edu.stanford.nlp.trees

Class UniversalEnglishGrammaticalStructure

java.lang.Object

edu.stanford.nlp.trees.GrammaticalStructure

edu.stanford.nlp.trees.UniversalEnglishGrammaticalStructure

All Implemented Interfaces:

java.io.Serializable

public class UniversalEnglishGrammaticalStructure
extends GrammaticalStructure

A GrammaticalStructure for Universal Dependencies English.
For feeding Stanford parser trees into this class, the Stanford parser should be run with the "-retainNPTmpSubcategories" option for best results!

Author:

Bill MacCartney, Marie-Catherine de Marneffe, Christopher Manning, Daniel Cer (CoNLLX format and alternative user selected dependency printer/reader interface), John Bauer, Sebastian Schuster

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	UniversalEnglishGrammaticalStructure.FromDependenciesFactory

Nested classes/interfaces inherited from class edu.stanford.nlp.trees.GrammaticalStructure

GrammaticalStructure.Extras

Field Summary

Fields

Modifier and Type	Field and Description
static EnhancementOptions	COLLAPSED_OPTIONS Deprecated.
static EnhancementOptions	ENHANCED_OPTIONS Options for "Enhanced" representation: Process multi-word prepositions: No Add prepositions to relation labels: Yes Add prepositions only to nmod relations: No Add coordinating conjunctions to relation labels: Yes Propagate dependents: Yes Add "referent" relations: Yes Add copy nodes for conjoined Ps and PPs: No Turn quantificational modifiers into flat MWEs: No Add relations between controlling subject and controlled verbs: Yes
static EnhancementOptions	ENHANCED_PLUS_PLUS_OPTIONS Options for "Enhanced++" representation: Process multi-word prepositions: Yes Add prepositions to relation labels: Yes Add prepositions only to nmod relations: No Add coordinating conjunctions to relation labels: Yes Propagate dependents: Yes Add "referent" relations: Yes Add copy nodes for conjoined Ps and PPs: Yes Turn quantificational modifiers into flat MWEs: Yes Add relations between controlling subject and controlled verbs: Yes

Fields inherited from class edu.stanford.nlp.trees.GrammaticalStructure

allTypedDependencies, CoNLLX_FieldCount, CoNLLX_GovField, CoNLLX_POSField, CoNLLX_RelnField, CoNLLX_WordField, puncFilter, tagFilter, typedDependencies

Constructor Summary

Constructors

Constructor and Description
UniversalEnglishGrammaticalStructure(java.util.List<TypedDependency> projectiveDependencies, TreeGraphNode root) Used for postprocessing CoNLL X dependencies
UniversalEnglishGrammaticalStructure(Tree t) Construct a new EnglishGrammaticalStructure from an existing parse tree.
UniversalEnglishGrammaticalStructure(Tree t, java.util.function.Predicate<java.lang.String> tagFilter) This gets used by GrammaticalStructureFactory (by reflection).
UniversalEnglishGrammaticalStructure(Tree t, java.util.function.Predicate<java.lang.String> tagFilter, HeadFinder hf) Construct a new GrammaticalStructure from an existing parse tree.

Method Summary

Modifier and Type	Method and Description
protected void	addEnhancements(java.util.List<TypedDependency> list, EnhancementOptions options) Destructively applies different enhancements to the dependency graph.
static void	addEnhancements(SemanticGraph sg, EnhancementOptions options) Manipulates the given SemanticGraph to add enhancements.
static UniversalEnglishGrammaticalStructure	buildCoNLLXGrammaticalStructure(java.util.List<java.util.List<java.lang.String>> tokenFields)
protected void	collapseDependencies(java.util.List<TypedDependency> list, boolean CCprocess, GrammaticalStructure.Extras includeExtras) Destructively modifies this Collection<TypedDependency> by collapsing several types of transitive pairs of dependencies or by adding additional information from the dependents to the relation of the governor.
protected void	collapseDependenciesTree(java.util.List<TypedDependency> list) Destructively modify the Collection&lt;TypedDependency&gt; to collapse language-dependent transitive dependencies but keeping a tree structure.
protected void	correctDependencies(java.util.List<TypedDependency> list) Destructively modify the TypedDependencyGraph to correct language-dependent dependencies.
protected static void	correctDependencies(SemanticGraph sg)
protected java.util.function.Predicate<TypedDependency>	extraTreeDepFilter() Returns a Filter which checks dependencies for usefulness as extra tree-based dependencies.
protected void	getExtras(java.util.List<TypedDependency> list) Get extra dependencies that do not depend on the tree structure, but rather only depend on the existing dependency structure.
protected void	getTreeDeps(java.util.List<TypedDependency> deps, DirectedMultiGraph<TreeGraphNode,GrammaticalRelation> completeGraph, java.util.function.Predicate<TypedDependency> puncTypedDepFilter, java.util.function.Predicate<TypedDependency> extraTreeDepFilter) Look through the tree t and adds to the List basicDep additional dependencies which aren't in the List but which satisfy the filter puncTypedDepFilter.
protected void	postProcessDependencies(java.util.List<TypedDependency> list) Post process the dependencies in whatever way this language requires.
protected static void	postProcessDependencies(SemanticGraph sg)
static java.util.List<GrammaticalStructure>	readCoNLLXGrammaticalStructureCollection(java.lang.String fileName)

Methods inherited from class edu.stanford.nlp.trees.GrammaticalStructure

allTypedDependencies, buildCoNLLXGrammaticalStructure, fromStringReps, getGrammaticalRelation, getGrammaticalRelation, getRoots, isConnected, main, readCoNLLXGrammaticalStructureCollection, root, toString, typedDependencies, typedDependencies, typedDependencies, typedDependenciesCCprocessed, typedDependenciesCCprocessed, typedDependenciesCCprocessed, typedDependenciesCollapsed, typedDependenciesCollapsed, typedDependenciesCollapsed, typedDependenciesCollapsedTree, typedDependenciesEnhanced, typedDependenciesEnhancedPlusPlus

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

ENHANCED_OPTIONS

public static final EnhancementOptions ENHANCED_OPTIONS

Options for "Enhanced" representation:

• Process multi-word prepositions: No
• Add prepositions to relation labels: Yes
• Add prepositions only to nmod relations: No
• Add coordinating conjunctions to relation labels: Yes
• Propagate dependents: Yes
• Add "referent" relations: Yes
• Add copy nodes for conjoined Ps and PPs: No
• Turn quantificational modifiers into flat MWEs: No
• Add relations between controlling subject and controlled verbs: Yes

ENHANCED_PLUS_PLUS_OPTIONS

public static final EnhancementOptions ENHANCED_PLUS_PLUS_OPTIONS

Options for "Enhanced++" representation:

• Process multi-word prepositions: Yes
• Add prepositions to relation labels: Yes
• Add prepositions only to nmod relations: No
• Add coordinating conjunctions to relation labels: Yes
• Propagate dependents: Yes
• Add "referent" relations: Yes
• Add copy nodes for conjoined Ps and PPs: Yes
• Turn quantificational modifiers into flat MWEs: Yes
• Add relations between controlling subject and controlled verbs: Yes

COLLAPSED_OPTIONS

@Deprecated
public static final EnhancementOptions COLLAPSED_OPTIONS

Deprecated.

Options for "Collapsed" representation. This representation is similar to the "collapsed" SD representation without any "Extra" relations.

• Process multi-word prepositions: Yes
• Add prepositions to relation labels: Yes
• Add prepositions only to nmod relations: Yes
• Add coordinating conjunctions to relation labels: Yes
• Propagate dependents: No
• Add "referent" relations: No
• Add copy nodes for conjoined Ps and PPs: Yes
• Turn quantificational modifiers into flat MWEs: No
• Add relations between controlling subject and controlled verbs: No

Constructor Detail

UniversalEnglishGrammaticalStructure

public UniversalEnglishGrammaticalStructure(Tree t)

Construct a new EnglishGrammaticalStructure from an existing parse tree. The new GrammaticalStructure has the same tree structure and label values as the given tree (but no shared storage). As part of construction, the parse tree is analyzed using definitions from {@code GrammaticalRelation} to populate the new GrammaticalStructure with as many labeled grammatical relations as it can.

Parameters:

t - Parse tree to make grammatical structure from

UniversalEnglishGrammaticalStructure

public UniversalEnglishGrammaticalStructure(Tree t,
                                            java.util.function.Predicate<java.lang.String> tagFilter)

This gets used by GrammaticalStructureFactory (by reflection). DON'T DELETE.

Parameters:

t - Parse tree to make grammatical structure from

tagFilter - Filter to remove punctuation dependencies

UniversalEnglishGrammaticalStructure

public UniversalEnglishGrammaticalStructure(Tree t,
                                            java.util.function.Predicate<java.lang.String> tagFilter,
                                            HeadFinder hf)

Construct a new GrammaticalStructure from an existing parse tree. The new GrammaticalStructure has the same tree structure and label values as the given tree (but no shared storage). As part of construction, the parse tree is analyzed using definitions from {@code GrammaticalRelation} to populate the new GrammaticalStructure with as many labeled grammatical relations as it can. This gets used by GrammaticalStructureFactory (by reflection). DON'T DELETE.

Parameters:

t - Parse tree to make grammatical structure from

tagFilter - Filter for punctuation tags

hf - HeadFinder to use when building it

UniversalEnglishGrammaticalStructure

public UniversalEnglishGrammaticalStructure(java.util.List<TypedDependency> projectiveDependencies,
                                            TreeGraphNode root)

Used for postprocessing CoNLL X dependencies

Method Detail

extraTreeDepFilter

protected java.util.function.Predicate<TypedDependency> extraTreeDepFilter()

Returns a Filter which checks dependencies for usefulness as extra tree-based dependencies. By default, everything is accepted. One example of how this can be useful is in the English dependencies, where the REL dependency is used as an intermediate and we do not want this to be added when we make a second pass over the trees for missing dependencies.

Overrides:

extraTreeDepFilter in class GrammaticalStructure

getTreeDeps

protected void getTreeDeps(java.util.List<TypedDependency> deps,
                           DirectedMultiGraph<TreeGraphNode,GrammaticalRelation> completeGraph,
                           java.util.function.Predicate<TypedDependency> puncTypedDepFilter,
                           java.util.function.Predicate<TypedDependency> extraTreeDepFilter)

Description copied from class: GrammaticalStructure

Look through the tree t and adds to the List basicDep additional dependencies which aren't in the List but which satisfy the filter puncTypedDepFilter.

Overrides:

getTreeDeps in class GrammaticalStructure

Parameters:

deps - The list of dependencies which may be augmented

completeGraph - a graph of all the tree dependencies found earlier

puncTypedDepFilter - The filter that may skip punctuation dependencies

extraTreeDepFilter - Additional dependencies are added only if they pass this filter

correctDependencies

protected void correctDependencies(java.util.List<TypedDependency> list)

Description copied from class: GrammaticalStructure

Destructively modify the TypedDependencyGraph to correct language-dependent dependencies. (e.g., nsubjpass in a relative clause)
Default is no-op; to be over-ridden in subclasses.

Overrides:

correctDependencies in class GrammaticalStructure

correctDependencies

protected static void correctDependencies(SemanticGraph sg)

postProcessDependencies

protected void postProcessDependencies(java.util.List<TypedDependency> list)

Description copied from class: GrammaticalStructure

Post process the dependencies in whatever way this language requires. For example, English might replace "rel" dependencies with either dobj or pobj depending on the surrounding dependencies.

Overrides:

postProcessDependencies in class GrammaticalStructure

postProcessDependencies

protected static void postProcessDependencies(SemanticGraph sg)

getExtras

protected void getExtras(java.util.List<TypedDependency> list)

Description copied from class: GrammaticalStructure

Get extra dependencies that do not depend on the tree structure, but rather only depend on the existing dependency structure. For example, the English xsubj dependency can be extracted that way.

Overrides:

getExtras in class GrammaticalStructure

addEnhancements

public static void addEnhancements(SemanticGraph sg,
                                   EnhancementOptions options)

Manipulates the given SemanticGraph to add enhancements. Note that this modifies the input graph.

addEnhancements

protected void addEnhancements(java.util.List<TypedDependency> list,
                               EnhancementOptions options)

Description copied from class: GrammaticalStructure

Destructively applies different enhancements to the dependency graph.
Default is no-op; to be over-ridden in subclasses.

Overrides:

addEnhancements in class GrammaticalStructure

Parameters:

list - A list of dependencies

options - Options that determine which enhancements are applied to the dependency graph.

collapseDependencies

protected void collapseDependencies(java.util.List<TypedDependency> list,
                                    boolean CCprocess,
                                    GrammaticalStructure.Extras includeExtras)

Destructively modifies this Collection<TypedDependency> by collapsing several types of transitive pairs of dependencies or by adding additional information from the dependents to the relation of the governor. If called with a tree of dependencies and both CCprocess and includeExtras set to false, then the tree structure is preserved.

nominal modifier dependencies: nmod

If there exist the relations case(hat, in) and nmod(in, hat) then the nmod relation is enhanced to nmod:in(cat, hat). The case(hat, in) relation is preserved.

clausal modifier of noun/adverbial clause modifier with case markers: acs/advcl

If there exist the relations case(attacking, of) and advcl(heard, attacking) then the nmod relation is enhanced to nmod:of(heard, attacking). The case(attacking, of) relation is preserved.

conjunct dependencies

If there exist the relations cc(investors, and) and conj(investors, regulators), then the conj relation is enhanced to conj:and(investors, regulators)

For relative clauses, it will collapse referent

ref(man, that) and dobj(love, that) are collapsed to dobj(love, man)

Overrides:

collapseDependencies in class GrammaticalStructure

Parameters:

list - A list of dependencies to process for possible collapsing

CCprocess - apply CC process?

collapseDependenciesTree

protected void collapseDependenciesTree(java.util.List<TypedDependency> list)

Description copied from class: GrammaticalStructure

Destructively modify the Collection<TypedDependency> to collapse language-dependent transitive dependencies but keeping a tree structure.
Default is no-op; to be over-ridden in subclasses.

Overrides:

collapseDependenciesTree in class GrammaticalStructure

Parameters:

list - A list of dependencies to process for possible collapsing

readCoNLLXGrammaticalStructureCollection

public static java.util.List<GrammaticalStructure> readCoNLLXGrammaticalStructureCollection(java.lang.String fileName)
                                                                                     throws java.io.IOException

Throws:

java.io.IOException

buildCoNLLXGrammaticalStructure

public static UniversalEnglishGrammaticalStructure buildCoNLLXGrammaticalStructure(java.util.List<java.util.List<java.lang.String>> tokenFields)