floraachy
/
vChewing-macOS
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

LMAssoc. // Also parse data exported from Windows IME.

This commit is contained in:
ShikiSuen 2022-07-17 13:41:14 +08:00
parent ebde0d088e
commit d7550f465b
1 changed files with 23 additions and 27 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

50
Source/Modules/LangModelRelated/SubLMs/lmAssociates.swift
View File

@ -27,7 +27,7 @@ import Foundation
extension vChewing { extension vChewing {
@frozen public struct LMAssociates { @frozen public struct LMAssociates {
var rangeMap: [String: [Range<String.Index>]] = [:] var rangeMap: [String: [(Range<String.Index>, Int)]] = [:]
var strData: String = "" var strData: String = ""
public var count: Int { public var count: Int {
@ -56,9 +56,11 @@ extension vChewing {
let neta = strData[$0].split(separator: " ") let neta = strData[$0].split(separator: " ")
if neta.count >= 2 { if neta.count >= 2 {
let theKey = String(neta[0]) let theKey = String(neta[0])
if !neta[0].isEmpty, !neta[1].isEmpty, theKey.first != "#" { if !theKey.isEmpty, theKey.first != "#" {
let theValue = $0 for (i, _) in neta.filter({ $0.first != "#" && !$0.isEmpty }).enumerated() {
rangeMap[theKey, default: []].append(theValue) if i == 0 { continue }
rangeMap[theKey, default: []].append(($0, i))
}
} }
} }
} }
@ -78,24 +80,16 @@ extension vChewing {
} }
public func dump() { public func dump() {
var strDump = "" // We remove this function in order to reduce out maintenance workload.
for entry in rangeMap { // This function will be implemented only if further hard-necessity comes.
let netaRanges: [Range<String.Index>] = entry.value
for netaRange in netaRanges {
let neta = strData[netaRange]
let addline = neta + "\n"
strDump += addline
}
}
IME.prtDebugIntel(strDump)
} }
public func valuesFor(key: String) -> [String] { public func valuesFor(key: String) -> [String] {
var pairs: [String] = [] var pairs: [String] = []
if let arrRangeRecords: [Range<String.Index>] = rangeMap[key] { if let arrRangeRecords: [(Range<String.Index>, Int)] = rangeMap[key] {
for netaRange in arrRangeRecords { for (netaRange, index) in arrRangeRecords {
let neta = strData[netaRange].split(separator: " ") let neta = strData[netaRange].split(separator: " ")
let theValue: String = .init(neta[1]) let theValue: String = .init(neta[index])
pairs.append(theValue) pairs.append(theValue)
} }
} }
@ -108,20 +102,22 @@ extension vChewing {
public func valuesFor(pair: Megrez.KeyValuePaired) -> [String] { public func valuesFor(pair: Megrez.KeyValuePaired) -> [String] {
var pairs: [String] = [] var pairs: [String] = []
if let arrRangeRecords: [Range<String.Index>] = rangeMap[pair.toNGramKey] { if let arrRangeRecords: [(Range<String.Index>, Int)] = rangeMap[pair.toNGramKey] {
for netaRange in arrRangeRecords { for (netaRange, index) in arrRangeRecords {
let neta = strData[netaRange].split(separator: " ") let neta = strData[netaRange].split(separator: " ")
let theValue: String = .init(neta[1]) let theValue: String = .init(neta[index])
pairs.append(theValue)
}
} else if let arrRangeRecords: [Range<String.Index>] = rangeMap[pair.value] {
for netaRange in arrRangeRecords {
let neta = strData[netaRange].split(separator: " ")
let theValue: String = .init(neta[1])
pairs.append(theValue) pairs.append(theValue)
} }
} }
return pairs if let arrRangeRecords: [(Range<String.Index>, Int)] = rangeMap[pair.value] {
for (netaRange, index) in arrRangeRecords {
let neta = strData[netaRange].split(separator: " ")
let theValue: String = .init(neta[index])
pairs.append(theValue)
}
}
var set = Set<String>()
return pairs.filter { set.insert($0).inserted }
} }
public func hasValuesFor(pair: Megrez.KeyValuePaired) -> Bool { public func hasValuesFor(pair: Megrez.KeyValuePaired) -> Bool {