This is the internal documentation for the Transducer
class. Wherever details are omitted, please refer to the original code.
self.rootWord
Type: String?
Value: nil
Placeholder for root word.
self.unkWord
Type: String
Value: _
Placeholder for unknown word.
self.rootTag
Type: String
Value: TOP
POS tag assigned to root node.
self.unkTag
Type: String
Value: _
POS tag assigned to unknown word.
self.rootDeprel
Type: String
Value: ROOT
Dependency tag between root node and head word.
self.unkDeprel
Type: String
Value: _
Dependency tag for unknown relation.
self.id2word
Type: [String?]
Wordlist as an array, plus the unknown and root word. Position of word corresponds to its ID.
self.id2tag
Type: [String?]
POS tags as an array, plus the tag for unknown and root word. Position of tag corresponds to its ID.
self.id2deprel
Type: [String?]
Dependency tags as an array, plus the tag for unknown and root word. Position of tag corresponds to its ID.
self.word2id
Type: [String: Int]
Dictionary used to translate between words and their corresponding IDs.
self.tag2id
Type: [String: Int]
Dictionary used to translate between POS tags and their corresponding IDs.
self.deprel2id
Type: [String: Int]
Dictionary used to translate between dependency tags and their corresponding IDs.
self.unkWordId
Type: Int
Word ID for an unknown word.
self.unkTagId
Type: Int
POS tag ID for an unknown word.
self.unkDeprelId
Type: Int
Dependency tag ID for an unknown word.
self.nullWordId
Type: Int
Word ID that denotes null.
self.nullTagId
Type: Int
POS tag ID that denotes null.
self.nullDeprelId
Type: Int
Dependency tag ID that denotes null.
Arguments
wordList
: wordlist as an array of String
.
tagList
: list of POS tags as an array of String
.
deprelList
: list of dependency tags as an array of String
.
Purpose Initialize instance variables with the appropriate values.
Returns None.
Arguments
array
: the array of Int
to convert .
Purpose
Converts an array of Int
into the CoreML-compatible MLMultiArray
.
Returns
Returns the converted array as MLMultiArray
.
Arguments
partial
: the PartialParse
object from which to construct a feature vector triple.
Purpose
Constructs a feature vector triple (word, pos, deprel)
from a PartialParse
object.
Returns
Returns (word, pos, deprel)
, all of which are of type MLMultiArray
.
The following two paragraphs will be difficult to make sense of without first understanding the paper our dependency parser comes from.
The word
and tag
vectors have 18 values each for a single sentence. They are: top 3 ids on stack, top 3 ids on buffer, 1st and 2nd leftmost and rightmost dependants from top two words on stack (8), and leftmost-leftmost and rightmost-rightmost of top two words on stack (4).
The deprel
vector has 12 values for a single sentence. They are: 1st and 2nd leftmost and rightmost dependants from top two words on stack (8), and the leftmost-leftmost and rightmost-rightmost of top two words on stack (4).
Arguments
partial
: the trans/deprel
vector to translate, as an MLMultiarray
.
Purpose
Converts a trans/deprel
vector into a trans,deprel
pair.
Returns
Returns (transition_id, deprel)
.