public static WordInfo Compine(ArrayList morphinfos)
{
WordInfo res = new WordInfo();
res.senseCounts = new int[5];
res.texts = new ArrayList();
foreach (WordInfo morph in morphinfos)
{
for (int i = 1; i < 5; i++)
{
res.senseCounts[i] += morph.senseCounts[i];
}
if (!res.texts.Contains(morph.texts[0]))
{
res.texts.Add(morph.texts[0]);
}
}
return(res);
}
private static WordInfo GetMorphInfo(ArrayList morphinfos, string morph)
{
// Attempt to find the morph string in the list.
// NOTE: Since the list should never get very large, a selection search will work just fine
foreach (WordInfo morphinfo in morphinfos)
{
if (morphinfo.text == morph)
{
return(morphinfo);
}
}
// if not found, create a new one
WordInfo wordinfo = new WordInfo();
wordinfo.text = morph;
wordinfo.SenseCounts = new int[Enums.Length];
return((WordInfo)morphinfos[morphinfos.Add(wordinfo)]);
}
public static WordInfo Compine(WordInfo word, WordInfo morph)
{
if (word.Strength == 0)
{
return(morph);
}
if (morph.Strength == 0)
{
return(word);
}
WordInfo result = new WordInfo();
result.texts = new ArrayList();
result.texts.AddRange(morph.texts);
if (!result.texts.Contains(word.texts[0]))
{
result.texts.Add(word.texts[0]);
}
int MaxLen = Math.Max(word.senseCounts.Length, morph.senseCounts.Length);
result.senseCounts = new int[MaxLen];
for (int i = 0; i < MaxLen; i++)
{
result.senseCounts[i] = word.senseCounts[i] + morph.senseCounts[i];
}
Wnlib.PartsOfSpeech[] enums =
(Wnlib.PartsOfSpeech[])Enum.GetValues(typeof(Wnlib.PartsOfSpeech));
int MaxCount = 0;
result.partOfSpeech = Wnlib.PartsOfSpeech.Unknown;
for (int i = 0; i < enums.Length; i++)
{
if (result.senseCounts[i] > MaxCount)
{
MaxCount = result.senseCounts[i];
result.partOfSpeech = enums[i];
}
}
return(result);
}
private static WordInfo getMorphInfo(ArrayList morphinfos, string morph)
{
// Attempt to find the morph string in the list.
// NOTE: Since the list should never get very large, a selection search will work just fine
foreach (WordInfo morphinfo in morphinfos)
if (morphinfo.text == morph)
return morphinfo;
// if not found, create a new one
WordInfo wordinfo = new WordInfo();
wordinfo.text = morph;
wordinfo.senseCounts = new int[enums.Length];
return (WordInfo)morphinfos[morphinfos.Add(wordinfo)];
}
private static WordInfo lookupWordMorphs(string word)
{
// OVERVIEW: This functions only gets called when the word was not found with
// an exact match. So, enumerate all the parts of speech, then enumerate
// all of the word's morphs in each category. Perform a lookup on each
// morph and save the morph/strength/part-of-speech data sets. Finally,
// loop over all the data sets and then pick the strongest one.
ArrayList wordinfos = new ArrayList();
// for each part of speech...
for (int i = 0; i < enums.Length; i++)
{
// get a valid part of speech
Wnlib.PartsOfSpeech pos = enums[i];
if (pos == Wnlib.PartsOfSpeech.Unknown)
continue;
// generate morph list
Wnlib.MorphStr morphs = new Wnlib.MorphStr(word, Wnlib.PartOfSpeech.of(pos));
string morph = "";
while ((morph = morphs.next()) != null)
{
// get an index to a synset collection
Wnlib.Index index = Wnlib.Index.lookup(morph, Wnlib.PartOfSpeech.of(pos));
// none found?
if (index == null)
continue;
// save the wordinfo
WordInfo wordinfo = getMorphInfo(wordinfos, morph);
wordinfo.senseCounts[i] = index.sense_cnt;
}
}
// search the wordinfo list for the best match
WordInfo bestWordInfo = new WordInfo();
int maxStrength = 0;
foreach (WordInfo wordinfo in wordinfos)
{
// for each part of speech...
int maxSenseCount = 0;
int strength = 0;
for (int i = 0; i < enums.Length; i++)
{
// get a valid part of speech
Wnlib.PartsOfSpeech pos = enums[i];
if (pos == Wnlib.PartsOfSpeech.Unknown)
continue;
// determine part of speech and strength
strength += wordinfo.senseCounts[i];
if (wordinfo.senseCounts[i] > maxSenseCount)
{
maxSenseCount = wordinfo.senseCounts[i];
wordinfo.partOfSpeech = pos;
}
}
// best match?
if (strength > maxStrength)
{
maxStrength = strength;
bestWordInfo = wordinfo;
}
}
return bestWordInfo;
}
private static WordInfo lookupWord(string word)
{
// OVERVIEW: For each part of speech, look for the word.
// Compare relative strengths of the synsets in each category
// to determine the most probable part of speech.
//
// PROBLEM: Word definitions are often context-based. It would be better
// to find a way to search in-context in stead of just singling
// out an individual word.
//
// SOLUTION: Modify FindPartOfSpeech to include a second argument, string
// context. The pass the entire sentence as the context for part
// of speech determination.
//
// PROBLEM: That's difficult to do so I'm going to keep this simple for now.
int maxCount = 0;
WordInfo wordinfo = new WordInfo();
wordinfo.partOfSpeech = Wnlib.PartsOfSpeech.Unknown;
// for each part of speech...
Wnlib.PartsOfSpeech[] enums = (Wnlib.PartsOfSpeech[])Enum.GetValues(typeof(Wnlib.PartsOfSpeech));
wordinfo.senseCounts = new int[enums.Length];
for (int i = 0; i < enums.Length; i++)
{
// get a valid part of speech
Wnlib.PartsOfSpeech pos = enums[i];
if (pos == Wnlib.PartsOfSpeech.Unknown)
continue;
// get an index to a synset collection
Wnlib.Index index = Wnlib.Index.lookup(word, Wnlib.PartOfSpeech.of(pos));
// none found?
if (index == null)
continue;
// does this part of speech have a higher sense count?
wordinfo.senseCounts[i] = index.sense_cnt;
if (wordinfo.senseCounts[i] > maxCount)
{
maxCount = wordinfo.senseCounts[i];
wordinfo.partOfSpeech = pos;
}
}
return wordinfo;
}
private static WordInfo LookupWordMorphs(string word)
{
// OVERVIEW: This functions only gets called when the word was not found with
// an exact match. So, enumerate all the parts of speech, then enumerate
// all of the word's morphs in each category. Perform a lookup on each
// morph and save the morph/strength/part-of-speech data sets. Finally,
// loop over all the data sets and then pick the strongest one.
ArrayList wordinfos = new ArrayList();
// for each part of speech...
for (int i = 0; i < Enums.Length; i++)
{
// get a valid part of speech
Wnlib.PartsOfSpeech pos = Enums[i];
if (pos == Wnlib.PartsOfSpeech.Unknown)
{
continue;
}
// generate morph list
Wnlib.MorphStr morphs = new Wnlib.MorphStr(word, Wnlib.PartOfSpeech.of(pos));
string morph = "";
while ((morph = morphs.next()) != null)
{
// get an index to a synset collection
Wnlib.Index index = Wnlib.Index.lookup(morph, Wnlib.PartOfSpeech.of(pos));
// none found?
if (index == null)
{
continue;
}
// save the wordinfo
WordInfo wordinfo = GetMorphInfo(wordinfos, morph);
wordinfo.SenseCounts[i] = index.SenseCnt;
}
}
// search the wordinfo list for the best match
WordInfo bestWordInfo = new WordInfo();
int maxStrength = 0;
foreach (WordInfo wordinfo in wordinfos)
{
// for each part of speech...
int maxSenseCount = 0;
int strength = 0;
for (int i = 0; i < Enums.Length; i++)
{
// get a valid part of speech
Wnlib.PartsOfSpeech pos = Enums[i];
if (pos == Wnlib.PartsOfSpeech.Unknown)
{
continue;
}
// determine part of speech and strength
strength += wordinfo.SenseCounts[i];
if (wordinfo.SenseCounts[i] > maxSenseCount)
{
maxSenseCount = wordinfo.SenseCounts[i];
wordinfo.partOfSpeech = pos;
}
}
// best match?
if (strength > maxStrength)
{
maxStrength = strength;
bestWordInfo = wordinfo;
}
}
return(bestWordInfo);
}
public static string findadjinfo(string word, ArrayList texts)
{
foreach (string s in texts)
{
if (Adj.Contains(s.ToUpper()))
{
ArrayList Adjects = (ArrayList)Adj[s.ToUpper()];
if (Adjects.Count == 1)
{
return("PADJ");
}
for (int i = 0; i < Adjects.Count; i++)
{
foreach (string r in ((string)Adjects[i]).Split(','))
{
if (word == r)
{
switch (i)
{
case 0:
return("PADJ");
case 1:
return("CADJ");
case 2:
return("SADJ");
}
}
}
}
}
else if (word == s.ToUpper() + "ER")
{
return("CADJ");
}
else if (word == s.ToUpper() + "EST")
{
return("SADJ");
}
else if (s.ToUpper().EndsWith("E") && (word == s.ToUpper().Substring(0, s.Length - 1) + "ER") || word == s.ToUpper().Substring(0, s.Length - 1) + "EST")
{
if (word == s.ToUpper().Substring(0, s.Length - 1) + "ER")
{
return("CADJ");
}
if (word == s.ToUpper().Substring(0, s.Length - 1) + "EST")
{
return("SADJ");
}
}
else
{
WordInfo wi = FindWordInfo(word, tagged_only, false);
if (wi.senseCounts[3] > 0)
{
return("PADJ");
}
}
}
return(null);
}
private static ArrayList findverbinfo(string word, ArrayList texts)
{
ArrayList verbinfo = new ArrayList();
foreach (string s in texts)
{
if (Verbs.Contains(s.ToUpper()))
{
ArrayList verbs = (ArrayList)Verbs[s.ToUpper()];
for (int i = 0; i < verbs.Count; i++)
{
foreach (string r in ((string)verbs[i]).Split(','))
{
if (word == r)
{
switch (i)
{
case 0:
if (!verbinfo.Contains("VINF"))
{
verbinfo.Add("VINF");
}
break;
case 1:
if (!verbinfo.Contains("V"))
{
verbinfo.Add("V");
}
break;
case 2:
if (!verbinfo.Contains("VPSP"))
{
verbinfo.Add("VPSP");
}
break;
case 3:
if (!verbinfo.Contains("VING"))
{
verbinfo.Add("VING");
}
break;
case 4:
if (!verbinfo.Contains("V"))
{
verbinfo.Add("V");
}
break;
}
}
}
}
if (word == s.ToUpper() + "ING")
{
if (!verbinfo.Contains("VING"))
{
verbinfo.Add("VING");
}
}
if (s.ToUpper().EndsWith("E") && (word == s.ToUpper().Substring(0, s.Length - 1) + "ING" || word == s.ToUpper().Substring(0, s.Length - 1) + "ES"))
{
if (word == s.ToUpper().Substring(0, s.Length - 1) + "ING")
{
if (!verbinfo.Contains("VING"))
{
verbinfo.Add("VING");
}
else if (word == s.ToUpper().Substring(0, s.Length - 1) + "ES")
{
if (!verbinfo.Contains("V"))
{
verbinfo.Add("V");
}
}
}
}
if (word == s.ToUpper() + "S" || word == s.ToUpper() + "ES")
{
if (!verbinfo.Contains("V"))
{
verbinfo.Add("V");
}
}
if (s.ToUpper().EndsWith("Y"))
{
if (word == s.ToUpper().Substring(0, s.Length - 1) + "IES")
{
if (!verbinfo.Contains("V"))
{
verbinfo.Add("V");
}
}
}
}
if (word == s.ToUpper() + "ED")
{
if (!verbinfo.Contains("VPSP"))
{
verbinfo.Add("VPSP");
}
}
else if (word == s.ToUpper() + "ING")
{
if (!verbinfo.Contains("VING"))
{
verbinfo.Add("VING");
}
}
else if (word == s.ToUpper() + "S" || word == s.ToUpper() + "ES")
{
if (!verbinfo.Contains("V"))
{
verbinfo.Add("V");
}
}
else if (s.ToUpper().EndsWith("E") && (word == s.ToUpper().Substring(0, s.Length - 1) + "ING" || word == s.ToUpper().Substring(0, s.Length - 1) + "ES" || word == s.ToUpper().Substring(0, s.Length - 1) + "ED"))
{
if (word == s.ToUpper().Substring(0, s.Length - 1) + "ING")
{
if (!verbinfo.Contains("VING"))
{
verbinfo.Add("VING");
}
}
else if (word == s.ToUpper().Substring(0, s.Length - 1) + "ES")
{
if (!verbinfo.Contains("V"))
{
verbinfo.Add("V");
}
}
else if (word == s.ToUpper().Substring(0, s.Length - 1) + "ED")
{
if (!verbinfo.Contains("VPSP"))
{
verbinfo.Add("VPSP");
}
}
}
else if (s.ToUpper().EndsWith("Y") && word == s.ToUpper().Substring(0, s.Length - 1) + "IES")
{
if (word == s.ToUpper().Substring(0, s.Length - 1) + "IES")
{
if (!verbinfo.Contains("V"))
{
verbinfo.Add("V");
}
}
}
else
{
WordInfo wi = FindWordInfo(word, tagged_only, false);
if (wi.senseCounts[2] > 0 && !verbinfo.Contains("VINF"))
{
verbinfo.Add("VINF");
}
}
}
if (!verbinfo.Contains("V") &&
(verbinfo.Contains("VINF") || verbinfo.Contains("VPSP")))
{
verbinfo.Add("V");
}
return(verbinfo);
}
public static bool IsPlural(string Word)
{
foreach (string s in Noun.Keys)
{
if (s.ToUpper() == Word)
{
return(true);
}
}
/*
* "s", "ses", "xes", "zes", "ches", "shes", "men", "ies",*/
/*"", "s", "x", "z", "ch", "sh", "man", "y",*/
if (Word.EndsWith("CHES"))
{
Word = Word.Remove(Word.Length - 4, 4);
Word += "CH";
}
else if (Word.EndsWith("SHES"))
{
Word = Word.Remove(Word.Length - 4, 4);
Word += "SH";
}
else if (Word.EndsWith("SES"))
{
Word = Word.Remove(Word.Length - 3, 3);
Word += "S";
}
else if (Word.EndsWith("XES"))
{
Word = Word.Remove(Word.Length - 3, 3);
Word += "X";
}
else if (Word.EndsWith("ZES"))
{
Word = Word.Remove(Word.Length - 3, 3);
Word += "Z";
}
else if (Word.EndsWith("MEN"))
{
Word = Word.Remove(Word.Length - 3, 3);
Word += "MAN";
}
else if (Word.EndsWith("IES"))
{
Word = Word.Remove(Word.Length - 3, 3);
Word += "Y";
}
else if (Word.EndsWith("S"))
{
Word = Word.Remove(Word.Length - 1, 1);
}
else
{
return(false);
}
WordInfo wi = FindWordInfo(Word, tagged_only, false);
if (wi.senseCounts[1] > 0)
{
return(true);
}
else
{
return(false);
}
}
private static WordInfo lookupWordMorphs(string word, bool tagged_only)
{
// OVERVIEW: This functions only gets called when the word was not found with
// an exact match. So, enumerate all the parts of speech, then enumerate
// all of the word's morphs in each category. Perform a lookup on each
// morph and save the morph/strength/part-of-speech data sets. Finally,
// loop over all the data sets and then pick the strongest one.
ArrayList wordinfos = new ArrayList();
// for each part of speech...
for (int i = 0; i < enums.Length; i++)
{
// get a valid part of speech
Wnlib.PartsOfSpeech pos = enums[i];
if (pos == Wnlib.PartsOfSpeech.Unknown)
{
continue;
}
// generate morph list
Wnlib.MorphStr morphs = new Wnlib.MorphStr(word, Wnlib.PartOfSpeech.of(pos));
string morph = "";
while ((morph = morphs.next()) != null)
{
// get an index to a synset collection
Wnlib.Index index = Wnlib.Index.lookup(morph, Wnlib.PartOfSpeech.of(pos));
// none found?
if (index == null)
{
continue;
}
// none tagged
if (tagged_only && index.tagsense_cnt == 0)
{
continue;
}
// save the wordinfo
WordInfo wordinfo = getMorphInfo(wordinfos, morph);
if (tagged_only)
{
wordinfo.senseCounts[i] = index.tagsense_cnt;
}
else
{
wordinfo.senseCounts[i] = index.sense_cnt;
}
}
}
return(WordInfo.Compine(wordinfos));
/*
* // search the wordinfo list for the best match
* WordInfo bestWordInfo = new WordInfo();
* int maxStrength = 0;
* foreach( WordInfo wordinfo in wordinfos )
* {
* // for each part of speech...
* int maxSenseCount = 0;
* int strength = 0;
* for( int i=0; i<enums.Length; i++ )
* {
* // get a valid part of speech
* Wnlib.PartsOfSpeech pos = enums[i];
* if( pos == Wnlib.PartsOfSpeech.Unknown )
* continue;
*
* // determine part of speech and strength
* strength += wordinfo.senseCounts[i];
* if( wordinfo.senseCounts[i] > maxSenseCount )
* {
* maxSenseCount = wordinfo.senseCounts[i];
* wordinfo.partOfSpeech = pos;
* }
* }
*
* // best match?
* if( strength > maxStrength )
* {
* maxStrength = strength;
* bestWordInfo = wordinfo;
* }
* }
*
* return bestWordInfo;
*/
}
private static WordInfo lookupWord(string word, bool tagged_only)
{
// OVERVIEW: For each part of speech, look for the word.
// Compare relative strengths of the synsets in each category
// to determine the most probable part of speech.
//
// PROBLEM: Word definitions are often context-based. It would be better
// to find a way to search in-context in stead of just singling
// out an individual word.
//
// SOLUTION: Modify FindPartOfSpeech to include a second argument, string
// context. The pass the entire sentence as the context for part
// of speech determination.
//
// PROBLEM: That's difficult to do so I'm going to keep this simple for now.
int maxCount = 0;
WordInfo wordinfo = new WordInfo();
wordinfo.partOfSpeech = Wnlib.PartsOfSpeech.Unknown;
wordinfo.texts = new ArrayList();
wordinfo.texts.Add(word);
// for each part of speech...
Wnlib.PartsOfSpeech[] enums = (Wnlib.PartsOfSpeech[])Enum.GetValues(typeof(Wnlib.PartsOfSpeech));
wordinfo.senseCounts = new int[enums.Length];
for (int i = 0; i < enums.Length; i++)
{
// get a valid part of speech
Wnlib.PartsOfSpeech pos = enums[i];
if (pos == Wnlib.PartsOfSpeech.Unknown)
{
continue;
}
// get an index to a synset collection
Wnlib.Index index = Wnlib.Index.lookup(word, Wnlib.PartOfSpeech.of(pos));
// none found?
if (index == null)
{
continue;
}
// none tagged
if (tagged_only && index.tagsense_cnt == 0)
{
continue;
}
// does this part of speech have a higher sense count?
if (tagged_only)
{
wordinfo.senseCounts[i] = index.tagsense_cnt;
}
else
{
wordinfo.senseCounts[i] = index.sense_cnt;
}
if (wordinfo.senseCounts[i] > maxCount)
{
maxCount = wordinfo.senseCounts[i];
wordinfo.partOfSpeech = pos;
}
}
return(wordinfo);
}
