dec 2006
Article Copyright 2005 by Richard Northedge
Everything else Copyright © CodeProject, 1999-2014
Terms of Service
Layout: fixed | fluid