Pre Merge pull request !60 from ShikiSuen/upd/1.8.0sp1
This commit is contained in:
commit
41ebe3dd54
|
@ -38,7 +38,12 @@ extension Megrez {
|
||||||
private var langModel: LangModelProtocol
|
private var langModel: LangModelProtocol
|
||||||
/// 允許查詢當前游標位置屬於第幾個幅位座標(從 0 開始算)。
|
/// 允許查詢當前游標位置屬於第幾個幅位座標(從 0 開始算)。
|
||||||
private(set) var cursorRegionMap: [Int: Int] = .init()
|
private(set) var cursorRegionMap: [Int: Int] = .init()
|
||||||
private(set) var walkedAnchors: [Megrez.NodeAnchor] = [] // 用以記錄爬過的節錨的陣列
|
/// 用以記錄爬過的節錨的陣列
|
||||||
|
private(set) var walkedAnchors: [NodeAnchor] = []
|
||||||
|
/// 該函式用以更新爬過的節錨的陣列
|
||||||
|
public func updateWalkedAnchors(with nodes: [Node]) {
|
||||||
|
walkedAnchors = nodes.map { Megrez.NodeAnchor(node: $0) }
|
||||||
|
}
|
||||||
|
|
||||||
/// 公開:多字讀音鍵當中用以分割漢字讀音的記號,預設為空。
|
/// 公開:多字讀音鍵當中用以分割漢字讀音的記號,預設為空。
|
||||||
public var joinSeparator: String = "-"
|
public var joinSeparator: String = "-"
|
||||||
|
@ -88,7 +93,7 @@ extension Megrez {
|
||||||
/// - separator: 多字讀音鍵當中用以分割漢字讀音的記號,預設為空。
|
/// - separator: 多字讀音鍵當中用以分割漢字讀音的記號,預設為空。
|
||||||
public init(lm: LangModelProtocol, length: Int = 10, separator: String = "-") {
|
public init(lm: LangModelProtocol, length: Int = 10, separator: String = "-") {
|
||||||
langModel = lm
|
langModel = lm
|
||||||
super.init(spanLength: abs(length)) // 防呆
|
super.init(spanLengthLimit: abs(length)) // 防呆
|
||||||
joinSeparator = separator
|
joinSeparator = separator
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -181,7 +186,7 @@ extension Megrez {
|
||||||
|
|
||||||
var paths = [[NodeAnchor]]()
|
var paths = [[NodeAnchor]]()
|
||||||
let nodes = nodesEndingAt(location: location).stableSorted {
|
let nodes = nodesEndingAt(location: location).stableSorted {
|
||||||
$0.scoreForSort > $1.scoreForSort
|
$0.node.score > $1.node.score
|
||||||
}
|
}
|
||||||
|
|
||||||
guard !nodes.isEmpty else { return .init() } // 防止下文出現範圍外索引的錯誤
|
guard !nodes.isEmpty else { return .init() } // 防止下文出現範圍外索引的錯誤
|
||||||
|
@ -270,7 +275,7 @@ extension Megrez {
|
||||||
if hasMatchedNode(location: p, spanLength: q, key: combinedReading) { continue }
|
if hasMatchedNode(location: p, spanLength: q, key: combinedReading) { continue }
|
||||||
let unigrams: [Unigram] = langModel.unigramsFor(key: combinedReading)
|
let unigrams: [Unigram] = langModel.unigramsFor(key: combinedReading)
|
||||||
if unigrams.isEmpty { continue }
|
if unigrams.isEmpty { continue }
|
||||||
let n = Node(key: combinedReading, unigrams: unigrams)
|
let n: Node = .init(key: combinedReading, spanLength: q, unigrams: unigrams)
|
||||||
insertNode(node: n, location: p, spanLength: q)
|
insertNode(node: n, location: p, spanLength: q)
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
|
@ -41,8 +41,8 @@ extension Megrez {
|
||||||
public var isEmpty: Bool { spans.isEmpty }
|
public var isEmpty: Bool { spans.isEmpty }
|
||||||
|
|
||||||
/// 初期化轨格。
|
/// 初期化轨格。
|
||||||
public init(spanLength: Int = 10) {
|
public init(spanLengthLimit: Int = 10) {
|
||||||
maxBuildSpanLength = spanLength
|
maxBuildSpanLength = spanLengthLimit
|
||||||
spans = [Megrez.SpanUnit]()
|
spans = [Megrez.SpanUnit]()
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -114,13 +114,7 @@ extension Megrez {
|
||||||
let span = spans[location]
|
let span = spans[location]
|
||||||
for i in 1...maxBuildSpanLength {
|
for i in 1...maxBuildSpanLength {
|
||||||
if let np = span.nodeOf(length: i) {
|
if let np = span.nodeOf(length: i) {
|
||||||
results.append(
|
results.append(.init(node: np))
|
||||||
.init(
|
|
||||||
node: np,
|
|
||||||
location: location,
|
|
||||||
spanLength: i
|
|
||||||
)
|
|
||||||
)
|
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
return results // 已證實不會有空節點產生。
|
return results // 已證實不會有空節點產生。
|
||||||
|
@ -137,13 +131,7 @@ extension Megrez {
|
||||||
let span = spans[i]
|
let span = spans[i]
|
||||||
if i + span.maxLength < location { continue }
|
if i + span.maxLength < location { continue }
|
||||||
if let np = span.nodeOf(length: location - i) {
|
if let np = span.nodeOf(length: location - i) {
|
||||||
results.append(
|
results.append(.init(node: np))
|
||||||
.init(
|
|
||||||
node: np,
|
|
||||||
location: i,
|
|
||||||
spanLength: location - i
|
|
||||||
)
|
|
||||||
)
|
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
return results // 已證實不會有空節點產生。
|
return results // 已證實不會有空節點產生。
|
||||||
|
@ -162,13 +150,7 @@ extension Megrez {
|
||||||
for j in 1...span.maxLength {
|
for j in 1...span.maxLength {
|
||||||
if i + j < location { continue }
|
if i + j < location { continue }
|
||||||
if let np = span.nodeOf(length: j) {
|
if let np = span.nodeOf(length: j) {
|
||||||
results.append(
|
results.append(.init(node: np))
|
||||||
.init(
|
|
||||||
node: np,
|
|
||||||
location: i,
|
|
||||||
spanLength: location - i
|
|
||||||
)
|
|
||||||
)
|
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
@ -193,7 +175,7 @@ extension Megrez {
|
||||||
@discardableResult public func fixNodeWithCandidateLiteral(_ value: String, at location: Int) -> NodeAnchor {
|
@discardableResult public func fixNodeWithCandidateLiteral(_ value: String, at location: Int) -> NodeAnchor {
|
||||||
let location = abs(location) // 防呆
|
let location = abs(location) // 防呆
|
||||||
var node = NodeAnchor()
|
var node = NodeAnchor()
|
||||||
for theAnchor in nodesOverlappedAt(location: location) {
|
for theAnchor in nodesCrossingOrEndingAt(location: location) {
|
||||||
let candidates = theAnchor.node.candidates
|
let candidates = theAnchor.node.candidates
|
||||||
// 將該位置的所有節點的候選字詞鎖定狀態全部重設。
|
// 將該位置的所有節點的候選字詞鎖定狀態全部重設。
|
||||||
theAnchor.node.resetCandidate()
|
theAnchor.node.resetCandidate()
|
||||||
|
@ -217,7 +199,7 @@ extension Megrez {
|
||||||
@discardableResult public func fixNodeWithCandidate(_ pair: KeyValuePaired, at location: Int) -> NodeAnchor {
|
@discardableResult public func fixNodeWithCandidate(_ pair: KeyValuePaired, at location: Int) -> NodeAnchor {
|
||||||
let location = abs(location) // 防呆
|
let location = abs(location) // 防呆
|
||||||
var node = NodeAnchor()
|
var node = NodeAnchor()
|
||||||
for theAnchor in nodesOverlappedAt(location: location) {
|
for theAnchor in nodesCrossingOrEndingAt(location: location) {
|
||||||
let candidates = theAnchor.node.candidates
|
let candidates = theAnchor.node.candidates
|
||||||
// 將該位置的所有節點的候選字詞鎖定狀態全部重設。
|
// 將該位置的所有節點的候選字詞鎖定狀態全部重設。
|
||||||
theAnchor.node.resetCandidate()
|
theAnchor.node.resetCandidate()
|
||||||
|
|
|
@ -30,28 +30,39 @@ extension Megrez {
|
||||||
public var isEmpty: Bool { node.key.isEmpty }
|
public var isEmpty: Bool { node.key.isEmpty }
|
||||||
/// 節點。一個節锚內不一定有節點。
|
/// 節點。一個節锚內不一定有節點。
|
||||||
public var node: Node = .init()
|
public var node: Node = .init()
|
||||||
/// 節锚所在的位置。
|
|
||||||
public var location: Int = 0
|
|
||||||
/// 指定的幅位長度。
|
/// 指定的幅位長度。
|
||||||
public var spanLength: Int = 0
|
public var spanLength: Int { node.spanLength }
|
||||||
|
/// 獲取用來比較的權重。
|
||||||
|
public var scoreForSort: Double { node.score }
|
||||||
/// 累計權重。
|
/// 累計權重。
|
||||||
public var mass: Double = 0.0
|
public var mass: Double = 0.0
|
||||||
|
/// 單元圖陣列。
|
||||||
|
public var unigrams: [Unigram] { node.unigrams }
|
||||||
|
/// 雙元圖陣列。
|
||||||
|
public var bigrams: [Bigram] { node.bigrams }
|
||||||
|
/// 鍵。
|
||||||
|
public var key: String { node.key }
|
||||||
/// 索引鍵的長度。
|
/// 索引鍵的長度。
|
||||||
public var keyLength: Int {
|
public var keyLength: Int {
|
||||||
isEmpty ? node.key.count : 0
|
isEmpty ? node.key.count : 0
|
||||||
}
|
}
|
||||||
|
|
||||||
|
/// 初期化一個節錨。
|
||||||
|
public init(node: Node = .init(), mass: Double? = nil) {
|
||||||
|
self.node = node
|
||||||
|
self.mass = mass ?? self.node.score
|
||||||
|
}
|
||||||
|
|
||||||
|
/// 將該節錨雜湊化。
|
||||||
public func hash(into hasher: inout Hasher) {
|
public func hash(into hasher: inout Hasher) {
|
||||||
hasher.combine(node)
|
hasher.combine(node)
|
||||||
hasher.combine(location)
|
|
||||||
hasher.combine(spanLength)
|
|
||||||
hasher.combine(mass)
|
hasher.combine(mass)
|
||||||
}
|
}
|
||||||
|
|
||||||
/// 將當前節锚列印成一個字串。
|
/// 將當前節锚列印成一個字串。
|
||||||
public var description: String {
|
public var description: String {
|
||||||
var stream = ""
|
var stream = ""
|
||||||
stream += "{@(" + String(location) + "," + String(spanLength) + "),"
|
stream += "{@(" + String(spanLength) + "),"
|
||||||
if node.key.isEmpty {
|
if node.key.isEmpty {
|
||||||
stream += node.description
|
stream += node.description
|
||||||
} else {
|
} else {
|
||||||
|
@ -60,11 +71,6 @@ extension Megrez {
|
||||||
stream += "}"
|
stream += "}"
|
||||||
return stream
|
return stream
|
||||||
}
|
}
|
||||||
|
|
||||||
/// 獲取用來比較的權重。
|
|
||||||
public var scoreForSort: Double {
|
|
||||||
isEmpty ? node.score : 0
|
|
||||||
}
|
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
|
@ -30,7 +30,7 @@ extension Megrez {
|
||||||
lhs.key == rhs.key && lhs.score == rhs.score && lhs.unigrams == rhs.unigrams && lhs.bigrams == rhs.bigrams
|
lhs.key == rhs.key && lhs.score == rhs.score && lhs.unigrams == rhs.unigrams && lhs.bigrams == rhs.bigrams
|
||||||
&& lhs.candidates == rhs.candidates && lhs.valueUnigramIndexMap == rhs.valueUnigramIndexMap
|
&& lhs.candidates == rhs.candidates && lhs.valueUnigramIndexMap == rhs.valueUnigramIndexMap
|
||||||
&& lhs.precedingBigramMap == rhs.precedingBigramMap && lhs.isCandidateFixed == rhs.isCandidateFixed
|
&& lhs.precedingBigramMap == rhs.precedingBigramMap && lhs.isCandidateFixed == rhs.isCandidateFixed
|
||||||
&& lhs.selectedUnigramIndex == rhs.selectedUnigramIndex
|
&& lhs.selectedUnigramIndex == rhs.selectedUnigramIndex && lhs.spanLength == rhs.spanLength
|
||||||
}
|
}
|
||||||
|
|
||||||
public func hash(into hasher: inout Hasher) {
|
public func hash(into hasher: inout Hasher) {
|
||||||
|
@ -38,6 +38,7 @@ extension Megrez {
|
||||||
hasher.combine(score)
|
hasher.combine(score)
|
||||||
hasher.combine(unigrams)
|
hasher.combine(unigrams)
|
||||||
hasher.combine(bigrams)
|
hasher.combine(bigrams)
|
||||||
|
hasher.combine(spanLength)
|
||||||
hasher.combine(candidates)
|
hasher.combine(candidates)
|
||||||
hasher.combine(valueUnigramIndexMap)
|
hasher.combine(valueUnigramIndexMap)
|
||||||
hasher.combine(precedingBigramMap)
|
hasher.combine(precedingBigramMap)
|
||||||
|
@ -50,9 +51,11 @@ extension Megrez {
|
||||||
/// 當前節點的當前被選中的候選字詞「在該節點內的」目前的權重。
|
/// 當前節點的當前被選中的候選字詞「在該節點內的」目前的權重。
|
||||||
private(set) var score: Double = 0
|
private(set) var score: Double = 0
|
||||||
/// 單元圖陣列。
|
/// 單元圖陣列。
|
||||||
private var unigrams: [Unigram]
|
private(set) var unigrams: [Unigram]
|
||||||
/// 雙元圖陣列。
|
/// 雙元圖陣列。
|
||||||
private var bigrams: [Bigram]
|
private(set) var bigrams: [Bigram]
|
||||||
|
/// 指定的幅位長度。
|
||||||
|
public var spanLength: Int = 0
|
||||||
/// 候選字詞陣列,以鍵值陣列的形式存在。
|
/// 候選字詞陣列,以鍵值陣列的形式存在。
|
||||||
private(set) var candidates: [KeyValuePaired] = []
|
private(set) var candidates: [KeyValuePaired] = []
|
||||||
/// 專門「用單元圖資料值來調查索引值」的辭典。
|
/// 專門「用單元圖資料值來調查索引值」的辭典。
|
||||||
|
@ -83,10 +86,11 @@ extension Megrez {
|
||||||
/// - key: 索引鍵。
|
/// - key: 索引鍵。
|
||||||
/// - unigrams: 單元圖陣列。
|
/// - unigrams: 單元圖陣列。
|
||||||
/// - bigrams: 雙元圖陣列(非必填)。
|
/// - bigrams: 雙元圖陣列(非必填)。
|
||||||
public init(key: String = "", unigrams: [Megrez.Unigram] = [], bigrams: [Megrez.Bigram] = []) {
|
public init(key: String = "", spanLength: Int = 0, unigrams: [Megrez.Unigram] = [], bigrams: [Megrez.Bigram] = []) {
|
||||||
self.key = key
|
self.key = key
|
||||||
self.unigrams = unigrams
|
self.unigrams = unigrams
|
||||||
self.bigrams = bigrams
|
self.bigrams = bigrams
|
||||||
|
self.spanLength = spanLength
|
||||||
|
|
||||||
self.unigrams.sort {
|
self.unigrams.sort {
|
||||||
$0.score > $1.score
|
$0.score > $1.score
|
||||||
|
|
Loading…
Reference in New Issue