public static List <Phrase> GetPrePreTerminalPhrases(this Tree root, Rhetorica.Sentence sentence = null, string ignore = "", string punctuation = null, AnalyzerOptions options = AnalyzerOptions.None)
{
var phrases = new List <Phrase>();
for (java.util.Iterator i = root.iterator(); i.hasNext();)
{
Tree tree = (Tree)i.next();
if (tree.isPreTerminal() || tree.isPrePreTerminal())
{
if (tree.isPreTerminal() && tree.parent(root) != null)
{
if (tree.parent(root).isPrePreTerminal())
{
continue;
}
}
var current = new Phrase(tree.GetTokens(root, sentence, ignore, punctuation, options));
// If current node matches previous node but for punctuation omission, replace previous with current:
bool omitFalseDuplicatePhrases = options.HasFlag(AnalyzerOptions.OmitFalseDuplicatePhrases);
if (omitFalseDuplicatePhrases)
{
if (phrases.Count > 0)
{
Phrase previous = phrases.Last();
if (previous.EqualExceptPunctuationOmission(current))
{
phrases[phrases.Count - 1] = current;
continue;
}
}
}
if (current.Count == 0)
{
continue;
}
phrases.Add(current);
}
}
// If "phrase" is a single token which is a preposition (IN) or infinitival to (TO), then join it to the subsequent phrase.
for (int i = 0; i < phrases.Count; ++i)
{
if (phrases[i].Count == 1 && Regex.IsMatch(phrases[i][0].TagEquivalent, @"^(IN|TO)$", RegexOptions.IgnoreCase) && i != phrases.Count - 1)
{
phrases[i + 1].Tokens.InsertRange(0, phrases[i].Tokens);
phrases.RemoveAt(i);
i = -1;
}
}
return(phrases);
}
public static List <Phrase> GetClauses(this Tree root, Rhetorica.Sentence sentence = null, string ignore = "", string punctuation = null, AnalyzerOptions options = AnalyzerOptions.None)
{
var phrases = new List <Phrase>();
for (java.util.Iterator i = root.iterator(); i.hasNext();)
{
Tree tree = (Tree)i.next();
var treeLabel = tree.label().value();
var clauseRe = @"^(S|SBAR|SBARQ|SINV|SQ|FRAG)$";
bool isClausal = Regex.IsMatch(treeLabel, clauseRe, RegexOptions.IgnoreCase);
if (isClausal)
{
var current = new Phrase(tree.GetTokens(root, sentence, ignore, punctuation, options));
// If current node matches previous node but for punctuation omission, replace previous with current:
bool omitFalseDuplicatePhrases = options.HasFlag(AnalyzerOptions.OmitFalseDuplicatePhrases);
if (omitFalseDuplicatePhrases)
{
if (phrases.Count > 0)
{
Phrase previous = phrases.Last();
if (previous.EqualExceptPunctuationOmission(current))
{
phrases[phrases.Count - 1] = current;
continue;
}
}
}
if (current.Count == 0)
{
continue;
}
phrases.Add(current);
}
}
if (phrases.Count == 0) // Since 'root' has been identified as a sentence, it should have at least one clause associated with it.
{
var pseudoClauses = root.GetPhrases(sentence, ignore, punctuation, options);
if (pseudoClauses.Count > 0)
{
phrases.Add(pseudoClauses[0]);
}
}
return(phrases);
}
public static List <Phrase> GetPhrases(this Tree root, Rhetorica.Sentence sentence = null, string ignore = "", string punctuation = null, AnalyzerOptions options = AnalyzerOptions.None)
{
var phrases = new List <Phrase>();
for (java.util.Iterator i = root.iterator(); i.hasNext();)
{
Tree tree = (Tree)i.next();
if (tree.isPhrasal())
{
java.util.List children = tree.getChildrenAsList();
if (children.size() == 1 && ((Tree)children.get(0)).isPhrasal())
{
continue;
}
var current = new Phrase(tree.GetTokens(root, sentence, ignore, punctuation, options));
// If current node matches previous node but for punctuation omission, replace previous with current:
bool omitFalseDuplicatePhrases = options.HasFlag(AnalyzerOptions.OmitFalseDuplicatePhrases);
if (omitFalseDuplicatePhrases)
{
if (phrases.Count > 0)
{
Phrase previous = phrases.Last();
if (previous.EqualExceptPunctuationOmission(current))
{
phrases[phrases.Count - 1] = current;
continue;
}
}
}
if (current.Count == 0)
{
continue;
}
phrases.Add(current);
}
}
return(phrases);
}
