/// <summary>
/// Create or modify a lexeme with no word form basis, gets tricky with best fit scenarios
/// </summary>
/// <param name="word">just the text of the word</param>
/// <returns>A lexeme</returns>
public static ILexeme CreateOrModifyLexeme(ILanguage language, string word, LexicalType wordType, ref List <string> processedWords)
{
word = word.ToLower();
Regex rgx = new Regex("[^a-z -]");
if (rgx.IsMatch(word))
{
return(null);
}
ILexeme newLex = ConfigDataCache.Get <ILexeme>(string.Format("{0}_{1}_{2}", ConfigDataType.Dictionary, language.Name, word));
if (newLex == null)
{
newLex = language.CreateOrModifyLexeme(word, wordType, new string[0]);
}
if ((newLex.IsSynMapped && newLex.MirriamIndexed) || processedWords.Any(wrd => wrd.Equals(word)))
{
if (!processedWords.Any(wrd => wrd.Equals(word)))
{
processedWords.Add(word);
}
}
else
{
LexicalType[] invalidTypes = new LexicalType[] { LexicalType.Article, LexicalType.Conjunction, LexicalType.ProperNoun, LexicalType.Pronoun, LexicalType.None };
processedWords.Add(word);
//This is wordnet processing, wordnet doesnt have any of the above and will return weird results if we let it
if (!invalidTypes.Contains(wordType))
{
var synSets = WordNetHarness.GetSynSets(word, new PartOfSpeech[] { PartOfSpeech.Adjective, PartOfSpeech.Adverb, PartOfSpeech.Noun, PartOfSpeech.Verb });
//We in theory have every single word form for this word now
if (synSets != null)
{
SemanticContext[] invalidContexts = new SemanticContext[]
{ SemanticContext.Group, SemanticContext.Event, SemanticContext.Location, SemanticContext.Competition, SemanticContext.Person
, SemanticContext.Plant, SemanticContext.Animal, SemanticContext.Time, SemanticContext.Artifact };
foreach (SynSet synSet in synSets)
{
if (synSet.PartOfSpeech == PartOfSpeech.None)
{
continue;
}
var synContext = TranslateContext(synSet.LexicographerFileName);
if (invalidContexts.Contains(synContext))
{
continue;
}
var newDict = newLex.GetForm(MapLexicalTypes(synSet.PartOfSpeech), -1);
if (newDict == null)
{
newLex = language.CreateOrModifyLexeme(word, MapLexicalTypes(synSet.PartOfSpeech), new string[0]);
newDict = newLex.GetForm(MapLexicalTypes(synSet.PartOfSpeech), -1);
newDict.Context = TranslateContext(synSet.LexicographerFileName);
}
//We're going to use the definition from here
if (!string.IsNullOrWhiteSpace(synSet.Gloss))
{
newDict.Definition = synSet.Gloss;
}
///wsns indicates hypo/hypernymity so
foreach (string synWord in synSet.Words)
{
var newWord = synWord.ToLower();
newWord = newWord.Replace("_", " ");
if (rgx.IsMatch(newWord))
{
continue;
}
int myElegance = Math.Max(0, newWord.SyllableCount() * 3);
processedWords.Add(newWord);
if (string.IsNullOrWhiteSpace(newWord) || newWord.All(ch => ch == '-') || newWord.IsNumeric())
{
continue;
}
var synLex = language.CreateOrModifyLexeme(newWord, MapLexicalTypes(synSet.PartOfSpeech), newDict.Semantics.ToArray());
var synDict = synLex.GetForm(MapLexicalTypes(synSet.PartOfSpeech), newDict.Semantics.ToArray(), false);
synDict.Elegance = 0;
synDict.Quality = 0;
synDict.Severity = 0;
synDict.Context = synContext;
synDict.Definition = newDict.Definition;
synLex.PersistToCache();
synLex.SystemSave();
if (!newWord.Equals(word))
{
newDict.MakeRelatedWord(language, newWord, true, synDict);
}
}
}
}
}
newLex.IsSynMapped = true;
newLex.SystemSave();
newLex.PersistToCache();
}
if (!newLex.MirriamIndexed)
{
var newDict = newLex.GetForm(0);
try
{
var dictEntry = MirriamWebsterAPI.GetDictionaryEntry(newLex.Name);
if (dictEntry != null)
{
//Stuff done to modify all forms of the lexeme
foreach (var dict in newLex.WordForms)
{
dict.Vulgar = dictEntry.meta.offensive;
}
//Stuff done based on the dictionary return data
foreach (var stemWord in dictEntry.uros)
{
if (newLex.GetForm(MapLexicalTypes(stemWord.fl)) == null)
{
var wordText = stemWord.ure.Replace("*", "");
ILexeme stemLex = ConfigDataCache.Get <ILexeme>(string.Format("{0}_{1}_{2}", ConfigDataType.Dictionary, language.Name, wordText));
if (stemLex == null)
{
stemLex = language.CreateOrModifyLexeme(wordText, MapLexicalTypes(stemWord.fl), null);
var stemDict = stemLex.GetForm(0);
stemDict.Elegance = newDict.Elegance;
stemDict.Quality = newDict.Quality;
stemDict.Severity = newDict.Severity;
stemDict.Context = newDict.Context;
stemDict.Definition = newDict.Definition;
stemDict.Semantics = newDict.Semantics;
processedWords.Add(wordText);
stemLex.SystemSave();
stemLex.PersistToCache();
}
}
}
newDict.Semantics = new HashSet <string>(dictEntry.sls);
}
}
catch
{
//just eating it
}
try
{
var thesEntry = MirriamWebsterAPI.GetThesaurusEntry(newLex.Name);
if (thesEntry != null)
{
foreach (var synonym in thesEntry.meta.syns.SelectMany(syn => syn))
{
var newWord = synonym.ToLower();
newWord = newWord.Replace("_", " ");
if (rgx.IsMatch(newWord) || string.IsNullOrWhiteSpace(newWord) || newWord.All(ch => ch == '-'))
{
continue;
}
var synLex = language.CreateOrModifyLexeme(newWord, MapLexicalTypes(thesEntry.fl), newDict.Semantics.ToArray());
var synDict = synLex.GetForm(MapLexicalTypes(thesEntry.fl), newDict.Semantics.ToArray(), false);
synDict.Elegance = 0;
synDict.Quality = 0;
synDict.Severity = 0;
synDict.Context = newDict.Context;
synDict.Definition = newDict.Definition;
synLex.PersistToCache();
synLex.SystemSave();
processedWords.Add(newWord);
if (!newWord.Equals(word))
{
newDict.MakeRelatedWord(language, newWord, true, synDict);
}
}
foreach (var antonym in thesEntry.meta.ants.SelectMany(syn => syn))
{
var newWord = antonym.ToLower();
newWord = newWord.Replace("_", " ");
if (rgx.IsMatch(newWord) || string.IsNullOrWhiteSpace(newWord) || newWord.All(ch => ch == '-'))
{
continue;
}
var synLex = language.CreateOrModifyLexeme(newWord, MapLexicalTypes(thesEntry.fl), newDict.Semantics.ToArray());
var synDict = synLex.GetForm(MapLexicalTypes(thesEntry.fl), newDict.Semantics.ToArray(), false);
synDict.Elegance = 0;
synDict.Quality = 0;
synDict.Severity = 0;
synDict.Context = newDict.Context;
synDict.Definition = newDict.Definition;
synLex.PersistToCache();
synLex.SystemSave();
processedWords.Add(newWord);
if (!newWord.Equals(word))
{
newDict.MakeRelatedWord(language, newWord, false, synDict);
}
}
}
}
catch
{
//just eating it
}
newLex.MirriamIndexed = true;
newLex.SystemSave();
newLex.PersistToCache();
}
if (!newLex.IsTranslated)
{
}
return(newLex);
}
/// <summary>
/// Create or modify a lexeme with no word form basis, gets tricky with best fit scenarios
/// </summary>
/// <param name="word">just the text of the word</param>
/// <returns>A lexeme</returns>
public static ILexeme CreateOrModifyLexeme(ILanguage language, string word, ref List <string> processedWords)
{
word = word.ToLower();
Regex rgx = new Regex("[^a-z -]");
word = rgx.Replace(word, "");
if (string.IsNullOrWhiteSpace(word) || word.All(ch => ch == '-'))
{
return(null);
}
ILexeme newLex = ConfigDataCache.Get <ILexeme>(string.Format("{0}_{1}", language.Name, word));
if (newLex == null)
{
newLex = language.CreateOrModifyLexeme(word, LexicalType.None, new string[0]);
}
if (newLex.IsSynMapped || processedWords.Any(wrd => wrd.Equals(word)))
{
if (!processedWords.Any(wrd => wrd.Equals(word)))
{
processedWords.Add(word);
}
return(newLex);
}
processedWords.Add(word);
bool exists = true;
SearchSet searchSet = null;
List <Search> results = new List <Search>();
WordNet.OverviewFor(word, string.Empty, ref exists, ref searchSet, results);
//We in theory have every single word form for this word now
if (exists && results != null)
{
LexicalType[] invalidTypes = new LexicalType[] { LexicalType.Article, LexicalType.Conjunction, LexicalType.ProperNoun, LexicalType.Pronoun, LexicalType.None };
foreach (SynonymSet synSet in results.SelectMany(result => result.senses))
{
//grab semantics somehow
List <string> semantics = new List <string>();
var indexSplit = synSet.defn.IndexOf(';');
string definition = synSet.defn.Substring(0, indexSplit < 0 ? synSet.defn.Length - 1 : indexSplit).Trim();
string[] defWords = definition.Split(' ');
foreach (string defWord in defWords)
{
var currentWord = defWord.ToLower();
currentWord = rgx.Replace(currentWord, "");
if (currentWord.Equals(word) || string.IsNullOrWhiteSpace(word) || word.All(ch => ch == '-') || word.IsNumeric())
{
continue;
}
var defLex = language.CreateOrModifyLexeme(currentWord, LexicalType.None, new string[0]);
if (defLex != null && !defLex.ContainedTypes().Any(typ => invalidTypes.Contains(typ)))
{
semantics.Add(currentWord);
}
}
var type = MapLexicalTypes(synSet.pos.Flag);
newLex = language.CreateOrModifyLexeme(word, type, semantics.ToArray());
var newDict = newLex.GetForm(type, semantics.ToArray(), false);
///wsns indicates hypo/hypernymity so
int baseWeight = synSet.words[Math.Max(0, synSet.whichword - 1)].wnsns;
newDict.Severity = baseWeight;
newDict.Elegance = Math.Max(0, newDict.Name.SyllableCount() * 3);
newDict.Quality = synSet.words.Count();
foreach (Lexeme synWord in synSet.words)
{
///wsns indicates hypo/hypernymity so
int mySeverity = synWord.wnsns;
int myElegance = Math.Max(0, synWord.word.SyllableCount() * 3);
int myQuality = synWord.semcor?.semcor ?? 0;
//Don't bother if this word is already the same word we started with
if (synWord.word.Equals(newDict.Name, StringComparison.InvariantCultureIgnoreCase))
{
continue;
}
//it's a phrase
if (synWord.word.Contains("_"))
{
string[] words = synWord.word.Split('_');
//foreach (string phraseWord in words)
//{
// //make the phrase? maybe later
//}
}
else
{
var newWord = synWord.word.ToLower();
newWord = rgx.Replace(newWord, "");
if (newWord.Equals(word) || string.IsNullOrWhiteSpace(newWord) || newWord.All(ch => ch == '-') || newWord.IsNumeric())
{
continue;
}
newDict.MakeRelatedWord(language, synWord.word, true);
}
}
}
}
newLex.IsSynMapped = true;
newLex.SystemSave();
newLex.PersistToCache();
return(newLex);
}
