435 lines
19 KiB
TypeScript
435 lines
19 KiB
TypeScript
import type { Access, FactBlock, FunctionFacts, ImportFacts, ImportItem, Node, PropertyRow, Segment, SourceSpan, TextBlock, TokenRole, Translator, VariableBlock } from '../contract'
|
||
|
||
/** 按所在语法节点区分译法的关键字，例如 Python 的 else：if 后「否则」，for / while 后「未中断时」。 */
|
||
export interface ContextualWord {
|
||
/** 没有匹配的语境时用的译法。 */
|
||
zh: string
|
||
/**
|
||
* 语境 → 译法。语境是关键字节点往上的节点类型链，外层在前、用 ` > ` 连接：
|
||
* `raise_statement` 指父节点，`for_statement > else_clause` 指父节点 else_clause 且祖父节点 for_statement。
|
||
* 同时匹配多条时取最长的那条。
|
||
*/
|
||
byContext: Record<string, string>
|
||
}
|
||
|
||
/**
|
||
* 节点模板的一项。字符串是插入的译词，按关键字着色；另外两种是子节点槽位：
|
||
* `{ field }` 按 tree-sitter 字段名取子节点，`{ namedChild }` 按具名子节点序号（0 起，不计注释）取，给语法没定义字段名的节点用。
|
||
*/
|
||
export type TemplatePart = string | { field: string } | { namedChild: number }
|
||
|
||
/** 一种语言的中文直译表。引擎替换表里的关键字、内置类型、运算符，按节点模板重排少数语法结构，其余原样照抄。 */
|
||
export interface ZhTable {
|
||
id: string
|
||
label: string
|
||
languageIds: string[]
|
||
/** 未翻译的词按源码着色的词性上色；关键字译文的词性也取它。 */
|
||
tokenRole: (leaf: Node) => TokenRole
|
||
/** 关键字：按语法树里匿名节点的类型查（Python 的 `not in` 这类组合关键字本身也是一个匿名节点）；随所在语法节点换译法的写成 ContextualWord。 */
|
||
keywords: Record<string, string | ContextualWord>
|
||
/** 字面量关键字：具名的叶子节点（true / None …），按节点类型查。 */
|
||
literals: Record<string, string>
|
||
/** 内置类型：只在 isTypePosition 成立的叶子上按原文查。 */
|
||
builtinTypes: Record<string, string>
|
||
isTypePosition: (leaf: Node) => boolean
|
||
/** 翻成文字的运算符：只认父节点在 operatorParents 里的运算符节点（TS 的非空断言 `x!` 不算）。 */
|
||
operators: Record<string, string>
|
||
operatorParents: string[]
|
||
/** 空字面量（没有具名子节点）整个换成一个词：空数组 / 空列表 / 空字典。 */
|
||
emptyLiterals: Record<string, string>
|
||
/** 整段原样照抄、按字符串着色的节点。 */
|
||
stringNodes: string[]
|
||
/** 注释（或 docstring）节点去掉注释符号后的原文，可以多行。 */
|
||
commentText: (node: Node) => string
|
||
/**
|
||
* 节点模板：节点类型 → 按中文语序重排的译文。命中的节点只输出模板里的译词和槽位，节点自己的关键字、标点不再输出；
|
||
* 槽位里的子节点照常递归翻译。模板引用的子节点缺了，或节点没整个落在要翻的这一段里，就不重排、照常翻。
|
||
*/
|
||
nodeTemplates: Record<string, TemplatePart[]>
|
||
}
|
||
|
||
interface Token {
|
||
text: string
|
||
role: TokenRole
|
||
translated: boolean
|
||
/** 源码里的原文，用来认出逗号、括号；模板插入的译词没有原文。 */
|
||
raw: string
|
||
/** 前面要空一格：源码里它前面有空白，或者模板在它前面插了别的项。 */
|
||
gap: boolean
|
||
/** 照抄的名字、字符串在源码里的起始下标，见 Segment.offset。 */
|
||
offset?: number
|
||
}
|
||
|
||
/** 按译文顺序收集词：cursor 是上一个词在源码里的结尾，gap 是下一个词前面要不要空一格。 */
|
||
interface Collector {
|
||
source: string
|
||
tokens: Token[]
|
||
cursor: number
|
||
gap: boolean
|
||
}
|
||
|
||
const CONTEXT_SEPARATOR = ' > '
|
||
const CLOSING = new Set([')', ']', '}'])
|
||
const HEADER_OPENERS = new Set(['{', ':'])
|
||
/** 源码着色里代表名字的词性：着这几种色的具名叶子是标识符、属性名、类型名，能查定义。 */
|
||
const NAME_ROLES = new Set<TokenRole>(['variable', 'function', 'type'])
|
||
|
||
function firstLeaf(node: Node): Node {
|
||
const first = node.child(0)
|
||
return first ? firstLeaf(first) : node
|
||
}
|
||
|
||
function commentsWithin(node: Node, span: SourceSpan): Node[] {
|
||
if (node.endIndex <= span.startIndex || node.startIndex >= span.endIndex)
|
||
return []
|
||
if (node.type === 'comment')
|
||
return [node]
|
||
return node.children.flatMap(child => child ? commentsWithin(child, span) : [])
|
||
}
|
||
|
||
/** 下一个要输出的词的原文（跳过注释）。 */
|
||
function nextRaw(tokens: Token[], i: number): string {
|
||
for (let j = i + 1; j < tokens.length; j++) {
|
||
if (tokens[j].text !== '')
|
||
return tokens[j].raw
|
||
}
|
||
return ''
|
||
}
|
||
|
||
function oneLine(text: string): string {
|
||
return text.replace(/\s*\n\s*/g, ' ')
|
||
}
|
||
|
||
/** 模板槽位引用的子节点；注释是语法之外的节点，不占具名子节点序号。 */
|
||
function slotChild(node: Node, slot: Exclude<TemplatePart, string>): Node | null | undefined {
|
||
if ('field' in slot)
|
||
return node.childForFieldName(slot.field)
|
||
return node.namedChildren.filter(child => child && !child.isExtra)[slot.namedChild]
|
||
}
|
||
|
||
export function createZhTemplateTranslator(table: ZhTable): Translator {
|
||
const operatorParents = new Set(table.operatorParents)
|
||
const stringNodes = new Set(table.stringNodes)
|
||
const contextDepth = Math.max(0, ...Object.values(table.keywords).flatMap(word =>
|
||
typeof word === 'string' ? [] : Object.keys(word.byContext).map(context => context.split(CONTEXT_SEPARATOR).length)))
|
||
|
||
/** 从父节点往上逐层拼出语境去查，更长（更具体）的匹配覆盖更短的。 */
|
||
function keywordText(word: string | ContextualWord, node: Node): string {
|
||
if (typeof word === 'string')
|
||
return word
|
||
let text = word.zh
|
||
let context = ''
|
||
let ancestor = node.parent
|
||
for (let level = 0; ancestor && level < contextDepth; level++, ancestor = ancestor.parent) {
|
||
context = context ? `${ancestor.type}${CONTEXT_SEPARATOR}${context}` : ancestor.type
|
||
text = word.byContext[context] ?? text
|
||
}
|
||
return text
|
||
}
|
||
|
||
/** 照抄这个节点时带不带源码下标：名字（按名字着色的具名叶子）和整段字符串带，标点、数字、关键字不带。 */
|
||
function offsetOf(node: Node): number | undefined {
|
||
const linkable = stringNodes.has(node.type) || (node.childCount === 0 && node.isNamed && NAME_ROLES.has(table.tokenRole(node)))
|
||
return linkable ? node.startIndex : undefined
|
||
}
|
||
|
||
/** 能整个换成中文的节点 → 译文和词性。 */
|
||
function translateNode(node: Node): { text: string, role: TokenRole } | undefined {
|
||
if (node.childCount === 0 && table.isTypePosition(node)) {
|
||
const type = table.builtinTypes[node.text]
|
||
if (type)
|
||
return { text: type, role: 'type' }
|
||
}
|
||
const parent = node.parent
|
||
if (!node.isNamed) {
|
||
if (parent && operatorParents.has(parent.type) && table.operators[node.type])
|
||
return { text: table.operators[node.type], role: 'operator' }
|
||
const word = table.keywords[node.type]
|
||
if (word !== undefined)
|
||
return { text: keywordText(word, node), role: table.tokenRole(firstLeaf(node)) }
|
||
return undefined
|
||
}
|
||
if (node.childCount === 0 && table.literals[node.type])
|
||
return { text: table.literals[node.type], role: table.tokenRole(node) }
|
||
if (table.emptyLiterals[node.type] && node.namedChildCount === 0)
|
||
return { text: table.emptyLiterals[node.type], role: 'literal' }
|
||
return undefined
|
||
}
|
||
|
||
/** 节点命中模板时，模板各项换成译词或子节点；不该重排时返回 undefined。 */
|
||
function templateParts(node: Node, span: SourceSpan): (string | Node)[] | undefined {
|
||
const template = table.nodeTemplates[node.type]
|
||
if (!template || node.startIndex < span.startIndex || node.endIndex > span.endIndex)
|
||
return undefined
|
||
const parts = template.map(part => typeof part === 'string' ? part : slotChild(node, part))
|
||
return parts.every((part): part is string | Node => part != null) ? parts : undefined
|
||
}
|
||
|
||
/** 源码里上一个词到 index 之间有空白，下一个词前面就空一格。 */
|
||
function advance(out: Collector, index: number): void {
|
||
out.gap ||= /\s/.test(out.source.slice(out.cursor, index))
|
||
out.cursor = index
|
||
}
|
||
|
||
function emit(out: Collector, token: Omit<Token, 'gap'>): void {
|
||
out.tokens.push({ ...token, gap: out.gap })
|
||
out.gap = false
|
||
}
|
||
|
||
function emitSource(out: Collector, node: Node, token: Omit<Token, 'gap' | 'raw'>): void {
|
||
advance(out, node.startIndex)
|
||
emit(out, { ...token, raw: node.text })
|
||
out.cursor = node.endIndex
|
||
}
|
||
|
||
function collect(node: Node, span: SourceSpan, out: Collector): void {
|
||
if (node.endIndex <= span.startIndex || node.startIndex >= span.endIndex || node.startIndex === node.endIndex)
|
||
return
|
||
if (node.type === 'comment') {
|
||
emitSource(out, node, { text: '', role: 'comment', translated: false })
|
||
return
|
||
}
|
||
const parts = templateParts(node, span)
|
||
if (parts) {
|
||
advance(out, node.startIndex)
|
||
parts.forEach((part, i) => {
|
||
out.gap ||= i > 0
|
||
if (typeof part === 'string') {
|
||
emit(out, { text: part, role: 'keyword', translated: true, raw: '' })
|
||
return
|
||
}
|
||
out.cursor = part.startIndex
|
||
collect(part, span, out)
|
||
})
|
||
out.cursor = node.endIndex
|
||
return
|
||
}
|
||
const translated = translateNode(node)
|
||
if (translated) {
|
||
emitSource(out, node, { ...translated, translated: true })
|
||
return
|
||
}
|
||
if (node.childCount === 0 || stringNodes.has(node.type)) {
|
||
const role = stringNodes.has(node.type) ? 'string' : table.tokenRole(node)
|
||
emitSource(out, node, { text: oneLine(node.text), role, translated: false, offset: offsetOf(node) })
|
||
return
|
||
}
|
||
for (const child of node.children) {
|
||
if (child)
|
||
collect(child, span, out)
|
||
}
|
||
}
|
||
|
||
/** 按译文顺序收集 node 落在 span 里的词。 */
|
||
function tokensIn(node: Node, span: SourceSpan, source: string): Token[] {
|
||
const out: Collector = { source, tokens: [], cursor: span.startIndex, gap: false }
|
||
collect(node, span, out)
|
||
return out.tokens
|
||
}
|
||
|
||
/** node 的源码原样照抄：叶子按源码着色（给了 role 就都按它），叶子之间的空白、标点照抄。 */
|
||
function verbatim(node: Node, source: string, role?: TokenRole): Segment[] {
|
||
const segments: Segment[] = []
|
||
let cursor = node.startIndex
|
||
const visit = (current: Node): void => {
|
||
if (current.startIndex === current.endIndex)
|
||
return
|
||
if (current.childCount > 0) {
|
||
for (const child of current.children) {
|
||
if (child)
|
||
visit(child)
|
||
}
|
||
return
|
||
}
|
||
pushSegment(segments, source.slice(cursor, current.startIndex), 'plain')
|
||
pushSegment(segments, current.text, role ?? table.tokenRole(current), offsetOf(current))
|
||
cursor = current.endIndex
|
||
}
|
||
visit(node)
|
||
pushSegment(segments, source.slice(cursor, node.endIndex), 'plain')
|
||
return segments
|
||
}
|
||
|
||
function translateComment(block: TextBlock): Segment[] {
|
||
const { startIndex, endIndex } = block.header
|
||
const root = block.node.tree.rootNode
|
||
const nodes = block.node.type === 'comment' ? commentsWithin(root.descendantForIndex(startIndex, endIndex) ?? root, block.header) : [block.node]
|
||
const text = nodes.map(table.commentText).join('\n').split('\n').map(line => line.trim()).filter(Boolean).join(' ')
|
||
return text ? [{ text, role: 'comment' }] : []
|
||
}
|
||
|
||
function translate(block: TextBlock, source: string): Segment[] {
|
||
if (block.kind === 'comment')
|
||
return translateComment(block)
|
||
const tokens = tokensIn(block.node, block.header, source)
|
||
const compound = block.children.length > 0 || block.header.endIndex < block.node.endIndex
|
||
const last = tokens.at(-1)
|
||
if (compound && last && !last.translated && HEADER_OPENERS.has(last.raw))
|
||
tokens.pop()
|
||
return joinTokens(tokens)
|
||
}
|
||
|
||
function describe(block: FactBlock, source: string): PropertyRow[] {
|
||
const { header } = block
|
||
const phrases: Phrases = {
|
||
// 值里带语句块主体的回调（useMemo(() => { … })）只翻到主体开始之前，和块头部一致
|
||
translated: (node) => {
|
||
const clipped = node.endIndex > header.endIndex
|
||
const span = {
|
||
startIndex: node.startIndex,
|
||
endIndex: clipped ? header.endIndex : node.endIndex,
|
||
startRow: node.startPosition.row,
|
||
endRow: clipped ? header.endRow : node.endPosition.row,
|
||
}
|
||
return joinTokens(tokensIn(node, span, source))
|
||
},
|
||
verbatim: node => verbatim(node, source),
|
||
// 点分模块名逐节照抄，每一节各自能跳到定义，和 VS Code 里一样
|
||
module: node => stringNodes.has(node.type) ? [segment(node.text, 'string', offsetOf(node))] : verbatim(node, source, 'type'),
|
||
name: (node, role) => [segment(node.text, role, offsetOf(node))],
|
||
}
|
||
switch (block.kind) {
|
||
case 'function':
|
||
return functionRows(block.facts, phrases)
|
||
case 'import':
|
||
return importRows(block.facts, phrases)
|
||
default:
|
||
return variableRows(block, phrases)
|
||
}
|
||
}
|
||
|
||
return { id: table.id, label: table.label, languageIds: table.languageIds, translate, describe }
|
||
}
|
||
|
||
function segment(text: string, role: TokenRole, offset: number | undefined): Segment {
|
||
return offset === undefined ? { text, role } : { text, role, offset }
|
||
}
|
||
|
||
function pushSegment(segments: Segment[], text: string, role: TokenRole, offset?: number): void {
|
||
if (!text)
|
||
return
|
||
const prev = segments.at(-1)
|
||
if (role === 'plain' && prev?.role === 'plain')
|
||
prev.text += text
|
||
else
|
||
segments.push(segment(text, role, offset))
|
||
}
|
||
|
||
/** 词连成译文：源码里隔着空白的词之间空一格，注释不进译文，右括号前的尾逗号去掉。 */
|
||
function joinTokens(tokens: Token[]): Segment[] {
|
||
const segments: Segment[] = []
|
||
let gap = false
|
||
tokens.forEach((token, i) => {
|
||
gap ||= token.gap
|
||
const trailingComma = token.raw === ',' && !token.translated && CLOSING.has(nextRaw(tokens, i))
|
||
if (token.text === '' || trailingComma) {
|
||
gap ||= token.role === 'comment'
|
||
return
|
||
}
|
||
if (gap && segments.length > 0)
|
||
pushSegment(segments, ' ', 'plain')
|
||
pushSegment(segments, token.text, token.role, token.offset)
|
||
// 「非」是个词，后面补一个空格，不和操作数粘在一起
|
||
gap = token.translated && token.raw === '!'
|
||
})
|
||
return segments
|
||
}
|
||
|
||
// ─── 属性表 ───────────────────────────────────────────────────────────────────
|
||
|
||
/** 属性表里值的三种来源。 */
|
||
interface Phrases {
|
||
/** 按表翻译，只翻落在块头部里的那一段。 */
|
||
translated: (node: Node) => Segment[]
|
||
/** 源码原样照抄、按源码着色（装饰器）。 */
|
||
verbatim: (node: Node) => Segment[]
|
||
/** 导入来源：字符串字面量按字符串着色，模块名按类型着色。 */
|
||
module: (node: Node) => Segment[]
|
||
/** 源码里的一个名字原样照抄、按 role 着色（函数名称）。 */
|
||
name: (node: Node, role: TokenRole) => Segment[]
|
||
}
|
||
|
||
const ACCESS_ZH: Record<Access, string> = { public: '公共', private: '私有', protected: '受保护' }
|
||
|
||
function hint(text: string): Segment {
|
||
return { text, role: 'hint' }
|
||
}
|
||
|
||
function yesNo(value: boolean): Segment[] {
|
||
return [{ text: value ? '是' : '否', role: 'storage' }]
|
||
}
|
||
|
||
function count(n: number): Segment[] {
|
||
return [hint(n > 0 ? `${n} 个` : '无')]
|
||
}
|
||
|
||
function joined(parts: Segment[][], separator: string): Segment[] {
|
||
return parts.flatMap((part, i) => i > 0 ? [{ text: separator, role: 'plain' as const }, ...part] : part)
|
||
}
|
||
|
||
function accessRows(access: Access | undefined): PropertyRow[] {
|
||
return access ? [{ label: '访问级别', value: [{ text: ACCESS_ZH[access], role: 'keyword' }] }] : []
|
||
}
|
||
|
||
function aliased(item: Segment[], alias: Node | undefined, say: Phrases): Segment[] {
|
||
if (!alias)
|
||
return item
|
||
return [...item, { text: ' ', role: 'plain' }, { text: '作为', role: 'keyword' }, { text: ' ', role: 'plain' }, ...say.translated(alias)]
|
||
}
|
||
|
||
/** 函数：名称 / 访问级别 / 异步 / 装饰器 / 输入参数（每个参数一行：名称、类型、默认值）/ 输出参数。 */
|
||
function functionRows(facts: FunctionFacts, say: Phrases): PropertyRow[] {
|
||
const decorators = facts.decorators.map(say.verbatim)
|
||
return [
|
||
{ label: '名称', value: facts.name ? say.name(facts.name, 'function') : [hint('匿名')] },
|
||
...accessRows(facts.access),
|
||
{ label: '异步', value: yesNo(facts.async) },
|
||
...(decorators.length > 0 ? [{ label: '装饰器', value: joined(decorators, '\n') }] : []),
|
||
{
|
||
label: '输入参数',
|
||
value: count(facts.parameters.length),
|
||
items: facts.parameters.map(param => ({
|
||
cells: [
|
||
param.optional ? [...say.translated(param.name), { text: '?', role: 'plain' }] : say.translated(param.name),
|
||
param.type ? say.translated(param.type) : [hint('未标注')],
|
||
param.default ? [hint('默认 '), ...say.translated(param.default)] : [],
|
||
],
|
||
})),
|
||
},
|
||
{ label: '输出参数', value: facts.returnType ? say.translated(facts.returnType) : [hint('未标注')] },
|
||
]
|
||
}
|
||
|
||
/** 导入：来源 / 仅类型（只有分这个的语言才有）/ 导入项（每项一行，有别名写成「名称 作为 别名」）。 */
|
||
function importRows(facts: ImportFacts, say: Phrases): PropertyRow[] {
|
||
const itemText = (item: ImportItem): Segment[] => {
|
||
switch (item.kind) {
|
||
case 'named':
|
||
return aliased(say.translated(item.name), item.alias, say)
|
||
case 'default':
|
||
return aliased([hint('默认导出')], item.alias, say)
|
||
case 'namespace':
|
||
return aliased([hint('全部')], item.alias, say)
|
||
case 'module':
|
||
// 一条语句导入多个模块时（Python 的 import a, b）要写出是哪个模块
|
||
return aliased(facts.sources.length > 1 ? say.module(item.module) : [hint('整个模块')], item.alias, say)
|
||
}
|
||
}
|
||
return [
|
||
{ label: '来源', value: facts.sources.length > 0 ? joined(facts.sources.map(say.module), ', ') : [hint('无')] },
|
||
...(facts.typeOnly === undefined ? [] : [{ label: '仅类型', value: yesNo(facts.typeOnly) }]),
|
||
{ label: '导入项', value: count(facts.items.length), items: facts.items.map(item => ({ cells: [itemText(item)] })) },
|
||
]
|
||
}
|
||
|
||
/** 常量 / 变量：每个声明的变量依次是名称 / 访问级别 / 类型（有标注才有）/ 值（变量叫初始值）。 */
|
||
function variableRows(block: VariableBlock, say: Phrases): PropertyRow[] {
|
||
const valueLabel = block.kind === 'const' ? '值' : '初始值'
|
||
return block.facts.declarators.flatMap(declarator => [
|
||
{ label: '名称', value: say.translated(declarator.name) },
|
||
...accessRows(block.facts.access),
|
||
...(declarator.type ? [{ label: '类型', value: say.translated(declarator.type) }] : []),
|
||
{ label: valueLabel, value: declarator.value ? say.translated(declarator.value) : [hint('无')] },
|
||
])
|
||
}
|