public static void OutputResults(SemgrexPattern pattern, SemanticGraph graph, params string[] ignored)
{
System.Console.Out.WriteLine("Matching pattern " + pattern + " to\n" + graph + " :" + (pattern.Matcher(graph).Matches() ? "matches" : "doesn't match"));
System.Console.Out.WriteLine();
pattern.PrettyPrint();
System.Console.Out.WriteLine();
SemgrexMatcher matcher = pattern.Matcher(graph);
while (matcher.Find())
{
System.Console.Out.WriteLine(" " + matcher.GetMatch());
ICollection <string> nodeNames = matcher.GetNodeNames();
if (nodeNames != null && nodeNames.Count > 0)
{
foreach (string name in nodeNames)
{
System.Console.Out.WriteLine(" " + name + ": " + matcher.GetNode(name));
}
}
ICollection <string> relNames = matcher.GetRelationNames();
if (relNames != null)
{
foreach (string name in relNames)
{
System.Console.Out.WriteLine(" " + name + ": " + matcher.GetRelnString(name));
}
}
}
}
public virtual void TestNamedNode()
{
SemanticGraph graph = MakeComplicatedGraph();
RunTest("{} >dobj ({} >expl {})", graph, "A");
SemgrexPattern pattern = SemgrexPattern.Compile("{} >dobj ({} >expl {}=foo)");
SemgrexMatcher matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(1, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("E", matcher.GetNode("foo").ToString());
NUnit.Framework.Assert.AreEqual("A", matcher.GetMatch().ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
pattern = SemgrexPattern.Compile("{} >dobj ({} >expl {}=foo) >mod {}");
matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(1, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("E", matcher.GetNode("foo").ToString());
NUnit.Framework.Assert.AreEqual("A", matcher.GetMatch().ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
pattern = SemgrexPattern.Compile("{} >dobj ({} >expl {}=foo) >mod ({} >mark {})");
matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(1, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("E", matcher.GetNode("foo").ToString());
NUnit.Framework.Assert.AreEqual("A", matcher.GetMatch().ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
pattern = SemgrexPattern.Compile("{} >dobj ({} >expl {}=foo) >mod ({} > {})");
matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(1, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("E", matcher.GetNode("foo").ToString());
NUnit.Framework.Assert.AreEqual("A", matcher.GetMatch().ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
pattern = SemgrexPattern.Compile("{} >dobj ({} >expl {}=foo) >mod ({} > {}=foo)");
matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(1, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("E", matcher.GetNode("foo").ToString());
NUnit.Framework.Assert.AreEqual("A", matcher.GetMatch().ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
pattern = SemgrexPattern.Compile("{} >dobj ({} >expl {}=foo) >mod ({}=foo > {})");
matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsFalse(matcher.Find());
}
public virtual void TestEqualsRelation()
{
SemanticGraph graph = SemanticGraph.ValueOf("[ate subj>Bill dobj>[muffins compound>blueberry]]");
SemgrexPattern pattern = SemgrexPattern.Compile("{} >> ({}=a == {}=b)");
SemgrexMatcher matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
// This split pattern should also work
pattern = SemgrexPattern.Compile("{} >> {}=a >> {}=b : {}=a == {}=b");
matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
}
public virtual void TestInitialConditions()
{
SemanticGraph graph = MakeComplicatedGraph();
SemgrexPattern pattern = SemgrexPattern.Compile("{}=a >> {}=b : {}=a >> {}=c");
IDictionary <string, IndexedWord> variables = new Dictionary <string, IndexedWord>();
variables["b"] = graph.GetNodeByIndex(5);
variables["c"] = graph.GetNodeByIndex(2);
SemgrexMatcher matcher = pattern.Matcher(graph, variables);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(3, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("A", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("E", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.AreEqual("B", matcher.GetNode("c").ToString());
NUnit.Framework.Assert.AreEqual("A", matcher.GetMatch().ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
}
public virtual void TestNotEquals()
{
SemanticGraph graph = SemanticGraph.ValueOf("[ate subj>Bill dobj>[muffins compound>blueberry]]");
SemgrexPattern pattern = SemgrexPattern.Compile("{} >> {}=a >> {}=b : {}=a !== {}=b");
SemgrexMatcher matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
// same as the first test, essentially, but with a more compact expression
pattern = SemgrexPattern.Compile("{} >> {}=a >> ({}=b !== {}=a)");
matcher = pattern.Matcher(graph);
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("Bill", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsTrue(matcher.Find());
NUnit.Framework.Assert.AreEqual(2, matcher.GetNodeNames().Count);
NUnit.Framework.Assert.AreEqual("ate", matcher.GetMatch().ToString());
NUnit.Framework.Assert.AreEqual("blueberry", matcher.GetNode("a").ToString());
NUnit.Framework.Assert.AreEqual("muffins", matcher.GetNode("b").ToString());
NUnit.Framework.Assert.IsFalse(matcher.Find());
}
/// <summary>Prints out all matches of a semgrex pattern on a file of dependencies.</summary>
/// <remarks>
/// Prints out all matches of a semgrex pattern on a file of dependencies.
/// <p>
/// Usage:<br />
/// java edu.stanford.nlp.semgraph.semgrex.SemgrexPattern [args]
/// <br />
/// See the help() function for a list of possible arguments to provide.
/// </remarks>
/// <exception cref="System.IO.IOException"/>
public static void Main(string[] args)
{
IDictionary <string, int> flagMap = Generics.NewHashMap();
flagMap[Pattern] = 1;
flagMap[TreeFile] = 1;
flagMap[Mode] = 1;
flagMap[Extras] = 1;
flagMap[ConlluFile] = 1;
flagMap[OutputFormatOption] = 1;
IDictionary <string, string[]> argsMap = StringUtils.ArgsToMap(args, flagMap);
// args = argsMap.get(null);
// TODO: allow patterns to be extracted from a file
if (!(argsMap.Contains(Pattern)) || argsMap[Pattern].Length == 0)
{
Help();
System.Environment.Exit(2);
}
Edu.Stanford.Nlp.Semgraph.Semgrex.SemgrexPattern semgrex = Edu.Stanford.Nlp.Semgraph.Semgrex.SemgrexPattern.Compile(argsMap[Pattern][0]);
string modeString = DefaultMode;
if (argsMap.Contains(Mode) && argsMap[Mode].Length > 0)
{
modeString = argsMap[Mode][0].ToUpper();
}
SemanticGraphFactory.Mode mode = SemanticGraphFactory.Mode.ValueOf(modeString);
string outputFormatString = DefaultOutputFormat;
if (argsMap.Contains(OutputFormatOption) && argsMap[OutputFormatOption].Length > 0)
{
outputFormatString = argsMap[OutputFormatOption][0].ToUpper();
}
SemgrexPattern.OutputFormat outputFormat = SemgrexPattern.OutputFormat.ValueOf(outputFormatString);
bool useExtras = true;
if (argsMap.Contains(Extras) && argsMap[Extras].Length > 0)
{
useExtras = bool.ValueOf(argsMap[Extras][0]);
}
IList <SemanticGraph> graphs = Generics.NewArrayList();
// TODO: allow other sources of graphs, such as dependency files
if (argsMap.Contains(TreeFile) && argsMap[TreeFile].Length > 0)
{
foreach (string treeFile in argsMap[TreeFile])
{
log.Info("Loading file " + treeFile);
MemoryTreebank treebank = new MemoryTreebank(new TreeNormalizer());
treebank.LoadPath(treeFile);
foreach (Tree tree in treebank)
{
// TODO: allow other languages... this defaults to English
SemanticGraph graph = SemanticGraphFactory.MakeFromTree(tree, mode, useExtras ? GrammaticalStructure.Extras.Maximal : GrammaticalStructure.Extras.None);
graphs.Add(graph);
}
}
}
if (argsMap.Contains(ConlluFile) && argsMap[ConlluFile].Length > 0)
{
CoNLLUDocumentReader reader = new CoNLLUDocumentReader();
foreach (string conlluFile in argsMap[ConlluFile])
{
log.Info("Loading file " + conlluFile);
IEnumerator <SemanticGraph> it = reader.GetIterator(IOUtils.ReaderFromString(conlluFile));
while (it.MoveNext())
{
SemanticGraph graph = it.Current;
graphs.Add(graph);
}
}
}
foreach (SemanticGraph graph_1 in graphs)
{
SemgrexMatcher matcher = semgrex.Matcher(graph_1);
if (!matcher.Find())
{
continue;
}
if (outputFormat == SemgrexPattern.OutputFormat.List)
{
log.Info("Matched graph:" + Runtime.LineSeparator() + graph_1.ToString(SemanticGraph.OutputFormat.List));
int i = 1;
bool found = true;
while (found)
{
log.Info("Match " + i + " at: " + matcher.GetMatch().ToString(CoreLabel.OutputFormat.ValueIndex));
IList <string> nodeNames = Generics.NewArrayList();
Sharpen.Collections.AddAll(nodeNames, matcher.GetNodeNames());
nodeNames.Sort();
foreach (string name in nodeNames)
{
log.Info(" " + name + ": " + matcher.GetNode(name).ToString(CoreLabel.OutputFormat.ValueIndex));
}
log.Info(" ");
found = matcher.Find();
}
}
else
{
if (outputFormat == SemgrexPattern.OutputFormat.Offset)
{
if (graph_1.VertexListSorted().IsEmpty())
{
continue;
}
System.Console.Out.Printf("+%d %s%n", graph_1.VertexListSorted()[0].Get(typeof(CoreAnnotations.LineNumberAnnotation)), argsMap[ConlluFile][0]);
}
}
}
}