public AddDep(string govNodeName, GrammaticalRelation relation, IndexedWord newNodePrototype)
{
this.newNodePrototype = newNodePrototype;
this.relation = relation;
this.govNodeName = govNodeName;
this.weight = 0;
}
public DepPattern(Token token, GrammaticalRelation relation)
: base(PatternFactory.PatternType.Dep)
{
this.relations = new List <Pair <Token, GrammaticalRelation> >();
relations.Add(new Pair <Token, GrammaticalRelation>(token, relation));
hashCode = this.ToString().GetHashCode();
}
/// <summary>Parse a JSON formatted tree into a SemanticGraph.</summary>
/// <param name="jsonString">
/// The JSON string tree to parse, e.g:
/// "[{\"\"dependent\"\": 7, \"\"dep\"\": \"\"root\"\", \"\"governorgloss\"\": \"\"root\"\", \"\"governor\"\": 0, \"\"dependentgloss\"\": \"\"sport\"\"}, {\"\"dependent\"\": 1, \"\"dep\"\": \"\"nsubj\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"chess\"\"}, {\"\"dependent\"\": 2, \"\"dep\"\": \"\"cop\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"is\"\"}, {\"\"dependent\"\": 3, \"\"dep\"\": \"\"neg\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"not\"\"}, {\"\"dependent\"\": 4, \"\"dep\"\": \"\"det\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"a\"\"}, {\"\"dependent\"\": 5, \"\"dep\"\": \"\"advmod\"\", \"\"governorgloss\"\": \"\"physical\"\", \"\"governor\"\": 6, \"\"dependentgloss\"\": \"\"predominantly\"\"}, {\"\"dependent\"\": 6, \"\"dep\"\": \"\"amod\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"physical\"\"}, {\"\"dependent\"\": 9, \"\"dep\"\": \"\"advmod\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"yet\"\"}, {\"\"dependent\"\": 10, \"\"dep\"\": \"\"nsubj\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"neither\"\"}, {\"\"dependent\"\": 11, \"\"dep\"\": \"\"cop\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"are\"\"}, {\"\"dependent\"\": 12, \"\"dep\"\": \"\"parataxis\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"shooting\"\"}, {\"\"dependent\"\": 13, \"\"dep\"\": \"\"cc\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"and\"\"}, {\"\"dependent\"\": 14, \"\"dep\"\": \"\"parataxis\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"curling\"\"}, {\"\"dependent\"\": 14, \"\"dep\"\": \"\"conj:and\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"curling\"\"}, {\"\"dependent\"\": 16, \"\"dep\"\": \"\"nsubjpass\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"which\"\"}, {\"\"dependent\"\": 18, \"\"dep\"\": \"\"case\"\", \"\"governorgloss\"\": \"\"fact\"\", \"\"governor\"\": 19, \"\"dependentgloss\"\": \"\"in\"\"}, {\"\"dependent\"\": 19, \"\"dep\"\": \"\"nmod:in\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"fact\"\"}, {\"\"dependent\"\": 21, \"\"dep\"\": \"\"aux\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"has\"\"}, {\"\"dependent\"\": 22, \"\"dep\"\": \"\"auxpass\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"been\"\"}, {\"\"dependent\"\": 23, \"\"dep\"\": \"\"dep\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"nicknamed\"\"}, {\"\"dependent\"\": 25, \"\"dep\"\": \"\"dobj\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"chess\"\"}, {\"\"dependent\"\": 26, \"\"dep\"\": \"\"case\"\", \"\"governorgloss\"\": \"\"ice\"\", \"\"governor\"\": 27, \"\"dependentgloss\"\": \"\"on\"\"}, {\"\"dependent\"\": 27, \"\"dep\"\": \"\"nmod:on\"\", \"\"governorgloss\"\": \"\"chess\"\", \"\"governor\"\": 25, \"\"dependentgloss\"\": \"\"ice\"\"}, {\"\"dependent\"\": 29, \"\"dep\"\": \"\"amod\"\", \"\"governorgloss\"\": \"\"chess\"\", \"\"governor\"\": 25, \"\"dependentgloss\"\": \"\"5\"\"}]");
/// </param>
/// <param name="tokens">The tokens of the sentence, to form the backing labels of the tree.</param>
/// <returns>A semantic graph of the sentence, according to the given tree.</returns>
public static SemanticGraph ParseJsonTree(string jsonString, IList <CoreLabel> tokens)
{
// Escape quoted string parts
IJsonReader json = Javax.Json.Json.CreateReader(new StringReader(jsonString));
SemanticGraph tree = new SemanticGraph();
IJsonArray array = json.ReadArray();
if (array == null || array.IsEmpty())
{
return(tree);
}
IndexedWord[] vertices = new IndexedWord[tokens.Count + 2];
// Add edges
for (int i = 0; i < array.Count; i++)
{
IJsonObject entry = array.GetJsonObject(i);
// Parse row
int dependentIndex = entry.GetInt("dependent");
if (vertices[dependentIndex] == null)
{
if (dependentIndex > tokens.Count)
{
// Bizarre mismatch in sizes; the malt parser seems to do this often
return(new SemanticGraph());
}
vertices[dependentIndex] = new IndexedWord(tokens[dependentIndex - 1]);
}
IndexedWord dependent = vertices[dependentIndex];
int governorIndex = entry.GetInt("governor");
if (governorIndex > tokens.Count)
{
// Bizarre mismatch in sizes; the malt parser seems to do this often
return(new SemanticGraph());
}
if (vertices[governorIndex] == null && governorIndex > 0)
{
vertices[governorIndex] = new IndexedWord(tokens[governorIndex - 1]);
}
IndexedWord governor = vertices[governorIndex];
string relation = entry.GetString("dep");
// Process row
if (governorIndex == 0)
{
tree.AddRoot(dependent);
}
else
{
tree.AddVertex(dependent);
if (!tree.ContainsVertex(governor))
{
tree.AddVertex(governor);
}
if (!"ref".Equals(relation))
{
tree.AddEdge(governor, dependent, GrammaticalRelation.ValueOf(Language.English, relation), double.NegativeInfinity, false);
}
}
}
return(tree);
}
/// <summary>Creates an EnglishGrammaticalRelation AddDep edit.</summary>
/// <param name="newNode">String representation of new dependent IndexedFeatureNode map.</param>
public static Edu.Stanford.Nlp.Semgraph.Semgrex.Ssurgeon.AddDep CreateEngAddDep(string govNodeName, string engRelation, string newNode)
{
GrammaticalRelation relation = EnglishGrammaticalRelations.ValueOf(engRelation);
// IndexedWord newNodeObj = new IndexedWord(CoreLabel.fromAbstractMapLabel(IndexedFeatureLabel.valueOf(newNode, MapFactory.HASH_MAP_FACTORY)));
IndexedWord newNodeObj = FromCheapString(newNode);
return(new Edu.Stanford.Nlp.Semgraph.Semgrex.Ssurgeon.AddDep(govNodeName, relation, newNodeObj));
}
/// <summary>Parse a CoNLL formatted tree into a SemanticGraph.</summary>
/// <param name="conll">The CoNLL tree to parse.</param>
/// <param name="tokens">The tokens of the sentence, to form the backing labels of the tree.</param>
/// <returns>A semantic graph of the sentence, according to the given tree.</returns>
public static SemanticGraph ParseTree(string conll, IList <CoreLabel> tokens)
{
SemanticGraph tree = new SemanticGraph();
if (conll == null || conll.IsEmpty())
{
return(tree);
}
string[] treeLines = newline.Split(conll);
IndexedWord[] vertices = new IndexedWord[tokens.Count + 2];
// Add edges
foreach (string line in treeLines)
{
// Parse row
string[] fields = tab.Split(line);
int dependentIndex = System.Convert.ToInt32(fields[0]);
if (vertices[dependentIndex] == null)
{
if (dependentIndex > tokens.Count)
{
// Bizarre mismatch in sizes; the malt parser seems to do this often
return(new SemanticGraph());
}
vertices[dependentIndex] = new IndexedWord(tokens[dependentIndex - 1]);
}
IndexedWord dependent = vertices[dependentIndex];
int governorIndex = System.Convert.ToInt32(fields[1]);
if (governorIndex > tokens.Count)
{
// Bizarre mismatch in sizes; the malt parser seems to do this often
return(new SemanticGraph());
}
if (vertices[governorIndex] == null && governorIndex > 0)
{
vertices[governorIndex] = new IndexedWord(tokens[governorIndex - 1]);
}
IndexedWord governor = vertices[governorIndex];
string relation = fields[2];
// Process row
if (governorIndex == 0)
{
tree.AddRoot(dependent);
}
else
{
tree.AddVertex(dependent);
if (!tree.ContainsVertex(governor))
{
tree.AddVertex(governor);
}
if (!"ref".Equals(relation))
{
tree.AddEdge(governor, dependent, GrammaticalRelation.ValueOf(Language.English, relation), double.NegativeInfinity, false);
}
}
}
return(tree);
}
public RemoveEdge(GrammaticalRelation relation, string govName, string depName)
{
// Name of the matched relation type
// Name of governor of this reln, in match
// Name of the dependent in this reln, in match
this.relation = relation;
this.govName = govName;
this.depName = depName;
}
/// <param name="source">The source IndexedWord for this edge</param>
/// <param name="target">The target IndexedWord for this edge</param>
/// <param name="relation">The relation between the two words represented by this edge</param>
/// <param name="weight">A score or weight to attach to the edge (not often used)</param>
/// <param name="isExtra">Whether or not the dependency this edge represents was "extra"</param>
public SemanticGraphEdge(IndexedWord source, IndexedWord target, GrammaticalRelation relation, double weight, bool isExtra)
{
// a hack for displaying SemanticGraph in JGraph. Should be redone better.
this.source = source;
this.target = target;
this.relation = relation;
this.weight = weight;
this.isExtra = isExtra;
}
/// <exception cref="System.Exception"/>
private static void DescendantsHelper(SemanticGraph g, IndexedWord curr, ICollection <IndexedWord> descendantSet, IList <string> allCutOffRels, IList <IndexedWord> doNotAddThese, IList <IndexedWord> seenNodes, bool ignoreCommonTags, IPredicate <CoreLabel
> acceptWord, CollectionValuedMap <int, string> feat)
{
if (seenNodes.Contains(curr))
{
return;
}
seenNodes.Add(curr);
if (descendantSet.Contains(curr) || (doNotAddThese != null && doNotAddThese.Contains(curr)) || !acceptWord.Test(curr.BackingLabel()))
{
return;
}
if (!ignoreCommonTags || !ignoreTags.Contains(curr.Tag().Trim()))
{
descendantSet.Add(curr);
}
foreach (IndexedWord child in g.GetChildren(curr))
{
bool dontuse = false;
if (doNotAddThese != null && doNotAddThese.Contains(child))
{
dontuse = true;
}
GrammaticalRelation rel = null;
if (dontuse == false)
{
rel = g.Reln(curr, child);
dontuse = CheckIfSatisfiesRelConstrains(g, curr, child, allCutOffRels, rel);
}
if (dontuse == false)
{
foreach (string cutOffTagRegex in cutoffTags)
{
if (child.Tag().Matches(cutOffTagRegex))
{
if (Debug >= 5)
{
System.Console.Out.WriteLine("ignored tag " + child + " because it satisfied " + cutOffTagRegex);
}
dontuse = true;
break;
}
}
}
if (dontuse == false)
{
if (!feat.Contains(curr.Index()))
{
feat[curr.Index()] = new List <string>();
}
GetPatternsFromDataMultiClass.GetFeatures(g, curr, false, feat[curr.Index()], rel);
//feat.add(curr.index(), "REL-" + rel.getShortName());
DescendantsHelper(g, child, descendantSet, allCutOffRels, doNotAddThese, seenNodes, ignoreCommonTags, acceptWord, feat);
}
}
}
public AddEdge(string govName, string depName, GrammaticalRelation relation)
{
// Name of governor of this reln, in match
// Name of the dependent in this reln, in match
// Type of relation to add between these edges
this.govName = govName;
this.depName = depName;
this.relation = relation;
this.weight = 0;
}
private static bool IfIgnoreRel(GrammaticalRelation rel)
{
if (ignoreRelsSet.Contains(rel))
{
return(true);
}
else
{
return(false);
}
}
public override void Evaluate(SemanticGraph sg, SemgrexMatcher sm)
{
string relation = sm.GetRelnString(edgeName);
IndexedWord govNode = GetNamedNode(govName, sm);
IndexedWord depNode = GetNamedNode(depName, sm);
SemanticGraphEdge edge = sg.GetEdge(govNode, depNode, GrammaticalRelation.ValueOf(relation));
if (edge != null)
{
sg.RemoveEdge(edge);
}
}
public static void SetUp(Properties props)
{
ArgumentParser.FillOptions(typeof(DepPatternFactory), props);
ArgumentParser.FillOptions(typeof(PatternFactory), props);
foreach (string s in ignoreRels.Split("[,;]"))
{
ignoreRelsSet.Add(GrammaticalRelation.ValueOf(s));
}
foreach (string s_1 in allowedTagsForTrigger.Split("[,;]"))
{
allowedTagPatternForTrigger.Add(Pattern.Compile(s_1));
}
}
internal static ICollection <DepPattern> GetContext(IndexedWord w, SemanticGraph graph, ICollection <CandidatePhrase> stopWords, DataInstance sent)
{
ICollection <DepPattern> patterns = new HashSet <DepPattern>();
IndexedWord node = w;
int depth = 1;
while (depth <= upDepth)
{
IndexedWord parent = graph.GetParent(node);
if (parent == null)
{
break;
}
GrammaticalRelation rel = graph.Reln(parent, node);
foreach (Pattern tagPattern in allowedTagPatternForTrigger)
{
if (tagPattern.Matcher(parent.Tag()).Matches())
{
if (!IfIgnoreRel(rel) && !stopWords.Contains(CandidatePhrase.CreateOrGet(parent.Word())) && parent.Word().Length > 1)
{
Pair <IndexedWord, GrammaticalRelation> pattern = new Pair <IndexedWord, GrammaticalRelation>(parent, rel);
DepPattern patterndep = PatternToDepPattern(pattern, sent);
if (depth <= upDepth)
{
patterns.Add(patterndep);
}
}
}
}
// if (depth <= maxDepth) {
// Counter<String> phrasesForPattern = phrasesForPatternForSent.get(patternStr);
// if (phrasesForPattern == null)
// phrasesForPattern = new ClassicCounter<String>();
// phrasesForPattern.incrementCount(phrase);
// phrasesForPatternForSent.put(patternStr, phrasesForPattern);
// }
// if (DEBUG >= 1)
// System.out.println("for phrase " + phrase + " pattern is " + patternStr);
node = parent;
depth++;
}
return(patterns);
}
private void ExtractNPorPRPFromDependency(ICoreMap s, IList <Mention> mentions, ICollection <IntPair> mentionSpanSet, ICollection <IntPair> namedEntitySpanSet)
{
IList <CoreLabel> sent = s.Get(typeof(CoreAnnotations.TokensAnnotation));
SemanticGraph basic = s.Get(typeof(SemanticGraphCoreAnnotations.BasicDependenciesAnnotation));
IList <IndexedWord> nounsOrPrp = basic.GetAllNodesByPartOfSpeechPattern("N.*|PRP.*|DT");
// DT is for "this, these, etc"
Tree tree = s.Get(typeof(TreeCoreAnnotations.TreeAnnotation));
foreach (IndexedWord w in nounsOrPrp)
{
SemanticGraphEdge edge = basic.GetEdge(basic.GetParent(w), w);
GrammaticalRelation rel = null;
string shortname = "root";
// if edge is null, it's root
if (edge != null)
{
rel = edge.GetRelation();
shortname = rel.GetShortName();
}
// TODO: what to remove? remove more?
if (shortname.Matches("det|compound"))
{
// // for debug ---------------
// Tree t = tree.getLeaves().get(w.index()-1);
// for(Tree p : tree.pathNodeToNode(t, tree)) {
// if(p.label().value().equals("NP")) {
// HeadFinder headFinder = new SemanticHeadFinder();
// Tree head = headFinder.determineHead(p);
// if(head == t.parent(tree)) {
// log.info();
// }
// break;
// }
// } // for debug -------------
continue;
}
else
{
ExtractMentionForHeadword(w, basic, s, mentions, mentionSpanSet, namedEntitySpanSet);
}
}
}
private static Pair <IndexedWord, GrammaticalRelation> GetGovAndReln(int govIdx, int copyCount, IndexedWord word, string relationName, IList <IndexedWord> sortedTokens)
{
IndexedWord gov;
GrammaticalRelation reln;
if (relationName.Equals("root"))
{
reln = GrammaticalRelation.Root;
}
else
{
reln = GrammaticalRelation.ValueOf(Language.UniversalEnglish, relationName);
}
if (govIdx == 0)
{
gov = new IndexedWord(word.DocID(), word.SentIndex(), 0);
gov.SetValue("ROOT");
}
else
{
gov = CoNLLUDocumentReader.SentenceProcessor.GetToken(sortedTokens, govIdx, copyCount);
}
return(Generics.NewPair(gov, reln));
}
public static Edu.Stanford.Nlp.Semgraph.Semgrex.Ssurgeon.AddEdge CreateEngAddEdge(string govName, string depName, string engRelnName, double weight)
{
GrammaticalRelation reln = EnglishGrammaticalRelations.ValueOf(engRelnName);
return(new Edu.Stanford.Nlp.Semgraph.Semgrex.Ssurgeon.AddEdge(govName, depName, reln, weight));
}
public AddEdge(string govName, string depName, GrammaticalRelation relation, double weight)
: this(govName, depName, relation)
{
this.weight = weight;
}
/// <summary>
/// Compares grammatical relations.
/// </summary>
/// <param name="gr1"></param>
/// <param name="gr2"></param>
/// <returns></returns>
public static bool IsGrammaticalRelation(this GrammaticalRelation gr1, GrammaticalRelation gr2)
{
return gr1.getLongName() == gr2.getLongName();
}
public virtual string PrintSemanticGraph(SemanticGraph sg, bool unescapeParenthesis)
{
bool isTree = SemanticGraphUtils.IsTree(sg);
StringBuilder sb = new StringBuilder();
/* Print comments. */
foreach (string comment in sg.GetComments())
{
sb.Append(comment).Append("\n");
}
foreach (IndexedWord token in sg.VertexListSorted())
{
/* Check for multiword tokens. */
if (token.ContainsKey(typeof(CoreAnnotations.CoNLLUTokenSpanAnnotation)))
{
IntPair tokenSpan = token.Get(typeof(CoreAnnotations.CoNLLUTokenSpanAnnotation));
if (tokenSpan.GetSource() == token.Index())
{
string range = string.Format("%d-%d", tokenSpan.GetSource(), tokenSpan.GetTarget());
sb.Append(string.Format("%s\t%s\t_\t_\t_\t_\t_\t_\t_\t_%n", range, token.OriginalText()));
}
}
/* Try to find main governor and additional dependencies. */
string govIdx = null;
GrammaticalRelation reln = null;
Dictionary <string, string> enhancedDependencies = new Dictionary <string, string>();
foreach (IndexedWord parent in sg.GetParents(token))
{
SemanticGraphEdge edge = sg.GetEdge(parent, token);
if (govIdx == null && !edge.IsExtra())
{
govIdx = parent.ToCopyIndex();
reln = edge.GetRelation();
}
enhancedDependencies[parent.ToCopyIndex()] = edge.GetRelation().ToString();
}
string additionalDepsString = isTree ? "_" : CoNLLUUtils.ToExtraDepsString(enhancedDependencies);
string word = token.Word();
string featuresString = CoNLLUUtils.ToFeatureString(token.Get(typeof(CoreAnnotations.CoNLLUFeats)));
string pos = token.GetString <CoreAnnotations.PartOfSpeechAnnotation>("_");
string upos = token.GetString <CoreAnnotations.CoarseTagAnnotation>("_");
string misc = token.GetString <CoreAnnotations.CoNLLUMisc>("_");
string lemma = token.GetString <CoreAnnotations.LemmaAnnotation>("_");
string relnName = reln == null ? "_" : reln.ToString();
/* Root. */
if (govIdx == null && sg.GetRoots().Contains(token))
{
govIdx = "0";
relnName = GrammaticalRelation.Root.ToString();
additionalDepsString = isTree ? "_" : "0:" + relnName;
}
else
{
if (govIdx == null)
{
govIdx = "_";
relnName = "_";
}
}
if (unescapeParenthesis)
{
word = word.ReplaceAll(LrbPattern, "(");
word = word.ReplaceAll(RrbPattern, ")");
lemma = lemma.ReplaceAll(LrbPattern, "(");
lemma = lemma.ReplaceAll(RrbPattern, ")");
}
sb.Append(string.Format("%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s%n", token.ToCopyIndex(), word, lemma, upos, pos, featuresString, govIdx, relnName, additionalDepsString, misc));
}
sb.Append("\n");
return(sb.ToString());
}
public virtual SemanticGraph Apply(string line)
{
if (line == null)
{
return(null);
}
IFunction <string, IndexedWord> func = new CoNLLUDocumentReader.WordProcessor();
ObjectBank <IndexedWord> words = ObjectBank.GetLineIterator(new StringReader(line), func);
IList <IndexedWord> wordList = new List <IndexedWord>(words);
IList <IndexedWord> sorted = new List <IndexedWord>(wordList.Count);
IList <string> comments = new LinkedList <string>();
/* Increase the line number in case there are comments before the actual sentence
* and add them to the list of comments. */
wordList.Stream().Filter(null).ForEach(null);
wordList.Stream().Filter(null).Sorted(byIndex.ThenComparing(byType)).ForEach(null);
IList <IndexedWord> sortedTokens = new List <IndexedWord>(wordList.Count);
sorted.Stream().Filter(null).Filter(null).ForEach(null);
sorted.Stream().Filter(null).Filter(null).ForEach(null);
/* Construct a semantic graph. */
IList <TypedDependency> deps = new List <TypedDependency>(sorted.Count);
IntPair tokenSpan = null;
string originalToken = null;
foreach (IndexedWord word in sorted)
{
lineNumberCounter++;
if (word.ContainsKey(typeof(CoreAnnotations.CoNLLUTokenSpanAnnotation)))
{
tokenSpan = word.Get(typeof(CoreAnnotations.CoNLLUTokenSpanAnnotation));
originalToken = word.Word();
}
else
{
/* Deal with multiword tokens. */
if (tokenSpan != null && tokenSpan.GetTarget() >= word.Index())
{
word.SetOriginalText(originalToken);
word.Set(typeof(CoreAnnotations.CoNLLUTokenSpanAnnotation), tokenSpan);
}
else
{
tokenSpan = null;
originalToken = null;
}
Dictionary <string, string> extraDeps = word.Get(typeof(CoreAnnotations.CoNLLUSecondaryDepsAnnotation));
if (extraDeps.IsEmpty())
{
int govIdx = word.Get(typeof(CoreAnnotations.CoNLLDepParentIndexAnnotation));
Pair <IndexedWord, GrammaticalRelation> govReln = GetGovAndReln(govIdx, 0, word, word.Get(typeof(CoreAnnotations.CoNLLDepTypeAnnotation)), sortedTokens);
IndexedWord gov = govReln.First();
GrammaticalRelation reln = govReln.Second();
TypedDependency dep = new TypedDependency(reln, gov, word);
word.Set(typeof(CoreAnnotations.LineNumberAnnotation), lineNumberCounter);
deps.Add(dep);
}
else
{
foreach (string extraGovIdxStr in extraDeps.Keys)
{
if (extraGovIdxStr.Contains("."))
{
string[] indexParts = extraGovIdxStr.Split("\\.");
int extraGovIdx = System.Convert.ToInt32(indexParts[0]);
int copyCount = System.Convert.ToInt32(indexParts[1]);
Pair <IndexedWord, GrammaticalRelation> govReln = GetGovAndReln(extraGovIdx, copyCount, word, extraDeps[extraGovIdxStr], sortedTokens);
IndexedWord gov = govReln.First();
GrammaticalRelation reln = govReln.Second();
TypedDependency dep = new TypedDependency(reln, gov, word);
dep.SetExtra();
deps.Add(dep);
}
else
{
int extraGovIdx = System.Convert.ToInt32(extraGovIdxStr);
int mainGovIdx = word.Get(typeof(CoreAnnotations.CoNLLDepParentIndexAnnotation)) != null?word.Get(typeof(CoreAnnotations.CoNLLDepParentIndexAnnotation)) : -1;
Pair <IndexedWord, GrammaticalRelation> govReln = GetGovAndReln(extraGovIdx, 0, word, extraDeps[extraGovIdxStr], sortedTokens);
IndexedWord gov = govReln.First();
GrammaticalRelation reln = govReln.Second();
TypedDependency dep = new TypedDependency(reln, gov, word);
if (extraGovIdx != mainGovIdx)
{
dep.SetExtra();
}
deps.Add(dep);
}
}
}
}
}
lineNumberCounter++;
SemanticGraph sg = new SemanticGraph(deps);
comments.ForEach(null);
return(sg);
}
public virtual void SetRelation(GrammaticalRelation relation)
{
this.relation = relation;
}
static ChineseGrammaticalRelations()
{
/*
* The "predicate" grammatical relation.
*/
// Fri Feb 20 15:42:54 2009 (pichuan)
// I'm surprise this relation has patterns.
// However it doesn't seem to match anything in CTB6.
/*
* public static final GrammaticalRelation PREDICATE =
* new GrammaticalRelation(Language.Chinese, "pred", "predicate",
* PredicateGRAnnotation.class, DEPENDENT, "IP",
* new String[]{
* " IP=target !> IP"
* });
* public static class PredicateGRAnnotation
* extends GrammaticalRelationAnnotation { }
*/
// Split the first rule to the second rule to avoid the duplication:
// ccomp(前来-12, 投资-13)
// conj(前来-12, 投资-13)
//
// (IP
// (VP
// (VP (VV 前来))
// (VP
// (VCD (VV 投资) (VV 办厂)))
// (CC 和)
// (VP (VV 洽谈)
// (NP (NN 生意))))))
// TODO: this following line has to be fixed.
// I think for now it just doesn't match anything.
//"VP|NP|ADJP|PP|ADVP|UCP < ( __=target $+ (PU < 、) )",
// Consider changing the rule ABOVE to these rules.
//"ADVP < ( /^AD/=target $+ ((PU < 、) $+ /^AD/))",
// This is for the 'conj's separated by commas.
// For now this creates too much duplicates with 'ccomp'.
// Need to look at more examples.
// Original version of this did not have the outer layer of
// the FRAG|INC|IP|VP. This caused a bug where the basic
// dependencies could have cycles.
// This following rule is too general and collides with 'ccomp'.
// Delete it for now.
// TODO: come up with a new rule. Does this exist in Chinese?
//"IP < (IP=target $+ ( VP !< VC))",
/*
* The "indirect object" grammatical relation.
*/
// deleted by pichuan: no real matches
/*
* public static final GrammaticalRelation INDIRECT_OBJECT =
* new GrammaticalRelation(Language.Chinese,
* "iobj", "indirect object",
* IndirectObjectGRAnnotation.class, OBJECT, "VP",
* new String[]{
* " CP !> VP < ( VV $+ ( NP|DP|QP|CLP=target . NP|DP ) )"
* });
* public static class IndirectObjectGRAnnotation
* extends GrammaticalRelationAnnotation { }
*/
// " VP|IP < ( VV|VC|VRD|VCD !$+ NP|QP|LCP ) > (IP < IP|VP|VRD|VCD=target) "
// "VP < (S=target < (VP !<, TO|VBG) !$-- NP)",
// pichuan: this is difficult to recognize in Chinese.
// remove the rules since it (always) collides with ccomp
// TODO: these rules seem to always collide with ccomp.
// Is this really desirable behavior?
//"VP !> (/^VP/ < /^VC$/ ) < (IP=target < (VP < P))",
//"ADJP < (IP=target <, (VP <, P))",
//"VP < (IP=target < (NP $+ NP|ADJP))",
//"VP < (/^VC/ $+ (VP=target < VC < NP))"
/*
* The "adjectival complement" grammatical relation.
* Example:
*/
// deleted by pichuan: no real matches
/*
* public static final GrammaticalRelation ADJECTIVAL_COMPLEMENT =
* new GrammaticalRelation(Language.Chinese,
* "acomp", "adjectival complement",
* AdjectivalComplementGRAnnotation.class, COMPLEMENT, "VP", tregexCompiler,
* new String[]{
* "VP < (ADJP=target !$-- NP)"
* });
* public static class AdjectivalComplementGRAnnotation
* extends GrammaticalRelationAnnotation { }
*/
/* This rule actually matches nothing.
* There's another tmod rule. This is removed for now.
* (pichuan) Sun Mar 8 18:22:40 2009
*/
/*
* public static final GrammaticalRelation TEMPORAL_MODIFIER =
* new GrammaticalRelation(Language.Chinese,
* "tmod", "temporal modifier",
* TemporalModifierGRAnnotation.class, MODIFIER, "VP|IP|ADJP", tregexCompiler,
* new String[]{
* " VC|VE ! >> VP|ADJP < NP=target < NT",
* "VC|VE !>>IP <( NP=target < NT $++ VP !< VC|VE )"
* });
* public static class TemporalModifierGRAnnotation
* extends GrammaticalRelationAnnotation { }
*/
// pichuan: previously "tclaus"
// TODO: we should figure out various ways to improve this pattern to
// improve both its precision and recall
//"DP < DT < QP=target"
/*
* The "possession modifier" grammatical relation.
*/
// Fri Feb 20 15:40:13 2009 (pichuan)
// I think this "poss" relation is just WRONG.
// DEC is a complementizer or a nominalizer,
// this rule probably originally want to capture "DEG".
// But it seems like it's covered by "assm" (associative marker).
/*
* public static final GrammaticalRelation POSSESSION_MODIFIER =
* new GrammaticalRelation(Language.Chinese,
* "poss", "possession modifier",
* PossessionModifierGRAnnotation.class, MODIFIER, "NP", tregexCompiler,
* new String[]{
* "NP < ( PN=target $+ DEC $+ NP )"
* });
* public static class PossessionModifierGRAnnotation
* extends GrammaticalRelationAnnotation { }
*/
/*
* The "possessive marker" grammatical relation.
*/
// Similar to the comments to "poss",
// I think this relation is wrong and will not appear.
/*
* public static final GrammaticalRelation POSSESSIVE_MODIFIER =
* new GrammaticalRelation(Language.Chinese, "possm", "possessive marker",
* PossessiveModifierGRAnnotation.class,
* MODIFIER, "NP", tregexCompiler,
* new String[]{
* "NP < ( PN $+ DEC=target ) "
* });
* public static class PossessiveModifierGRAnnotation
* extends GrammaticalRelationAnnotation { }
*/
// pichuan: previously "clmpd"
//ADJECTIVAL_COMPLEMENT,
//INDIRECT_OBJECT,
//POSSESSION_MODIFIER,
//POSSESSIVE_MODIFIER,
//PREDICATE,
// Cache frequently used views of the values list
for (int i = 0; i < rawValues.Length; i++)
{
GrammaticalRelation gr = rawValues[i];
synchronizedValues.Add(gr);
}
ValuesLock().Lock();
try
{
foreach (GrammaticalRelation gr in Edu.Stanford.Nlp.Trees.International.Pennchinese.ChineseGrammaticalRelations.Values())
{
shortNameToGRel[gr.GetShortName()] = gr;
}
}
finally
{
ValuesLock().Unlock();
}
}
public static GrammaticalRelation ValueOf(string s)
{
return(GrammaticalRelation.ValueOf(s, Values(), ValuesLock()));
}
public NotenizerRelation(GrammaticalRelation grammaticalRelation)
{
_longName = grammaticalRelation.getLongName();
_shortName = grammaticalRelation.getShortName();
_specific = grammaticalRelation.getSpecific();
}
public virtual SemanticGraph ConvertIntermediateGraph(IList <CoreLabel> sentence)
{
SemanticGraph graph = new SemanticGraph();
// First construct the actual nodes; keep them indexed by their index and copy count.
// Sentences such as "I went over the river and through the woods" have
// two copies for "went" in the collapsed dependencies.
TwoDimensionalMap <int, int, IndexedWord> nodeMap = TwoDimensionalMap.HashMap();
foreach (AnnotationSerializer.IntermediateNode @in in nodes)
{
CoreLabel token = sentence[@in.index - 1];
// index starts at 1!
IndexedWord word;
if (@in.copyAnnotation > 0)
{
// TODO: if we make a copy wrapper CoreLabel, use it here instead
word = new IndexedWord(new CoreLabel(token));
word.SetCopyCount(@in.copyAnnotation);
}
else
{
word = new IndexedWord(token);
}
// for backwards compatibility - new annotations should have
// these fields set, but annotations older than August 2014 might not
if (word.DocID() == null && @in.docId != null)
{
word.SetDocID(@in.docId);
}
if (word.SentIndex() < 0 && @in.sentIndex >= 0)
{
word.SetSentIndex(@in.sentIndex);
}
if (word.Index() < 0 && @in.index >= 0)
{
word.SetIndex(@in.index);
}
nodeMap.Put(word.Index(), word.CopyCount(), word);
graph.AddVertex(word);
if (@in.isRoot)
{
graph.AddRoot(word);
}
}
// add all edges to the actual graph
foreach (AnnotationSerializer.IntermediateEdge ie in edges)
{
IndexedWord source = nodeMap.Get(ie.source, ie.sourceCopy);
if (source == null)
{
throw new RuntimeIOException("Failed to find node " + ie.source + "-" + ie.sourceCopy);
}
IndexedWord target = nodeMap.Get(ie.target, ie.targetCopy);
if (target == null)
{
throw new RuntimeIOException("Failed to find node " + ie.target + "-" + ie.targetCopy);
}
// assert(target != null);
lock (Lock)
{
// this is not thread-safe: there are static fields in GrammaticalRelation
GrammaticalRelation rel = GrammaticalRelation.ValueOf(ie.dep);
graph.AddEdge(source, target, rel, 1.0, ie.isExtra);
}
}
// compute root nodes if they weren't stored in the graph
if (!graph.IsEmpty() && graph.GetRoots().Count == 0)
{
graph.ResetRoots();
}
return(graph);
}
internal static bool CheckIfSatisfiesRelConstrains(SemanticGraph g, IndexedWord curr, IndexedWord child, IList <string> allCutOffRels, GrammaticalRelation rel)
{
string relName = rel.GetShortName();
string relSpecificName = rel.ToString();
string relFullName = rel.GetLongName();
if (allCutOffRels != null)
{
foreach (string check in allCutOffRels)
{
if (relName.Matches(check) || (relSpecificName != null && relSpecificName.Matches(check)) || (relFullName != null && relFullName.Matches(check)))
{
return(true);
}
}
}
return(false);
}
/// <summary>Fix some bizarre peculiarities with certain trees.</summary>
/// <remarks>
/// Fix some bizarre peculiarities with certain trees.
/// So far, these include:
/// <ul>
/// <li>Sometimes there's a node from a word to itself. This seems wrong.</li>
/// </ul>
/// </remarks>
/// <param name="tree">The tree to clean (in place!).</param>
/// <returns>A list of extra edges, which are valid but were removed.</returns>
public static IList <SemanticGraphEdge> CleanTree(SemanticGraph tree)
{
// assert !isCyclic(tree);
// Clean nodes
IList <IndexedWord> toDelete = new List <IndexedWord>();
foreach (IndexedWord vertex in tree.VertexSet())
{
// Clean punctuation
if (vertex.Tag() == null)
{
continue;
}
char tag = vertex.BackingLabel().Tag()[0];
if (tag == '.' || tag == ',' || tag == '(' || tag == ')' || tag == ':')
{
if (!tree.OutgoingEdgeIterator(vertex).MoveNext())
{
// This should really never happen, but it does.
toDelete.Add(vertex);
}
}
}
toDelete.ForEach(null);
// Clean edges
IEnumerator <SemanticGraphEdge> iter = tree.EdgeIterable().GetEnumerator();
IList <Triple <IndexedWord, IndexedWord, SemanticGraphEdge> > toAdd = new List <Triple <IndexedWord, IndexedWord, SemanticGraphEdge> >();
toDelete.Clear();
while (iter.MoveNext())
{
SemanticGraphEdge edge = iter.Current;
if (edge.GetDependent().Index() == edge.GetGovernor().Index())
{
// Clean up copy-edges
if (edge.GetDependent().IsCopy(edge.GetGovernor()))
{
foreach (SemanticGraphEdge toCopy in tree.OutgoingEdgeIterable(edge.GetDependent()))
{
toAdd.Add(Triple.MakeTriple(edge.GetGovernor(), toCopy.GetDependent(), toCopy));
}
toDelete.Add(edge.GetDependent());
}
if (edge.GetGovernor().IsCopy(edge.GetDependent()))
{
foreach (SemanticGraphEdge toCopy in tree.OutgoingEdgeIterable(edge.GetGovernor()))
{
toAdd.Add(Triple.MakeTriple(edge.GetDependent(), toCopy.GetDependent(), toCopy));
}
toDelete.Add(edge.GetGovernor());
}
// Clean self-edges
iter.Remove();
}
else
{
if (edge.GetRelation().ToString().Equals("punct"))
{
// Clean punctuation (again)
if (!tree.OutgoingEdgeIterator(edge.GetDependent()).MoveNext())
{
// This should really never happen, but it does.
iter.Remove();
}
}
}
}
// (add edges we wanted to add)
toDelete.ForEach(null);
foreach (Triple <IndexedWord, IndexedWord, SemanticGraphEdge> edge_1 in toAdd)
{
tree.AddEdge(edge_1.first, edge_1.second, edge_1.third.GetRelation(), edge_1.third.GetWeight(), edge_1.third.IsExtra());
}
// Handle extra edges.
// Two cases:
// (1) the extra edge is a subj/obj edge and the main edge is a conj:.*
// in this case, keep the extra
// (2) otherwise, delete the extra
IList <SemanticGraphEdge> extraEdges = new List <SemanticGraphEdge>();
foreach (SemanticGraphEdge edge_2 in tree.EdgeIterable())
{
if (edge_2.IsExtra())
{
IList <SemanticGraphEdge> incomingEdges = tree.IncomingEdgeList(edge_2.GetDependent());
SemanticGraphEdge toKeep = null;
foreach (SemanticGraphEdge candidate in incomingEdges)
{
if (toKeep == null)
{
toKeep = candidate;
}
else
{
if (toKeep.GetRelation().ToString().StartsWith("conj") && candidate.GetRelation().ToString().Matches(".subj.*|.obj.*"))
{
toKeep = candidate;
}
else
{
if (!candidate.IsExtra() && !(candidate.GetRelation().ToString().StartsWith("conj") && toKeep.GetRelation().ToString().Matches(".subj.*|.obj.*")))
{
toKeep = candidate;
}
}
}
}
foreach (SemanticGraphEdge candidate_1 in incomingEdges)
{
if (candidate_1 != toKeep)
{
extraEdges.Add(candidate_1);
}
}
}
}
extraEdges.ForEach(null);
// Add apposition edges (simple coref)
foreach (SemanticGraphEdge extraEdge in new List <SemanticGraphEdge>(extraEdges))
{
// note[gabor] prevent concurrent modification exception
foreach (SemanticGraphEdge candidateAppos in tree.IncomingEdgeIterable(extraEdge.GetDependent()))
{
if (candidateAppos.GetRelation().ToString().Equals("appos"))
{
extraEdges.Add(new SemanticGraphEdge(extraEdge.GetGovernor(), candidateAppos.GetGovernor(), extraEdge.GetRelation(), extraEdge.GetWeight(), extraEdge.IsExtra()));
}
}
foreach (SemanticGraphEdge candidateAppos_1 in tree.OutgoingEdgeIterable(extraEdge.GetDependent()))
{
if (candidateAppos_1.GetRelation().ToString().Equals("appos"))
{
extraEdges.Add(new SemanticGraphEdge(extraEdge.GetGovernor(), candidateAppos_1.GetDependent(), extraEdge.GetRelation(), extraEdge.GetWeight(), extraEdge.IsExtra()));
}
}
}
// Brute force ensure tree
// Remove incoming edges from roots
IList <SemanticGraphEdge> rootIncomingEdges = new List <SemanticGraphEdge>();
foreach (IndexedWord root in tree.GetRoots())
{
foreach (SemanticGraphEdge incomingEdge in tree.IncomingEdgeIterable(root))
{
rootIncomingEdges.Add(incomingEdge);
}
}
rootIncomingEdges.ForEach(null);
// Loop until it becomes a tree.
bool changed = true;
while (changed)
{
// I just want trees to be trees; is that so much to ask!?
changed = false;
IList <IndexedWord> danglingNodes = new List <IndexedWord>();
IList <SemanticGraphEdge> invalidEdges = new List <SemanticGraphEdge>();
foreach (IndexedWord vertex_1 in tree.VertexSet())
{
// Collect statistics
IEnumerator <SemanticGraphEdge> incomingIter = tree.IncomingEdgeIterator(vertex_1);
bool hasIncoming = incomingIter.MoveNext();
bool hasMultipleIncoming = false;
if (hasIncoming)
{
incomingIter.Current;
hasMultipleIncoming = incomingIter.MoveNext();
}
// Register actions
if (!hasIncoming && !tree.GetRoots().Contains(vertex_1))
{
danglingNodes.Add(vertex_1);
}
else
{
if (hasMultipleIncoming)
{
foreach (SemanticGraphEdge edge in new IterableIterator <SemanticGraphEdge>(incomingIter))
{
invalidEdges.Add(edge_2);
}
}
}
}
// Perform actions
foreach (IndexedWord vertex_2 in danglingNodes)
{
tree.RemoveVertex(vertex_2);
changed = true;
}
foreach (SemanticGraphEdge edge_3 in invalidEdges)
{
tree.RemoveEdge(edge_3);
changed = true;
}
}
// Edge case: remove duplicate dobj to "that."
// This is a common parse error.
foreach (IndexedWord vertex_3 in tree.VertexSet())
{
SemanticGraphEdge thatEdge = null;
int dobjCount = 0;
foreach (SemanticGraphEdge edge in tree.OutgoingEdgeIterable(vertex_3))
{
if (Sharpen.Runtime.EqualsIgnoreCase("that", edge_2.GetDependent().Word()))
{
thatEdge = edge_2;
}
if ("dobj".Equals(edge_2.GetRelation().ToString()))
{
dobjCount += 1;
}
}
if (dobjCount > 1 && thatEdge != null)
{
// Case: there are two dobj edges, one of which goes to the word "that"
// Action: rewrite the dobj edge to "that" to be a "mark" edge.
tree.RemoveEdge(thatEdge);
tree.AddEdge(thatEdge.GetGovernor(), thatEdge.GetDependent(), GrammaticalRelation.ValueOf(thatEdge.GetRelation().GetLanguage(), "mark"), thatEdge.GetWeight(), thatEdge.IsExtra());
}
}
// Return
System.Diagnostics.Debug.Assert(IsTree(tree));
return(extraEdges);
}
/// <summary>Returns all of the entailed shortened clauses (as per natural logic) from the given clause.</summary>
/// <remarks>
/// Returns all of the entailed shortened clauses (as per natural logic) from the given clause.
/// This runs the forward entailment component of the OpenIE system only.
/// It is usually chained together with the clause splitting component:
/// <see cref="ClausesInSentence(Edu.Stanford.Nlp.Util.ICoreMap)"/>
/// .
/// </remarks>
/// <param name="clause">The premise clause, as a sentence fragment in itself.</param>
/// <returns>A list of entailed clauses.</returns>
public virtual IList <SentenceFragment> EntailmentsFromClause(SentenceFragment clause)
{
if (clause.parseTree.IsEmpty())
{
return(Java.Util.Collections.EmptyList());
}
else
{
// Get the forward entailments
IList <SentenceFragment> list = new List <SentenceFragment>();
if (entailmentsPerSentence > 0)
{
Sharpen.Collections.AddAll(list, forwardEntailer.Apply(clause.parseTree, true).Search().Stream().Map(null).Collect(Collectors.ToList()));
}
list.Add(clause);
// A special case for adjective entailments
IList <SentenceFragment> adjFragments = new List <SentenceFragment>();
SemgrexMatcher matcher = adjectivePattern.Matcher(clause.parseTree);
while (matcher.Find())
{
// (get nodes)
IndexedWord subj = matcher.GetNode("subj");
IndexedWord be = matcher.GetNode("be");
IndexedWord adj = matcher.GetNode("adj");
IndexedWord obj = matcher.GetNode("obj");
IndexedWord pobj = matcher.GetNode("pobj");
string prep = matcher.GetRelnString("prep");
// (if the adjective, or any earlier adjective, is privative, then all bets are off)
foreach (SemanticGraphEdge edge in clause.parseTree.OutgoingEdgeIterable(obj))
{
if ("amod".Equals(edge.GetRelation().ToString()) && edge.GetDependent().Index() <= adj.Index() && Edu.Stanford.Nlp.Naturalli.Util.PrivativeAdjectives.Contains(edge.GetDependent().Word().ToLower()))
{
goto OUTER_continue;
}
}
// (create the core tree)
SemanticGraph tree = new SemanticGraph();
tree.AddRoot(adj);
tree.AddVertex(subj);
tree.AddVertex(be);
tree.AddEdge(adj, be, GrammaticalRelation.ValueOf(Language.English, "cop"), double.NegativeInfinity, false);
tree.AddEdge(adj, subj, GrammaticalRelation.ValueOf(Language.English, "nsubj"), double.NegativeInfinity, false);
// (add pp attachment, if it existed)
if (pobj != null)
{
System.Diagnostics.Debug.Assert(prep != null);
tree.AddEdge(adj, pobj, GrammaticalRelation.ValueOf(Language.English, prep), double.NegativeInfinity, false);
}
// (check for monotonicity)
if (adj.Get(typeof(NaturalLogicAnnotations.PolarityAnnotation)).IsUpwards() && be.Get(typeof(NaturalLogicAnnotations.PolarityAnnotation)).IsUpwards())
{
// (add tree)
adjFragments.Add(new SentenceFragment(tree, clause.assumedTruth, false));
}
OUTER_continue :;
}
OUTER_break :;
Sharpen.Collections.AddAll(list, adjFragments);
return(list);
}
}
/// <summary>Given a string entry, converts it into a SsurgeonEdit object.</summary>
public static SsurgeonEdit ParseEditLine(string editLine)
{
// Extract the operation name first
string[] tuples1 = editLine.Split("\\s+", 2);
if (tuples1.Length < 2)
{
throw new ArgumentException("Error in SsurgeonEdit.parseEditLine: invalid number of arguments");
}
string command = tuples1[0];
string[] argsArray = ParseArgs(tuples1[1]);
Ssurgeon.SsurgeonArgs argsBox = new Ssurgeon.SsurgeonArgs();
for (int argIndex = 0; argIndex < argsArray.Length; ++argIndex)
{
switch (argsArray[argIndex])
{
case GovNodenameArg:
{
argsBox.govNodeName = argsArray[argIndex + 1];
argIndex += 2;
break;
}
case DepNodenameArg:
{
argsBox.dep = argsArray[argIndex + 1];
argIndex += 2;
break;
}
case EdgeNameArg:
{
argsBox.edge = argsArray[argIndex + 1];
argIndex += 2;
break;
}
case RelnArg:
{
argsBox.reln = argsArray[argIndex + 1];
argIndex += 2;
break;
}
case NodenameArg:
{
argsBox.node = argsArray[argIndex + 1];
argIndex += 2;
break;
}
case NodeProtoArg:
{
argsBox.nodeString = argsArray[argIndex + 1];
argIndex += 2;
break;
}
case WeightArg:
{
argsBox.weight = double.ValueOf(argsArray[argIndex + 1]);
argIndex += 2;
break;
}
case NameArg:
{
argsBox.name = argsArray[argIndex + 1];
argIndex += 2;
break;
}
default:
{
throw new ArgumentException("Parsing Ssurgeon args: unknown flag " + argsArray[argIndex]);
}
}
}
// Parse the arguments based upon the type of command to execute.
// TODO: this logic really should be moved into the individual classes. The string-->class
// mappings should also be stored in more appropriate data structure.
SsurgeonEdit retEdit;
if (Sharpen.Runtime.EqualsIgnoreCase(command, AddDep.Label))
{
retEdit = AddDep.CreateEngAddDep(argsBox.govNodeName, argsBox.reln, argsBox.nodeString);
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(command, AddNode.Label))
{
retEdit = AddNode.CreateAddNode(argsBox.nodeString, argsBox.name);
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(command, AddEdge.Label))
{
retEdit = AddEdge.CreateEngAddEdge(argsBox.govNodeName, argsBox.dep, argsBox.reln);
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(command, DeleteGraphFromNode.Label))
{
retEdit = new DeleteGraphFromNode(argsBox.node);
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(command, RemoveEdge.Label))
{
retEdit = new RemoveEdge(GrammaticalRelation.ValueOf(argsBox.reln), argsBox.govNodeName, argsBox.dep);
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(command, RemoveNamedEdge.Label))
{
retEdit = new RemoveNamedEdge(argsBox.edge, argsBox.govNodeName, argsBox.dep);
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(command, SetRoots.Label))
{
string[] names = tuples1[1].Split("\\s+");
IList <string> newRoots = Arrays.AsList(names);
retEdit = new SetRoots(newRoots);
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(command, KillNonRootedNodes.Label))
{
retEdit = new KillNonRootedNodes();
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(command, KillAllIncomingEdges.Label))
{
retEdit = new KillAllIncomingEdges(argsBox.node);
}
else
{
throw new ArgumentException("Error in SsurgeonEdit.parseEditLine: command '" + command + "' is not supported");
}
}
}
}
}
}
}
}
}
return(retEdit);
}
public AddDep(string govNodeName, GrammaticalRelation relation, IndexedWord newNodePrototype, double weight)
: this(govNodeName, relation, newNodePrototype)
{
this.weight = weight;
}
/// <summary>
/// The "prep" grammatical relation. Used to collapse prepositions.
/// They will be turned into prep_word, where "word" is a preposition
/// </summary>
/// <param name="prepositionString">The preposition to make a GrammaticalRelation out of</param>
/// <returns>A grammatical relation for this preposition</returns>
public static GrammaticalRelation GetPrep(string prepositionString)
{
GrammaticalRelation result = Preps[prepositionString];
if (result == null)
{
lock (Preps)
{
result = Preps[prepositionString];
if (result == null)
{
result = new GrammaticalRelation(GrammaticalRelation.Language.English, "prep", "prep_collapsed",
PrepositionalModifier, prepositionString);
Preps.TryAdd(prepositionString, result);
//threadSafeAddRelation(result);
}
}
}
return result;
}
/// <summary>
/// The "prepc" grammatical relation. Used to collapse preposition complements.
/// They will be turned into prep_word, where "word" is a preposition
/// </summary>
/// <param name="prepositionString">The preposition to make a GrammaticalRelation out of</param>
/// <returns>A grammatical relation for this preposition</returns>
public static GrammaticalRelation GetPrepC(string prepositionString)
{
GrammaticalRelation result = PrepsC[prepositionString];
if (result == null)
{
lock (PrepsC)
{
result = PrepsC[prepositionString];
if (result == null)
{
result = new GrammaticalRelation(GrammaticalRelation.Language.English, "prepc",
"prepc_collapsed", GrammaticalRelation.Dependent, prepositionString);
PrepsC.TryAdd(prepositionString, result);
//threadSafeAddRelation(result);
}
}
}
return result;
}
// collapseMultiwordPreps(list);
private static void CollapsePrepAndPoss(ICollection <TypedDependency> list)
{
ICollection <TypedDependency> newTypedDeps = new List <TypedDependency>();
// Construct a map from words to the set of typed
// dependencies in which the word appears as governor.
IDictionary <IndexedWord, ICollection <TypedDependency> > map = Generics.NewHashMap();
foreach (TypedDependency typedDep in list)
{
if (!map.Contains(typedDep.Gov()))
{
map[typedDep.Gov()] = Generics.NewHashSet <TypedDependency>();
}
map[typedDep.Gov()].Add(typedDep);
}
//log.info("here's the map: " + map);
foreach (TypedDependency td1 in list)
{
if (td1.Reln() != GrammaticalRelation.Kill)
{
IndexedWord td1Dep = td1.Dep();
string td1DepPOS = td1Dep.Tag();
// find all other typedDeps having our dep as gov
ICollection <TypedDependency> possibles = map[td1Dep];
if (possibles != null)
{
// look for the "second half"
foreach (TypedDependency td2 in possibles)
{
// TreeGraphNode td2Dep = td2.dep();
// String td2DepPOS = td2Dep.parent().value();
if (td1.Reln() == GrammaticalRelation.Dependent && td2.Reln() == GrammaticalRelation.Dependent && td1DepPOS.Equals("P"))
{
GrammaticalRelation td3reln = UniversalChineseGrammaticalRelations.ValueOf(td1Dep.Value());
if (td3reln == null)
{
td3reln = GrammaticalRelation.ValueOf(Language.UniversalChinese, td1Dep.Value());
}
TypedDependency td3 = new TypedDependency(td3reln, td1.Gov(), td2.Dep());
//log.info("adding: " + td3);
newTypedDeps.Add(td3);
td1.SetReln(GrammaticalRelation.Kill);
// remember these are "used up"
td2.SetReln(GrammaticalRelation.Kill);
}
}
// remember these are "used up"
// Now we need to see if there any TDs that will be "orphaned"
// by this collapse. Example: if we have:
// dep(drew, on)
// dep(on, book)
// dep(on, right)
// the first two will be collapsed to on(drew, book), but then
// the third one will be orphaned, since its governor no
// longer appears. So, change its governor to 'drew'.
if (td1.Reln().Equals(GrammaticalRelation.Kill))
{
foreach (TypedDependency td2_1 in possibles)
{
if (!td2_1.Reln().Equals(GrammaticalRelation.Kill))
{
//log.info("td1 & td2: " + td1 + " & " + td2);
td2_1.SetGov(td1.Gov());
}
}
}
}
}
}
// now copy remaining unkilled TDs from here to new
foreach (TypedDependency td in list)
{
if (!td.Reln().Equals(GrammaticalRelation.Kill))
{
newTypedDeps.Add(td);
}
}
list.Clear();
// forget all (esp. killed) TDs
Sharpen.Collections.AddAll(list, newTypedDeps);
}
/// <summary>
/// The "conj" grammatical relation. Used to collapse conjunct relations.
/// They will be turned into conj_word, where "word" is a conjunction.
/// </summary>
/// <param name="conjunctionString">The conjunction to make a GrammaticalRelation out of</param>
/// <returns>A grammatical relation for this conjunction</returns>
public static GrammaticalRelation GetConj(string conjunctionString)
{
GrammaticalRelation result = Conjs[conjunctionString];
if (result == null)
{
lock (Conjs)
{
result = Conjs[conjunctionString];
if (result == null)
{
result = new GrammaticalRelation(GrammaticalRelation.Language.English, "conj", "conj_collapsed",
Conjunct, conjunctionString);
Conjs.TryAdd(conjunctionString, result);
//threadSafeAddRelation(result);
}
}
}
return result;
}
