initial commit

.editorconfig

@@ -0,0 +1,9 @@
+root = true
+
+[*]
+insert_final_newline = true
+indent_style = space
+indent_size = 2
+end_of_line = lf
+charset = utf-8
+trim_trailing_whitespace = true

.gitattributes

@@ -0,0 +1,7 @@
+* text eol=lf
+
+*.png -text
+*.jpg -text
+*.ico -text
+*.gif -text
+*.webp -text

.gitignore

@@ -0,0 +1,18 @@
+lib
+dist
+external
+
+node_modules
+npm-debug.log
+yarn-debug.log
+yarn-error.log
+tsconfig.tsbuildinfo
+
+.eslintcache
+.DS_Store
+.idea
+.vscode
+*.suo
+*.ntvs*
+*.njsproj
+*.sln

.npmignore

@@ -0,0 +1,2 @@
+.DS_Store
+tsconfig.tsbuildinfo

package.json

@@ -0,0 +1,25 @@
+{
+  "name": "@dgck81lnn/koishi-plugin-xdi8",
+  "description": "希顶语与汉字互转",
+  "version": "0.1.0",
+  "main": "lib/index.js",
+  "typings": "lib/index.d.ts",
+  "files": [
+    "lib",
+    "dist"
+  ],
+  "license": "MIT",
+  "scripts": {},
+  "keywords": [
+    "chatbot",
+    "koishi",
+    "plugin",
+    "shidinn"
+  ],
+  "peerDependencies": {
+    "koishi": "^4.16.1"
+  },
+  "dependencies": {
+    "xdi8-transcriber": "file:../../../../soul/xdi8-transcriber/package"
+  }
+}

readme.md

@@ -0,0 +1,78 @@
+# @dgck81lnn/koishi-plugin-xdi8
+
+[![npm](https://img.shields.io/npm/v/@dgck81lnn/koishi-plugin-xdi8?style=flat-square)](https://www.npmjs.com/package/@dgck81lnn/koishi-plugin-xdi8)
+
+希顶语与汉字互转
+
+基于本人维护的 npm 包 [xdi8-transcriber](https://github.com/DGCK81LNN/xdi8-transcriber)
+
+## 用法
+
+  > 指令：`xdi8 <text...>`
+  >
+  > 汉字希顶互转（实验性）
+  >
+  > 在文本前添加选项“-a”（与文本和指令名“xdi8”之间用空格隔开）来显示隐藏的结果，包括过时拼写和例外结果。加“-A”可显示例外结果，但不包含过时拼写。
+  >
+  > “例外结果”包括希转汉时的部分繁体字和汉转希时多音字的罕见读音。
+  >
+  > 当文本为单个汉字或不包含空格的希顶“词”时，会默认显示所有隐藏结果。
+  >
+  > 可用的选项有：
+  >
+  >   * -a, --all  总是显示所有隐藏结果
+  >   * -A, --almost-all  总是显示例外结果（不包括过时拼写）
+
+## 示例
+
+`xdi8 萤火虫`
+
+```
+nu3k ho ci3
+```
+
+`xdi8 地`
+
+```
+地:
+dde（助词 de）
+dDE（名词 dì）
+```
+
+`xdi8 怒发冲冠  奋发图强`
+
+```
+nAF jbia¹ mzu3² HB2  hu8H pio³ mb1 qT
+[1] 发:
+jbia（“髪”的简化字 fà）
+pio（“發”的简化字 fā）
+[2] 冲:
+mzu3（“衝”的简化字）
+Du3E（同“沖” 山间的平地；用于地名）
+[3] 发:
+pio（“發”的简化字 fā）
+jbia（“髪”的简化字 fà）
+```
+
+`xdi8 tof H6H huT vnuV nYH`
+
+```
+施氏食狮史
+```
+
+`xdi8 NAh`
+
+```
+NAh:
+曲（“麴”的简化字 酿酒或制酱时引起发醇的东西）
+麹
+麴
+```
+
+<code>xdi8<br>nAF nE wiY vnuV<br>nE wiY 4i6 n5i6<br>wiY 4i6 wA 56</code>
+
+```
+怒尼威狮
+尼威又鸡
+威又乌犀
+```

src/index.ts

@@ -0,0 +1,161 @@
+import { Context, Schema, Session } from "koishi"
+import {
+  Alternation,
+  HanziToXdi8Transcriber,
+  TranscribeResult,
+  Xdi8ToHanziTranscriber,
+} from "xdi8-transcriber"
+
+export const name = "xdi8"
+
+export interface Config {}
+
+export const Config: Schema<Config> = Schema.object({})
+
+function supNum(n: number) {
+  return Array.from(String(0 | n), (d: `${number}`) => "⁰¹²³⁴⁵⁶⁷⁸⁹"[d]).join("")
+}
+
+let hxTranscriber: HanziToXdi8Transcriber
+let xhTranscriber: Xdi8ToHanziTranscriber
+
+export function apply(ctx: Context) {
+  /**
+   * Dict of shidinn spellings and their preferred hanzi forms.
+   *
+   * Some entries have hanzi froms that include PUA characters which will not
+   * display in plain-text environments. Thus, when characters in this dict are
+   * transcribed to hanzi, only the preferred hanzi forms will be shown --
+   * unless the `all` flag is set, in which case they are moved to the end of
+   * the alternations array instead.
+   */
+  const ahoFixes: Record<string, string[]> = {
+    aho: ["纟火", "糹火"],
+  }
+
+  function stringifyResult<T extends "h" | "x">(
+    session: Session,
+    result: TranscribeResult,
+    sourceType: T,
+    { all = false, almostAll = false }
+  ) {
+    const single = result.length === 1 && Array.isArray(result[0])
+
+    const showLegacy = single || all
+    const showExceptional = single || all || almostAll
+    result = result.flatMap(seg => {
+      if (Array.isArray(seg)) {
+        if (!all) {
+          seg = seg.filter(alt => {
+            if (!showLegacy && alt.legacy) return false
+            if (!showExceptional && alt.exceptional) return false
+            // aho fix when `all` flag is set: only keep preferred forms
+            if (
+              sourceType === "x" &&
+              alt.content.some(
+                seg => Object.hasOwn(ahoFixes, seg.x) && !ahoFixes[seg.x].includes(seg.v)
+              )
+            )
+              return false
+            return true
+          })
+        } else if (
+          seg.some(alt => alt.content.some(seg => Object.hasOwn(ahoFixes, seg.x)))
+        ) {
+          // aho fix when `all` flag is not set: move non-preferred forms to bottom
+          const good: Alternation[] = []
+          const bad: Alternation[] = []
+          for (const alt of seg) {
+            const isBad = alt.content.some(
+              seg => Object.hasOwn(ahoFixes, seg.x) && !ahoFixes[seg.x].includes(seg.v)
+            )
+            ;(isBad ? bad : good).push(alt)
+          }
+          seg = good.concat(bad)
+        }
+        if (seg.length === 1) return seg[0].content
+      }
+      return [seg]
+    })
+
+    const alts: (Alternation[] & { $: string })[] = []
+    const text = result
+      .map(seg => {
+        if (typeof seg === "string") return seg
+        if (Array.isArray(seg)) {
+          const j = JSON.stringify(seg)
+          let index = alts.findIndex(s => s.$ === j)
+          if (index === -1) {
+            index = alts.length
+            alts.push(Object.assign(seg, { $: j }))
+          }
+          return seg[0].content.map(seg => seg.v).join("") + supNum(index + 1)
+        }
+        return seg.v
+      })
+      .join("")
+
+    const footnotes = alts.map(seg => {
+      const source = seg[0].content.map(seg => seg[sourceType]).join("")
+      const alts = seg.map(alt => {
+        let line = alt.content.map(seg => seg.v).join("")
+        if (alt.note)
+          line += session.text("general.paren", [alt.note.replace(/\n/g, "；")])
+        return line
+      })
+      return `${source}:\n${alts.join("\n")}`
+    })
+
+    if (single && footnotes.length === 1) return footnotes[0]
+    return `${text}\n${footnotes.map((fn, i) => `[${i + 1}] ${fn}`).join("\n")}`
+  }
+
+  function getResultScore(result: TranscribeResult) {
+    if (!result) return 0
+    return result.reduce(
+      (score, seg) =>
+        score + (Array.isArray(seg) ? seg[0].content.length : +(typeof seg === "object")),
+      0
+    )
+  }
+
+  const cmdXdi8 = ctx.command("xdi8 <text:text>", {
+    checkArgCount: true,
+    checkUnknown: true,
+    showWarning: true,
+  })
+  cmdXdi8
+    .option("all", "-a")
+    .option("almostAll", "-A")
+    .action(({ options, session }, text) => {
+      hxTranscriber ||= new HanziToXdi8Transcriber()
+      xhTranscriber ||= new Xdi8ToHanziTranscriber()
+
+      const hxResult = hxTranscriber.transcribe(text, { ziSeparator: " " })
+      const hxScore = getResultScore(hxResult)
+      const xhResult = xhTranscriber.transcribe(text, { alphaFilter: null })
+      const xhScore = getResultScore(xhResult)
+
+      if (!hxScore && !xhScore) return session.text(".no-result")
+
+      if (hxScore > xhScore) return stringifyResult(session, hxResult, "h", options)
+
+      const xhResultCompact = xhResult.filter(seg => seg !== " ")
+      return stringifyResult(session, xhResultCompact, "x", options)
+    })
+
+  ctx.i18n.define("zh", "commands.xdi8", {
+    description: "汉字希顶互转",
+    usage:
+      "在文本前添加选项“-a”（与文本和指令名“xdi8”之间用空格隔开）来显示隐藏的结果，包括过时拼写和例外结果。加“-A”可显示例外结果，但不包含过时拼写。\n" +
+      "“例外结果”包括希转汉时的部分繁体字和汉转希时多音字的罕见读音。\n" +
+      "当文本为单个汉字或不包含空格的希顶“词”时，会默认显示所有隐藏结果。",
+    options: {
+      all: "总是显示所有隐藏结果",
+      almostAll: "总是显示例外结果（不包括过时拼写）",
+    },
+    messages: {
+      "no-result": "未找到可转换的字词。",
+    },
+  })
+}

tsconfig.json

@@ -0,0 +1,23 @@
+{
+  "compilerOptions": {
+    "rootDir": "src",
+    "outDir": "lib",
+    "target": "es2022",
+    "module": "commonjs",
+    "declaration": true,
+    "composite": true,
+    "incremental": true,
+    "skipLibCheck": true,
+    "esModuleInterop": true,
+    "moduleResolution": "node",
+    "jsx": "react-jsx",
+    "jsxImportSource": "@satorijs/element",
+    "types": [
+      "node",
+      "yml-register/types"
+    ]
+  },
+  "include": [
+    "src"
+  ]
+}