CCGParser
Synopsis
Syntax parsing with CCG Parser .
Description
CCGParser applies the CCG Parser to sentences specified as annotations from the sentenceLayer layer. Sentence words are specified by annotations in the wordLayer layer. For each sentence, only words entirely included in the sentence will be considered; WoSMig and SeSMig should create these layers with the appropriate annotations. Additionally CCGParser takes advantage of word POS tag specified in the posFeature feature.
CCGParser creates a relation named dependencyRelation in each section and a tuple in this relation for each dependency. This relation is ternary:
- dependencySentenceRole : the first argument is the sentence in which the dependency was found;
- headRole : the second argument is the head word of the dependency;
- dependentRole : the third argument is the dependent word of the dependency.
CCGParser adds to each dependency tuple a feature dependencyLabelFeature with the label of the dependency.
Snippet
<ccgparser class="CCGParser">
<executable></executable>
<parserModel></parserModel>
<superModel></superModel>
</ccgparser>
Mandatory parameters
executable
Path to the CCG Parser executable.
parserModel
Path to the parser model file.
superModel
Path to the CCG supertagger model file.
Optional parameters
constantRelationFeatures
Constant features to add to each relation created by this module.
constantTupleFeatures
Constant features to add to each tuple created by this module.
stanfordMarkedUpScript
Path to the markedup script for Stanford tagset output. See Biomedical parsing for CCG .
stanfordScript
Post-processing script for Stanford tagset output. See Biomedical parsing for CCG .
dependencyLabelFeature
Name of the feature where to store the dependency label.
dependencyRelation
Name of the relation where to store dependency tuples.
dependencySentenceRole
Name of the role of the dependency tuple argument that references the parsed sentence.
dependentRole
Name of the role of the dependency tuple argument that references the modifier (dependent) token.
documentFilter
Only process document that satisfy this expression.
formFeature
Name of the feature in word annotations that contains the surface form.
headRole
Name of the role of the dependency tuple argument that references the head (governor) token.
internalEncoding
Character encoding to use for CCG input and output files.
lpTransformation
Either to translate into LP tag-set.
maxRuns
Maximal number of CCG runs.
maxSuperCats
Maximum number of supercats before the parse explodes (cited from CCG documentation).
posFeature
Name of the feature in word annotations where to write POS tags.
sectionFilter
Process only sections that satisfy this expression.
sentenceFilter
Process only sentences that satisfy this filter.
sentenceLayer
Name of the layer containing sentence annotations.
supertagFeature
Name of the feature containing the supertag label.
wordLayer
Name of the layer containing word annotations.
Deprecated parameters
formFeatureName
Deprecated alias for formFeature .
labelFeatureName
Deprecated alias for dependencyLabelFeature .
posFeatureName
Deprecated alias for posFeature .
relationName
Deprecated alias for dependencyRelation .
sentenceLayerName
Deprecated alias for sentenceLayer .
supertagFeatureName
Deprecated alias for supertagFeature .
wordLayerName
Deprecated alias for wordLayer .