} // for serialization
public SynSet(int offset, PartOfSpeech p, string wd, Search sch, int sens)
{
pos = p;
hereiam = offset;
search = sch;
sense = sens;
StreamReader f = WNDB.data(p);
f.DiscardBufferedData();
//f.BaseStream.Seek(off,SeekOrigin.Begin);
f.BaseStream.Position = offset;
string rec = f.ReadLine();
//string str;
if (!rec.StartsWith(offset.ToString("D8")))
{
Console.WriteLine("Error reading " + p.name + " file! " + offset + ": " + rec);
WNDB.reopen(p);
f = WNDB.data(p);
f.DiscardBufferedData();
//f.BaseStream.Seek(off, SeekOrigin.Begin);
f.BaseStream.Position = offset;
rec = f.ReadLine();
//if (!rec.StartsWith(off.ToString("D8")))
// str = "";
//else
// Console.WriteLine("Recovered...");
}
Parse(rec, pos, wd);
}
public Indexes(string str, PartOfSpeech pos)
{
string[] strings = new string[stringscount];
str = str.ToLower();
strings[0] = str;
strings[1] = str.Replace('_', '-');
strings[2] = str.Replace('-', '_');
strings[3] = str.Replace("-", "").Replace("_", "");
strings[4] = str.Replace(".", "");
// TDMS - 14 Aug 2005 - added two more possibilities
// to allow for spaces to be transformed
// an example is a search for "11 plus", without this
// transformation no result would be found
strings[5] = str.Replace(' ', '-');
strings[6] = str.Replace(' ', '_');
offsets[0] = Index.lookup(str, pos);
// TDMS - 14 Aug 2005 - changed 5 to 7 to allow for two
// new possibilities
for (int i = 1; i < stringscount; i++)
{
if (str != strings[i])
{
offsets[i] = Index.lookup(strings[i], pos);
}
}
fpos = pos;
}
/// <summary>
/// Trace meronyms.
/// </summary>
/// <param name="pbase"></param>
/// <param name="fpos"></param>
/// <param name="depth"></param>
void traceInherit(PointerType pbase, PartOfSpeech fpos, int depth)
{
for (int i = 0; i < ptrs.Length; i++)
{
Pointer pt = ptrs[i];
if (pt.ptp.Ident == HYPERPTR && (pt.sce == 0 || pt.sce == whichword))
{
spaces("TRACEI", depth);
SynSet cursyn = new SynSet(pt.off, pt.pos, this);
search.wordsFrom(cursyn);
cursyn.str("=> ", "\n", 1, 0, 0, 1);
// TDMS 6 Oct 2005 - build hierarchical results
// TODO: verify this
if (this.senses == null)
{
this.senses = new SynSetList();
}
cursyn.thisptr = pt; // TDMS 17 Nov 2005 - add this pointer type
// TODO: This is adding senses incorrectly
this.senses.Add(cursyn);
cursyn.tracePtrs(pbase, PartOfSpeech.of("noun"), depth);
cursyn.tracePtrs(pbase + 1, PartOfSpeech.of("noun"), depth);
cursyn.tracePtrs(pbase + 2, PartOfSpeech.of("noun"), depth);
if (depth > 0)
{
depth = depthcheck(depth);
cursyn.traceInherit(pbase, cursyn.pos, depth + 1);
}
}
}
search.trunc();
}
public static void reopen(PartOfSpeech p)
{
index(p).Close();
data(p).Close();
indexfps[p.key] = GetStreamReader(IndexFile(p));
datafps[p.key] = GetStreamReader(DataFile(p));
}
void catword(int wdnum, int adjmarker, int antflag)
{
search.buf += deadjify(words[wdnum].word);
/* Print additional lexicographer information and WordNet sense
* number as indicated by flags */
if (words[wdnum].uniq != 0)
{
search.buf += "" + words[wdnum].uniq;
}
int s = getsearchsense(wdnum + 1);
words[wdnum].wnsns = s;
if (WNOpt.opt("-s").flag)
{
search.buf += "#" + s;
}
/* For adjectives, append adjective marker if present, and
* print antonym if flag is passed */
if (pos.name == "adj")
{
if (adjmarker > 0)
{
search.buf += "" + adj_marker.mark;
}
if (antflag > 0)
{
strAnt(PartOfSpeech.of("adj"), wdnum + 1, "(vs. ", ")");
}
}
}
internal Search(string w, PartOfSpeech p, SearchType s, int sn)
{
word = w;
pos = p;
sch = s;
whichsense = sn;
}
internal Search(string w, PartOfSpeech p, SearchType s, int sn)
{
word = w;
pos = p;
sch = s;
whichsense = sn;
}
/* From search.c:
* Find word in index file and return parsed entry in data structure.
Input word must be exact match of string in database. */
// From the WordNet Manual (http://wordnet.princeton.edu/man/wnsearch.3WN.html)
// index_lookup() finds searchstr in the index file for pos and returns a pointer
// to the parsed entry in an Index data structure. searchstr must exactly match the
// form of the word (lower case only, hyphens and underscores in the same places) in
// the index file. NULL is returned if a match is not found.
public static Index lookup(string word, PartOfSpeech pos)
{
int j;
if (word == "")
return null;
// TDMS 14 Aug 2005 - changed to allow for numbers as well
// because the database contains searches that can start with
// numerals
//if (!char.IsLetter(word[0]))
if (!char.IsLetter(word[0]) && !char.IsNumber(word[0]))
return null;
string line = WNDB.binSearch(word, pos);
if (line == null)
return null;
Index idx = new Index();
StrTok st = new StrTok(line);
idx.Wd = st.next(); /* the word */
idx.PartOfSpeech = PartOfSpeech.of(st.next()); /* the part of speech */
idx.SenseCnt = int.Parse(st.next()); /* collins count */
int ptruse_cnt = int.Parse(st.next()); /* number of pointers types */
idx.Ptruse = new PointerType[ptruse_cnt];
for (j = 0; j < ptruse_cnt; j++)
idx.Ptruse[j] = PointerType.of(st.next());
int off_cnt = int.Parse(st.next());
idx.SynsetOffsets = new int[off_cnt];
idx.TaggedSensesCount = int.Parse(st.next());
for (j = 0; j < off_cnt; j++)
idx.SynsetOffsets[j] = int.Parse(st.next());
return idx;
}
internal static ArrayList wngrep(string wordPassed, PartOfSpeech pos)
{
ArrayList r = new ArrayList();
StreamReader inputFile = index(pos);
inputFile.BaseStream.Seek(0L, SeekOrigin.Begin);
inputFile.DiscardBufferedData();
string word = wordPassed.Replace(" ", "_");
string line;
while ((line = inputFile.ReadLine()) != null)
{
int lineLen = line.IndexOf(' ');
line = line.Substring(0, lineLen);
try
{
if (line.IndexOf(word) >= 0)
{
r.Add(line.Replace("_", " "));
}
}
catch
{
}
}
return(r);
}
public Exceptions(string word, PartOfSpeech pos)
{
line = WNDB.binSearch(word, excfps[pos.ident]);
if (line != null)
{
endlp = line.IndexOf(' ');
}
}
public Exceptions(string word, PartOfSpeech pos)
{
line = WNDB.binSearch(word, (StreamReader)excfps[pos.key]);
if (line != null)
{
endlp = line.IndexOf(' ');
}
}
static bool AddSearchFor(string s, PartOfSpeech pos, ArrayList list)
{
var se = new Search(s, false, pos, new SearchType(false, "OVERVIEW"), 0);
if (se.lexemes.Count > 0)
list.Add(se);
return se.lexemes.Count > 0;
}
public Search(string w, bool doMorphs, PartOfSpeech p, SearchType s, int sn)
: this(w, p, s, sn)
{
if (p != null)
{
do_search(doMorphs);
}
}
public Search(string w, bool doMorphs, PartOfSpeech p, SearchType s, int sn, CustomGrep cg)
: this(w, p, s, sn)
{
customgrep = cg;
if (p != null)
{
do_search(doMorphs);
}
}
public static void reopen(PartOfSpeech p)
{
WNDBpart w = WNDBpart.of(p);
w.index.Close();
w.data.Close();
w.index = new StreamReader(IndexFile(p));
w.data = new StreamReader(DataFile(p));
}
static Exceptions()
{
excfps = new StreamReader[PartOfSpeech.parts.Count];
IDictionaryEnumerator d = PartOfSpeech.parts.GetEnumerator();
while (d.MoveNext())
{
PartOfSpeech p = (PartOfSpeech)(d.Value);
excfps[p.ident] = new StreamReader(WNDB.ExcFile(p));
}
}
bool AddSearchFor(string s,PartOfSpeech pos, ArrayList list)
{
Search se = new Search(s,false,pos,new SearchType(false,"OVERVIEW"),0);
if(se.lexemes.Count > 0)
list.Add(se);
if(se.lexemes.Count > 0)
return true; // results were found
else
return false;
}
/// <summary>
/// Performs a search based on the parameters setup
/// in the Search constructor.
/// </summary>
/// <param name="m">Specify if morphs should be searched</param>
/// <param name="p">The Part Of Speech to perform the search on</param>
public void do_search(bool m, string p)
{
if (parts == null)
{
parts = new Hashtable();
}
Search s = new Search(word, PartOfSpeech.of(p), sch, whichsense);
s.do_search(m);
parts[p] = s;
buf += s.buf;
}
/// <summary>
/// Performs a search based on the parameters setup
/// in the Search constructor.
/// </summary>
/// <param name="doMorphs">Specify if morphs should be searched</param>
/// <param name="thePartOfSpeech">The Part Of Speech to perform the search on</param>
public void do_search(bool doMorphs, string thePartOfSpeech)
{
if (parts == null)
{
parts = new Hashtable();
}
Search s = new Search(word, PartOfSpeech.of(thePartOfSpeech), sch, whichsense);
s.do_search(doMorphs);
parts[thePartOfSpeech] = s;
buf += s.buf;
}
internal void strAnt(PartOfSpeech pos, int wdnum, string head, string tail)
{
int i, j, wdoff;
/* Go through all the pointers looking for anotnyms from the word
* indicated by wdnum. When found, print all the antonym's
* antonym pointers which point back to wdnum. */
for (i = 0; i < ptrs.Length; i++)
{
Pointer pt = ptrs[i];
if (pt.ptp.Ident == ANTPTR && pt.sce == wdnum)
{
SynSet psyn = new SynSet(pt.off, pos, this);
for (j = 0; j < psyn.ptrs.Length; j++)
{
Pointer ppt = psyn.ptrs[j];
if (ppt.ptp.Ident == ANTPTR &&
ppt.dst == wdnum &&
ppt.off == hereiam)
{
wdoff = ppt.sce > 0 ? ppt.sce - 1 : 0;
search.buf += head;
/* Construct buffer containing formatted antonym,
* then add it onto end of return buffer */
search.buf += deadjify(psyn.words[wdoff].word);
/* Print additional lexicographer information and
* WordNet sense number as indicated by flags */
isDirty = true; // TDMS 19 July 2006 - attempt to tie the logic which
// populates buf to the logic that defines whether the
// synset is populated with relevant information
if (search.prlexid && psyn.words[wdoff].uniq != 0)
{
search.buf += psyn.words[wdoff].uniq;
}
int s = getsearchsense(wdoff + 1);
psyn.words[wdoff].wnsns = s;
if (WnOpt.Opt("-s").Flag)
{
search.buf += "#" + s;
isDirty = true; // TDMS 19 July 2006 - attempt to tie the logic which
// populates buf to the logic that defines whether the
// synset is populated with relevant information
}
search.buf += tail;
}
}
}
}
}
void strAnt(string tail, AdjSynSetType attype, int definition)
{
int i, wdcnt;
bool first = true;
if (WnOpt.Opt("-o").Flag)
{
search.buf += "(" + hereiam + ") ";
}
if (WnOpt.Opt("-a").Flag)
{
search.buf += "<" + WNDB.lexfiles[fnum] + "> ";
search.prlexid = true;
}
else
{
search.prlexid = false;
}
/* print antonyms from cluster head (of indirect ant) */
search.buf += "INDIRECT (VIA ";
for (i = 0, wdcnt = words.Length; i < wdcnt; i++)
{
if (first)
{
strAnt(PartOfSpeech.of("adj"), i + 1, "", ", ");
first = false;
}
else
{
strAnt(PartOfSpeech.of("adj"), i + 1, ", ", ", ");
}
}
search.buf += ") -> ";
/* now print synonyms from cluster head (of indirect ant) */
for (i = 0, wdcnt = words.Length; i < wdcnt; i++)
{
catword(i, 0, 0);
if (i < wdcnt - 1)
{
search.buf += ", ";
}
}
if (WnOpt.Opt("-g").Flag&& defn != null && definition != 0)
{
search.buf += " -- " + defn;
isDirty = true; // TDMS 19 July 2006 - attempt to tie the logic which
// populates buf to the logic that defines whether the
// synset is populated with relevant information
}
search.buf += tail;
}
public MorphStr(string s, PartOfSpeech p)
{
string origstr = s;
pos = p;
if (pos.clss == "SATELLITE")
pos = PartOfSpeech.of("adj");
/* Assume string hasnt had spaces substitued with _ */
str = origstr.Replace(' ', '_').ToLower();
searchstr = "";
cnt = str.Split('_').Length;
svprep = 0;
firsttime = true;
}
static Exceptions()
{
IDictionaryEnumerator d = PartOfSpeech.parts.GetEnumerator();
while (d.MoveNext())
{
PartOfSpeech p = (PartOfSpeech)(d.Value);
if (!excfps.ContainsKey(p.key))
{
excfps[p.key] = WNDB.GetStreamReader(WNDB.ExcFile(p));
}
}
}
/// <summary>
/// Trace coordinate terms.
/// </summary>
/// <param name="ptp"></param>
/// <param name="fpos"></param>
/// <param name="depth"></param>
internal void traceCoords(PointerType ptp, PartOfSpeech fpos, int depth)
{
for (int i = 0; i < ptrs.Length; i++)
{
Pointer pt = ptrs[i];
/* WN2.0
* if (pt.ptp.mnemonic=="HYPERPTR" &&
* (pt.sce==0 ||
* pt.sce==whichword))
*/
// WN2.1 if statement change - TDMS
if ((pt.ptp.Ident == HYPERPTR || pt.ptp.Ident == INSTANCE) &&
((pt.sce == 0) ||
(pt.sce == whichword)))
{
if (!search.prflag)
{
strsns(sense + 1);
search.prflag = true;
}
spaces("TRACEC", depth);
SynSet cursyn = new SynSet(pt.off, pt.pos, this);
search.wordsFrom(cursyn);
cursyn.str("-> ", "\n", 1, 0, 0, 1);
cursyn.tracePtrs(ptp, cursyn.pos, depth);
// TDMS 6 Oct 2005 - build hierarchical results
if (this.senses == null)
{
this.senses = new SynSetList();
}
cursyn.thisptr = pt; // TDMS 17 Nov 2005 - add this pointer type
this.senses.Add(cursyn);
if (depth > 0)
{
depth = depthcheck(depth);
cursyn.traceCoords(ptp, cursyn.pos, depth + 1);
// TDMS 6 Oct 2005 - build hierarchical results
// TODO: verify this
if (this.senses == null)
{
this.senses = new SynSetList();
}
cursyn.thisptr = pt; // TDMS 17 Nov 2005 - add this pointer type
this.senses.Add(cursyn);
}
}
}
}
public WNDBpart(PartOfSpeech p)
{
try
{
index = new StreamReader(WNDB.IndexFile(p));
data = new StreamReader(WNDB.DataFile(p));
fps[p] = this;
}
catch
{
// TODO: handle bad dict path
// ignore errors - as the user is locating the dictionary location
// wordnet classes are trying to instantiate based on an incorrect dict path
}
}
public Exceptions(string word,PartOfSpeech pos)
{
if(laststr != ((System.IO.FileStream)(excfps[pos.ident].BaseStream)).Name)
{
laststr = ((System.IO.FileStream)(excfps[pos.ident].BaseStream)).Name;
StreamReader fs = excfps[pos.ident];
Byte[] b = System.Text.Encoding.Unicode.GetBytes(fs.ReadToEnd());
excfps[pos.ident].BaseStream.Seek(0, SeekOrigin.Begin);
exc = new MemoryStream(b);
}
line = WNDB.binSearch(word,exc); //excfps[pos.ident]);
if (line!=null)
endlp = line.IndexOf(' ');
}
void strAnt(string tail, AdjSynSetType attype, int definition)
{
int i, wdcnt;
bool first = true;
if (WNOpt.opt("-o").flag)
{
search.buf += "(" + hereiam + ") ";
}
if (WNOpt.opt("-a").flag)
{
search.buf += "<" + WNDB.lexfiles[fnum] + "> ";
search.prlexid = true;
}
else
{
search.prlexid = false;
}
/* print antonyms from cluster head (of indirect ant) */
search.buf += "INDIRECT (VIA ";
for (i = 0, wdcnt = words.Length; i < wdcnt; i++)
{
if (first)
{
strAnt(PartOfSpeech.of("adj"), i + 1, "", ", ");
first = false;
}
else
{
strAnt(PartOfSpeech.of("adj"), i + 1, ", ", ", ");
}
}
search.buf += ") -> ";
/* now print synonyms from cluster head (of indirect ant) */
for (i = 0, wdcnt = words.Length; i < wdcnt; i++)
{
catword(i, 0, 0);
if (i < wdcnt - 1)
{
search.buf += ", ";
}
}
if (WNOpt.opt("-g").flag&& defn != null && definition != 0)
{
search.buf += " -- " + defn;
}
search.buf += tail;
}
public static MemoryStream index(PartOfSpeech p) //StreamReader index(PartOfSpeech p)
{
if (laststr == ((System.IO.FileStream)(WNDBpart.of(p).index.BaseStream)).Name)
{
return(ms); //WNDBpart.of(p).index;
}
else
{
laststr = ((System.IO.FileStream)(WNDBpart.of(p).index.BaseStream)).Name;
StreamReader fs = WNDBpart.of(p).index;
fs.BaseStream.Position = 0;
Byte[] b = System.Text.Encoding.Unicode.GetBytes(fs.ReadToEnd());
ms = new MemoryStream(b);
return(ms); //WNDBpart.of(p).index;
}
}
public MorphStr(string s, PartOfSpeech p)
{
string origstr = s;
pos = p;
if (pos.clss == "SATELLITE")
{
pos = PartOfSpeech.of("adj");
}
/* Assume string hasnt had spaces substitued with _ */
str = origstr.Replace(' ', '_').ToLower();
searchstr = "";
cnt = str.Split('_').Length;
svprep = 0;
firsttime = true;
}
public void beginDisambiguate()
{
Disambiguate(SParseTrees);
///////////////////////////get the text of senses ///////////////////////
for (int i = 0; i < NewParseTreeSenses.Count; i++)
{
MyWordInfo[] mwiArr = (MyWordInfo[])NewParseTreeSenses[i];
ParseTree pt;
pt = (ParseTree)NewSParseTrees[i];
AddArrStems(NewSParseTrees);
for (int j = 0; j < mwiArr.Length; j++)
{
Wnlib.PartOfSpeech p = Wnlib.PartOfSpeech.of((Wnlib.PartsOfSpeech)mwiArr[j].Pos);
try
{
//i need the stems here to get index
Wnlib.Index index = Wnlib.Index.lookup(mwiArr[j].Word.ToLower(), p);
SynSet sense = new SynSet(index, mwiArr[j].Sense, null);
NodesSenses.Add(sense.defn);
SensesNos.Add(mwiArr[j].Sense);
string s = mwiArr[j].Word.ToLower() + " : " + sense.defn;
DisambRes.Add(s);
}
catch
{
try
{
Wnlib.Index index = Wnlib.Index.lookup(Stems[j], p);
SynSet sense = new SynSet(index, mwiArr[j].Sense, null);
NodesSenses.Add(sense.defn);
SensesNos.Add(mwiArr[j].Sense);
string s = Stems[j].ToLower() + " : " + sense.defn;
DisambRes.Add(s);
}
catch
{ };
};
}
Senses = NodesSenses;
}
//////////////////////////add sense text & sense no to the nodes//////////////////////////////
AddNodesSenses(NewSParseTrees);
//////////////////////////put the output parsetrees in SparseTree again//////////////////////////////
SParseTrees = NewSParseTrees;
}
public Exceptions(string word, PartOfSpeech pos)
{
if (laststr != ((System.IO.FileStream)(excfps[pos.ident].BaseStream)).Name)
{
laststr = ((System.IO.FileStream)(excfps[pos.ident].BaseStream)).Name;
StreamReader fs = excfps[pos.ident];
Byte[] b = System.Text.Encoding.Unicode.GetBytes(fs.ReadToEnd());
excfps[pos.ident].BaseStream.Seek(0, SeekOrigin.Begin);
exc = new MemoryStream(b);
}
line = WNDB.binSearch(word, exc); //excfps[pos.ident]);
if (line != null)
{
endlp = line.IndexOf(' ');
}
}
Opt(string a, string m, string p, int h, string b)
{
arg = a;
if (m[0] == '-')
{
sch = new SearchType(true, m.Substring(1));
}
else
{
sch = new SearchType(false, m);
}
pos = PartOfSpeech.of(p.ToLower());
helpx = h;
label = b;
id = opts.Count;
opts.Add(this);
}
internal void strAnt(PartOfSpeech pos, int wdnum, string head, string tail)
{
int i, j, wdoff;
/* Go through all the pointers looking for anotnyms from the word
* indicated by wdnum. When found, print all the antonym's
* antonym pointers which point back to wdnum. */
for (i = 0; i < ptrs.Length; i++)
{
Pointer pt = ptrs[i];
if (pt.ptp.mnemonic == "ANTPTR" && pt.sce == wdnum)
{
SynSet psyn = new SynSet(pt.off, pos, this);
for (j = 0; j < psyn.ptrs.Length; j++)
{
Pointer ppt = psyn.ptrs[j];
if (ppt.ptp.mnemonic == "ANTPTR" &&
ppt.dst == wdnum &&
ppt.off == hereiam)
{
wdoff = ppt.sce > 0?ppt.sce - 1:0;
search.buf += head;
/* Construct buffer containing formatted antonym,
* then add it onto end of return buffer */
search.buf += deadjify(psyn.words[wdoff].word);
/* Print additional lexicographer information and
* WordNet sense number as indicated by flags */
if (search.prlexid && psyn.words[wdoff].uniq != 0)
{
search.buf += psyn.words[wdoff].uniq;
}
int s = getsearchsense(wdoff + 1);
psyn.words[wdoff].wnsns = s;
if (WNOpt.opt("-s").flag)
{
search.buf += "#" + s;
}
search.buf += tail;
}
}
}
}
}
/* From search.c:
* Find word in index file and return parsed entry in data structure.
* Input word must be exact match of string in database. */
// From the WordNet Manual (http://wordnet.princeton.edu/man/wnsearch.3WN.html)
// index_lookup() finds searchstr in the index file for pos and returns a pointer
// to the parsed entry in an Index data structure. searchstr must exactly match the
// form of the word (lower case only, hyphens and underscores in the same places) in
// the index file. NULL is returned if a match is not found.
public static Index lookup(string word, PartOfSpeech pos)
{
int j;
if (word == "")
{
return(null);
}
// TDMS 14 Aug 2005 - changed to allow for numbers as well
// because the database contains searches that can start with
// numerals
//if (!char.IsLetter(word[0]))
if (!char.IsLetter(word[0]) && !char.IsNumber(word[0]))
{
return(null);
}
string line = WNDB.binSearch(word, pos);
if (line == null)
{
return(null);
}
Index idx = new Index();
StrTok st = new StrTok(line);
idx.wd = st.next(); /* the word */
idx.pos = PartOfSpeech.of(st.next()); /* the part of speech */
idx.sense_cnt = int.Parse(st.next()); /* collins count */
int ptruse_cnt = int.Parse(st.next()); /* number of pointers types */
idx.ptruse = new PointerType[ptruse_cnt];
for (j = 0; j < ptruse_cnt; j++)
{
idx.ptruse[j] = PointerType.of(st.next());
}
int off_cnt = int.Parse(st.next());
idx.offs = new int[off_cnt];
idx.tagsense_cnt = int.Parse(st.next());
for (j = 0; j < off_cnt; j++)
{
idx.offs[j] = int.Parse(st.next());
}
return(idx);
}
static WNDB()
{
//Console.WriteLine("WNDB");
path = WNCommon.path;
IDictionaryEnumerator d = PartOfSpeech.parts.GetEnumerator();
while (d.MoveNext())
{
PartOfSpeech p = (PartOfSpeech)(d.Value);
if (!indexfps.ContainsKey(p.key))
{
indexfps[p.key] = GetStreamReader(IndexFile(p));
}
if (!datafps.ContainsKey(p.key))
{
datafps[p.key] = GetStreamReader(DataFile(p));
}
}
}
public static PartOfSpeech of(PartsOfSpeech f)
{
if (f == PartsOfSpeech.Noun)
{
return(PartOfSpeech.of("noun"));
}
if (f == PartsOfSpeech.Verb)
{
return(PartOfSpeech.of("verb"));
}
if (f == PartsOfSpeech.Adj)
{
return(PartOfSpeech.of("adj"));
}
if (f == PartsOfSpeech.Adv)
{
return(PartOfSpeech.of("adv"));
}
return(null); // unknown or not unique
}
/// <summary>
/// Traces pointer hierarchy.
/// </summary>
/// <param name="stp"></param>
/// <param name="fpos"></param>
/// <param name="depth"></param>
internal void tracePtrs(SearchType stp, PartOfSpeech fpos, int depth)
{
int i;
SynSet cursyn;
PointerType ptp = stp.ptp;
string prefix;
int realptr; // WN2.1
for (i = 0; i < ptrs.Length; i++)
{
Pointer pt = ptrs[i];
// following if statement is WN2.1 - TDMS
if ((ptp.ident == HYPERPTR && (pt.ptp.ident == HYPERPTR ||
pt.ptp.ident == INSTANCE)) ||
(ptp.ident == HYPOPTR && (pt.ptp.ident == HYPOPTR ||
pt.ptp.ident == INSTANCES)) ||
((pt.ptp == ptp) &&
((pt.sce == 0) ||
(pt.sce == whichword))))
{
realptr = pt.ptp.ident; /* WN2.1 deal with INSTANCE */
if (!search.prflag) // print sense number and synset
strsns(sense + 1);
search.prflag = true;
spaces("TRACEP", depth + (stp.rec ? 2 : 0));
// switch (ptp.mnemonic)
// TDMS 11 JUL 2006 - changed switch to ident switch (pt.ptp.mnemonic) // TDMS - WN2.1 MOD
switch (pt.ptp.ident) // TDMS 11 JUL 2006 - changed switch to ident
{
case PERTPTR:
if (fpos.name == "adv") // TDMS "adverb")
prefix = "Derived from " + pt.pos.name + " ";
else
prefix = "Pertains to " + pt.pos.name + " ";
break;
case ANTPTR: // TDMS 26/8/05
if (fpos.name == "adj") //TODO: which adjective will fall into the below?
prefix = "Antonym of ";
else
prefix = "";
break;
case PPLPTR:
prefix = "Participle of verb";
break;
case INSTANCE:
prefix = "INSTANCE OF=> ";
break;
case INSTANCES:
prefix = "HAS INSTANCE=> ";
break;
case HASMEMBERPTR:
prefix = " HAS MEMBER: ";
break;
case HASSTUFFPTR:
prefix = " HAS SUBSTANCE: ";
break;
case HASPARTPTR:
prefix = " HAS PART: ";
break;
case ISMEMBERPTR:
prefix = " MEMBER OF: ";
break;
case ISSTUFFPTR: // TDMS 26/8/05
prefix = " SUBSTANCE OF: ";
break;
case ISPARTPTR: // TDMS 26/8/05
prefix = " PART OF: ";
break;
default:
prefix = "=> ";
break;
}
/* Read synset pointed to */
cursyn = new SynSet(pt.off, pt.pos, this);
search.wordsFrom(cursyn);
// TDMS 6 Oct 2005 - build hierarchical results
if (this.senses == null)
this.senses = new SynSetList();
cursyn.thisptr = pt; // TDMS 17 Nov 2005 - add this pointer type
this.senses.Add(cursyn);
/* For Pertainyms and Participles pointing to a specific
sense, indicate the sense then retrieve the synset
pointed to and other info as determined by type.
Otherwise, just print the synset pointed to. */
if ((ptp.ident == PERTPTR || ptp.ident == PPLPTR) &&
pt.dst != 0)
{
string tbuf = " (Sense " + cursyn.getsearchsense(pt.dst) + ")";
cursyn.str(prefix, tbuf, 0, pt.dst, 0, 1);
if (ptp.ident == PPLPTR) // adj pointing to verb
{
cursyn.str(" =>", "\n", 1, 0, 1, 1);
cursyn.tracePtrs(PointerType.of("HYPERPTR"), cursyn.pos, 0);
}
else if (fpos.name == "adv") // adverb pointing to adjective
{
cursyn.str(" =>", "\n", 0, 0, (pos.clss == "SATELLITE") ? 0 : 1, 1);
// cursyn.traceptrs(HYPERPTR,pos,0);
}
else // adjective pointing to noun
{
cursyn.str(" =>", "\n", 1, 0, 1, 1);
cursyn.tracePtrs(PointerType.of("HYPERPTR"), pos, 0);
}
}
else
cursyn.str(prefix, "\n", 1, 0, 1, 1);
/* For HOLONYMS and MERONYMS, keep track of last one
printed in buffer so results can be truncated later. */
if (ptp.ident >= PointerType.of("ISMEMBERPTR").ident &&
ptp.ident <= PointerType.of("HASPARTPTR").ident)
{
search.mark();
}
if (depth > 0)
{
depth = cursyn.depthcheck(depth);
cursyn.tracePtrs(ptp, cursyn.pos, depth + 1);
}
}
}
}
/// <summary>
/// Trace coordinate terms.
/// </summary>
/// <param name="ptp"></param>
/// <param name="fpos"></param>
/// <param name="depth"></param>
internal void traceCoords(PointerType ptp, PartOfSpeech fpos, int depth)
{
for (int i = 0; i < ptrs.Length; i++)
{
Pointer pt = ptrs[i];
/* WN2.0
if (pt.ptp.mnemonic=="HYPERPTR" &&
(pt.sce==0 ||
pt.sce==whichword))
*/
// WN2.1 if statement change - TDMS
if ((pt.ptp.ident == HYPERPTR || pt.ptp.ident == INSTANCE) &&
((pt.sce == 0) ||
(pt.sce == whichword)))
{
if (!search.prflag)
{
strsns(sense + 1);
search.prflag = true;
}
spaces("TRACEC", depth);
SynSet cursyn = new SynSet(pt.off, pt.pos, this);
search.wordsFrom(cursyn);
cursyn.str("-> ", "\n", 1, 0, 0, 1);
cursyn.tracePtrs(ptp, cursyn.pos, depth);
// TDMS 6 Oct 2005 - build hierarchical results
if (this.senses == null)
this.senses = new SynSetList();
cursyn.thisptr = pt; // TDMS 17 Nov 2005 - add this pointer type
this.senses.Add(cursyn);
if (depth > 0)
{
depth = depthcheck(depth);
cursyn.traceCoords(ptp, cursyn.pos, depth + 1);
// TDMS 6 Oct 2005 - build hierarchical results
// TODO: verify this
if (this.senses == null)
this.senses = new SynSetList();
cursyn.thisptr = pt; // TDMS 17 Nov 2005 - add this pointer type
this.senses.Add(cursyn);
}
}
}
}
public static string binSearch(string word,PartOfSpeech pos)
{
return binSearch(word,WNDB.index(pos));
}
/// <summary>
/// Trace meronyms.
/// </summary>
/// <param name="pbase"></param>
/// <param name="fpos"></param>
/// <param name="depth"></param>
void traceInherit(PointerType pbase, PartOfSpeech fpos, int depth)
{
for (int i = 0; i < ptrs.Length; i++)
{
Pointer pt = ptrs[i];
if (pt.ptp.ident == HYPERPTR && (pt.sce == 0 || pt.sce == whichword))
{
spaces("TRACEI", depth);
SynSet cursyn = new SynSet(pt.off, pt.pos, this);
search.wordsFrom(cursyn);
cursyn.str("=> ", "\n", 1, 0, 0, 1);
// TDMS 6 Oct 2005 - build hierarchical results
// TODO: verify this
if (this.senses == null)
this.senses = new SynSetList();
cursyn.thisptr = pt; // TDMS 17 Nov 2005 - add this pointer type
// TODO: This is adding senses incorrectly
this.senses.Add(cursyn);
cursyn.tracePtrs(pbase, PartOfSpeech.of("noun"), depth);
cursyn.tracePtrs(pbase + 1, PartOfSpeech.of("noun"), depth);
cursyn.tracePtrs(pbase + 2, PartOfSpeech.of("noun"), depth);
if (depth > 0)
{
depth = depthcheck(depth);
cursyn.traceInherit(pbase, cursyn.pos, depth + 1);
}
}
}
search.trunc();
}
public static void reopen(PartOfSpeech p)
{
WNDBpart w = WNDBpart.of(p);
w.index.Close();
w.data.Close();
w.index = new StreamReader(IndexFile(p));
w.data = new StreamReader(DataFile(p));
}
internal static ArrayList wngrep(string wordPassed, PartOfSpeech pos)
{
ArrayList r = new ArrayList();
StreamReader fp = index(pos);
fp.BaseStream.Position = 0;
fp.DiscardBufferedData();
string word = wordPassed.Replace(" ", "_");
string line;
while ((line = fp.ReadLine()) != null)
{
int lineLen = line.IndexOf(' ');
line = line.Substring(0, lineLen);
try
{
if (line.IndexOf(word) >= 0)
r.Add(line.Replace("_", " "));
}
catch
{
}
}
return r;
}
public static WNDBpart of(PartOfSpeech p)
{
return (WNDBpart)fps[p];
}
public Search(string w, bool doMorphs, PartOfSpeech p, SearchType s, int sn)
: this(w, p, s, sn)
{
if (p != null)
do_search(doMorphs);
}
internal static ArrayList wngrep(string wordPassed,PartOfSpeech pos)
{
ArrayList r = new ArrayList();
MemoryStream ms = index(pos);
StreamReader inputFile = new StreamReader(ms); //= index(pos);
inputFile.BaseStream.Seek(0L,SeekOrigin.Begin);
inputFile.DiscardBufferedData();
string word = wordPassed.Replace(" ","_");
string line;
while ((line=inputFile.ReadLine())!=null)
{
int lineLen = line.IndexOf(' ');
line = line.Substring(0,lineLen);
try
{
if (line.IndexOf(word)>=0)
r.Add(line.Replace("_"," "));
}
catch
{
}
}
return r;
}
internal static string ExcFile(PartOfSpeech n)
{
return path+n.name+".exc";
}
public static StreamReader index(PartOfSpeech p)
{
return (StreamReader)indexfps[p.key];
}
public static StreamReader data(PartOfSpeech p)
{
return (StreamReader)datafps[p.key];
}
public static void reopen(PartOfSpeech p)
{
index(p).Close();
data(p).Close();
indexfps[p.key] = GetStreamReader(IndexFile(p));
datafps[p.key] = GetStreamReader(DataFile(p));
}
Index[] offsets = new Index[stringscount]; // of Index
#endregion Fields
#region Constructors
public Indexes(string str,PartOfSpeech pos)
{
string[] strings = new string[stringscount];
str = str.ToLower();
strings[0] = str;
strings[1] = str.Replace('_','-');
strings[2] = str.Replace('-','_');
strings[3] = str.Replace("-","").Replace("_","");
strings[4] = str.Replace(".","");
// TDMS - 14 Aug 2005 - added two more possibilities
// to allow for spaces to be transformed
// an example is a search for "11 plus", without this
// transformation no result would be found
strings[5] = str.Replace(' ','-');
strings[6] = str.Replace(' ','_');
offsets[0] = Index.lookup(str,pos);
// TDMS - 14 Aug 2005 - changed 5 to 7 to allow for two
// new possibilities
for (int i=1;i<stringscount;i++)
if (str!=strings[i])
offsets[i] = Index.lookup(strings[i],pos);
fpos = pos;
}
Opt(string a,string m,string p,int h,string b)
{
arg = a;
if (m[0]=='-')
sch = new SearchType(true,m.Substring(1));
else
sch = new SearchType(false,m);
pos = PartOfSpeech.of(p.ToLower());
helpx = h;
label = b;
id = opts.Count;
opts.Add(this);
}
internal static string DataFile(PartOfSpeech n)
{
return path+"data."+n.name; // WN2.1 - TDMS
}
public WNHelp(SearchType s, PartOfSpeech p)
: this((s.rec ? "-" : "") + s.ptp.mnemonic, p)
{
}
// TDMS 16 July 2006 - removed this method.
// Method removed because if called externally
// WNDBPart was not correctly constructed.
// Calling is_defined(string searchstr,PartOfSpeech fpos)
// correctly constructs WNDBPart.
/*
private static SearchSet is_defined(string word,string p)
{
Console.WriteLine("is_defined string, string");
return is_defined(word,PartOfSpeech.of(p));
}
*/
/// <summary>
/// Determines if a word is defined in the WordNet database and returns
/// all possible searches of the word.
/// </summary>
/// <example> This sample displays a message stating whether the
/// word "car" exists as the part of speech "noun".
/// <code>
/// Wnlib.WNCommon.path = "C:\Program Files\WordNet\2.1\dict\"
/// Dim wrd As String = "car"
/// Dim POS As String = "noun"
/// Dim b As Boolean = Wnlib.WNDB.is_defined(wrd, Wnlib.PartOfSpeech.of(POS)).NonEmpty.ToString
///
/// If b Then
/// MessageBox.Show("The word " & wrd & " exists as a " & POS & ".")
/// Else
/// MessageBox.Show("The word " & wrd & " does not exist as a " & POS & ".")
/// End If
/// </code>
/// </example>
/// <param name="searchstr">The word to search for</param>
/// <param name="fpos">Part of Speech (noun, verb, adjective, adverb)</param>
/// <returns>A SearchSet or null if the word does not exist in the dictionary</returns>
public static SearchSet is_defined(string searchstr,PartOfSpeech fpos)
{
Indexes ixs = new Indexes(searchstr,fpos);
Index index;
int i;
int CLASS = 22; /* - */
int LASTTYPE = CLASS;
Search s = new Search(searchstr,fpos,new SearchType(false,"FREQ"),0);
SearchSet retval = new SearchSet();
while ((index=ixs.next())!=null)
{
retval=retval+"SIMPTR"+"FREQ"+"SYNS"+"WNGREP"+"OVERVIEW"; // added WNGREP - TDMS
for (i=0;i<index.ptruse.Length;i++)
{
PointerType pt = index.ptruse[i];
// retval=retval+pt;
// WN2.1 - TDMS
if (pt.ident <= LASTTYPE) {
retval = retval + pt;
} else if (pt.mnemonic == "INSTANCE") {
retval = retval + "HYPERPTR";
} else if (pt.mnemonic == "INSTANCES") {
retval = retval + "HYPOPTR";
}
// WN2.1 - TDMS
if (pt.mnemonic == "SIMPTR") {
retval = retval + "ANTPTR";
}
if (fpos.name=="noun")
{
/* set generic HOLONYM and/or MERONYM bit if necessary */
if (pt>="ISMEMBERPTR" && pt<="ISPARTPTR")
retval=retval+"HOLONYM";
else if (pt>="HASMEMBERPTR" && pt<="HASPARTPTR")
retval=retval+"MERONYM";
}
// WN2.1 - TDMS else if (fpos.name=="adj" && pt.mnemonic=="SIMPTR")
// retval=retval+"ANTPTR";
}
if (fpos.name=="noun")
{
retval=retval+"RELATIVES";
if (index.HasHoloMero("HMERONYM",s))
retval=retval+"HMERONYM";
if (index.HasHoloMero("HHOLONYM",s))
retval=retval+"HHOLONYM";
if (retval["HYPERPTR"])
retval = retval+"COORDS";
}
else if (fpos.name=="verb")
retval=retval+"RELATIVES"+"FRAMES"; // added frames - TDMS
}
return retval;
}
public WNDBpart(PartOfSpeech p)
{
try
{
Console.WriteLine("WNDBpart");
index = new StreamReader(WNDB.IndexFile(p));
data = new StreamReader(WNDB.DataFile(p));
fps[p] = this;
}
catch
{
MessageBox.Show("Bad dict path");
// TODO: handle bad dict path
// ignore errors - as the user is locating the dictionary location
// wordnet classes are trying to instantiate based on an incorrect dict path
}
}
internal void strAnt(PartOfSpeech pos, int wdnum, string head, string tail)
{
int i, j, wdoff;
/* Go through all the pointers looking for anotnyms from the word
indicated by wdnum. When found, print all the antonym's
antonym pointers which point back to wdnum. */
for (i = 0; i < ptrs.Length; i++)
{
Pointer pt = ptrs[i];
if (pt.ptp.ident == ANTPTR && pt.sce == wdnum)
{
SynSet psyn = new SynSet(pt.off, pos, this);
for (j = 0; j < psyn.ptrs.Length; j++)
{
Pointer ppt = psyn.ptrs[j];
if (ppt.ptp.ident == ANTPTR &&
ppt.dst == wdnum &&
ppt.off == hereiam)
{
wdoff = ppt.sce > 0 ? ppt.sce - 1 : 0;
search.buf += head;
/* Construct buffer containing formatted antonym,
then add it onto end of return buffer */
search.buf += deadjify(psyn.words[wdoff].word);
/* Print additional lexicographer information and
WordNet sense number as indicated by flags */
isDirty = true; // TDMS 19 July 2006 - attempt to tie the logic which
// populates buf to the logic that defines whether the
// synset is populated with relevant information
if (search.prlexid && psyn.words[wdoff].uniq != 0)
search.buf += psyn.words[wdoff].uniq;
int s = getsearchsense(wdoff + 1);
psyn.words[wdoff].wnsns = s;
if (WNOpt.opt("-s").flag)
{
search.buf += "#" + s;
isDirty = true; // TDMS 19 July 2006 - attempt to tie the logic which
// populates buf to the logic that defines whether the
// synset is populated with relevant information
}
search.buf += tail;
}
}
}
}
}
public WNHelp(string s, PartOfSpeech p)
{
help = (string)p.help[s];
}
//StreamReader index(PartOfSpeech p)
public static MemoryStream index(PartOfSpeech p)
{
if(laststr == ((System.IO.FileStream)(WNDBpart.of(p).index.BaseStream)).Name)
{
return ms; //WNDBpart.of(p).index;
}
else
{
laststr = ((System.IO.FileStream)(WNDBpart.of(p).index.BaseStream)).Name;
StreamReader fs = WNDBpart.of(p).index;
fs.BaseStream.Position = 0;
Byte[] b = System.Text.Encoding.Unicode.GetBytes(fs.ReadToEnd());
ms = new MemoryStream(b);
return ms; //WNDBpart.of(p).index;
}
}
public static StreamReader data(PartOfSpeech p)
{
return WNDBpart.of(p).data;
}
internal static string IndexFile(PartOfSpeech n)
{
return path+"index."+n.name; // WN2.1 - TDMS
}