dsh-cait 1.0.3 → 1.0.4

package/lib/tm.js

@@ -1,160 +1,160 @@
-import { CatError } from './errors.js'
-import { DEFAULT_TM_ID, MAX_SEGMENTS, canonicalLanguage, cleanLang, clamp, clampInt, makeId, now } from './store-utils.js'
-import { classifyMatchType, normalizeSource, normalizeText, similarity } from './similarity.js'
-
-export const tmMethods = {
-  mountedTmCount() {
-    const mounted = new Set(this.mounts.tmIds)
-    return this.tm.filter((entry) => mounted.has(entry.tmId || DEFAULT_TM_ID)).length
-  }
-,
-  upsertTM(source, target, sourceLang, targetLang, tmId) {
-    const cleanSource = normalizeSource(source)
-    const cleanTarget = String(target ?? '').trim()
-    if (!cleanSource || !cleanTarget) {
-      throw new CatError('EMPTY_TM_ENTRY', '记忆库条目不能为空。')
-    }
-    const libraryId = this.requireTmLibrary(tmId)
-    const key = normalizeText(cleanSource)
-    const existing = this.tm.find(
-      (entry) =>
-        (entry.tmId || DEFAULT_TM_ID) === libraryId &&
-        entry.key === key &&
-        entry.sourceLang === sourceLang &&
-        entry.targetLang === targetLang,
-    )
-    if (existing) {
-      existing.source = cleanSource
-      existing.target = cleanTarget
-      existing.updatedAt = now()
-      return { entry: existing, created: false }
-    }
-    const entry = {
-      id: makeId('tm'),
-      tmId: libraryId,
-      key,
-      source: cleanSource,
-      target: cleanTarget,
-      sourceLang,
-      targetLang,
-      createdAt: now(),
-      updatedAt: now(),
-    }
-    this.tm.push(entry)
-    return { entry, created: true }
-  }
-,
-  lookupTM({ query, threshold, limit, sourceLang, targetLang } = {}) {
-    const text = normalizeText(query)
-    const defaultThreshold = this.settings.fuzzyThreshold
-    const scoreFloor = clamp(Number(threshold ?? defaultThreshold), 0, 1)
-    const max = clampInt(limit ?? this.settings.tmLimit, 1, 30, 5)
-    const src = canonicalLanguage(sourceLang || this.settings.sourceLang)
-    const tgt = canonicalLanguage(targetLang || this.settings.targetLang)
-
-    if (!text) return { query: String(query ?? ''), matches: [] }
-
-    const mountedTm = new Set(this.mounts.tmIds)
-    const matches = this.tm
-      .filter(
-        (entry) =>
-          mountedTm.has(entry.tmId || DEFAULT_TM_ID) &&
-          entry.sourceLang === src &&
-          entry.targetLang === tgt,
-      )
-      .map((entry) => {
-        const exact = entry.key === text
-        const score = exact ? 1 : similarity(text, entry.key)
-        return { entry, exact, score }
-      })
-      .filter((item) => item.score >= scoreFloor)
-      .sort((a, b) => b.score - a.score || b.entry.updatedAt.localeCompare(a.entry.updatedAt))
-      .slice(0, max)
-      .map(({ entry, exact, score }) => ({
-        id: entry.id,
-        tmId: entry.tmId || DEFAULT_TM_ID,
-        source: entry.source,
-        target: entry.target,
-        sourceLang: entry.sourceLang,
-        targetLang: entry.targetLang,
-        score,
-        percent: Math.round(score * 100),
-        exact,
-        matchType: classifyMatchType({ exact, score, query: text, source: entry.key }),
-        updatedAt: entry.updatedAt,
-      }))
-
-    return { query: String(query ?? ''), matches }
-  }
-,
-  addTM({ source, target, sourceLang, targetLang, tmId } = {}) {
-    const src = cleanLang(sourceLang || this.settings.sourceLang, this.settings.sourceLang)
-    const tgt = cleanLang(targetLang || this.settings.targetLang, this.settings.targetLang)
-    const result = this.upsertTM(source, target, src, tgt, tmId)
-    this.touch()
-    this.persist()
-    return {
-      entry: {
-        id: result.entry.id,
-        tmId: result.entry.tmId,
-        source: result.entry.source,
-        target: result.entry.target,
-        sourceLang: result.entry.sourceLang,
-        targetLang: result.entry.targetLang,
-      },
-      created: result.created,
-      count: this.tm.length,
-    }
-  }
-,
-  importTM({ pairs, sourceLang, targetLang, tmId } = {}) {
-    if (!Array.isArray(pairs)) throw new CatError('INVALID_TM_IMPORT', '记忆库导入数据格式不正确。')
-    if (pairs.length > MAX_SEGMENTS) throw new CatError('TOO_MANY_TM', `一次最多导入 ${MAX_SEGMENTS} 条记忆。`)
-
-    const src = cleanLang(sourceLang || this.settings.sourceLang, this.settings.sourceLang)
-    const tgt = cleanLang(targetLang || this.settings.targetLang, this.settings.targetLang)
-    const libraryId = this.requireTmLibrary(tmId)
-    let added = 0
-    let updated = 0
-
-    for (const pair of pairs) {
-      if (!pair || !String(pair.source ?? '').trim() || !String(pair.target ?? '').trim()) continue
-      const result = this.upsertTM(
-        pair.source,
-        pair.target,
-        pair.sourceLang || src,
-        pair.targetLang || tgt,
-        libraryId,
-      )
-      if (result.created) added += 1
-      else updated += 1
-    }
-
-    if (added || updated) this.touch()
-    this.persist()
-    return { added, updated, total: this.tm.length }
-  }
-,
-  clearTM({ tmIds, all = false } = {}) {
-    let targets
-    let scope
-    if (all === true) {
-      targets = new Set(this.tmLibraries.map((library) => library.tmId))
-      scope = 'all'
-    } else if (Array.isArray(tmIds) && tmIds.length) {
-      targets = new Set(tmIds.map((id) => this.requireMountedTmLibrary(id)))
-      scope = 'explicit'
-    } else {
-      targets = new Set(this.mounts.tmIds)
-      scope = 'mounted'
-    }
-
-    const before = this.tm.length
-    this.tm = this.tm.filter((entry) => !targets.has(entry.tmId || DEFAULT_TM_ID))
-    const removed = before - this.tm.length
-    this.touch()
-    this.persist()
-    return { removed, scope, tmIds: [...targets] }
-  }
-
-}
+import { CatError } from './errors.js'
+import { DEFAULT_TM_ID, MAX_SEGMENTS, canonicalLanguage, cleanLang, clamp, clampInt, makeId, now } from './store-utils.js'
+import { classifyMatchType, normalizeSource, normalizeText, similarity } from './similarity.js'
+
+export const tmMethods = {
+  mountedTmCount() {
+    const mounted = new Set(this.mounts.tmIds)
+    return this.tm.filter((entry) => mounted.has(entry.tmId || DEFAULT_TM_ID)).length
+  }
+,
+  upsertTM(source, target, sourceLang, targetLang, tmId) {
+    const cleanSource = normalizeSource(source)
+    const cleanTarget = String(target ?? '').trim()
+    if (!cleanSource || !cleanTarget) {
+      throw new CatError('EMPTY_TM_ENTRY', '记忆库条目不能为空。')
+    }
+    const libraryId = this.requireTmLibrary(tmId)
+    const key = normalizeText(cleanSource)
+    const existing = this.tm.find(
+      (entry) =>
+        (entry.tmId || DEFAULT_TM_ID) === libraryId &&
+        entry.key === key &&
+        entry.sourceLang === sourceLang &&
+        entry.targetLang === targetLang,
+    )
+    if (existing) {
+      existing.source = cleanSource
+      existing.target = cleanTarget
+      existing.updatedAt = now()
+      return { entry: existing, created: false }
+    }
+    const entry = {
+      id: makeId('tm'),
+      tmId: libraryId,
+      key,
+      source: cleanSource,
+      target: cleanTarget,
+      sourceLang,
+      targetLang,
+      createdAt: now(),
+      updatedAt: now(),
+    }
+    this.tm.push(entry)
+    return { entry, created: true }
+  }
+,
+  lookupTM({ query, threshold, limit, sourceLang, targetLang } = {}) {
+    const text = normalizeText(query)
+    const defaultThreshold = this.settings.fuzzyThreshold
+    const scoreFloor = clamp(Number(threshold ?? defaultThreshold), 0, 1)
+    const max = clampInt(limit ?? this.settings.tmLimit, 1, 30, 5)
+    const src = canonicalLanguage(sourceLang || this.settings.sourceLang)
+    const tgt = canonicalLanguage(targetLang || this.settings.targetLang)
+
+    if (!text) return { query: String(query ?? ''), matches: [] }
+
+    const mountedTm = new Set(this.mounts.tmIds)
+    const matches = this.tm
+      .filter(
+        (entry) =>
+          mountedTm.has(entry.tmId || DEFAULT_TM_ID) &&
+          entry.sourceLang === src &&
+          entry.targetLang === tgt,
+      )
+      .map((entry) => {
+        const exact = entry.key === text
+        const score = exact ? 1 : similarity(text, entry.key)
+        return { entry, exact, score }
+      })
+      .filter((item) => item.score >= scoreFloor)
+      .sort((a, b) => b.score - a.score || b.entry.updatedAt.localeCompare(a.entry.updatedAt))
+      .slice(0, max)
+      .map(({ entry, exact, score }) => ({
+        id: entry.id,
+        tmId: entry.tmId || DEFAULT_TM_ID,
+        source: entry.source,
+        target: entry.target,
+        sourceLang: entry.sourceLang,
+        targetLang: entry.targetLang,
+        score,
+        percent: Math.round(score * 100),
+        exact,
+        matchType: classifyMatchType({ exact, score, query: text, source: entry.key }),
+        updatedAt: entry.updatedAt,
+      }))
+
+    return { query: String(query ?? ''), matches }
+  }
+,
+  addTM({ source, target, sourceLang, targetLang, tmId } = {}) {
+    const src = cleanLang(sourceLang || this.settings.sourceLang, this.settings.sourceLang)
+    const tgt = cleanLang(targetLang || this.settings.targetLang, this.settings.targetLang)
+    const result = this.upsertTM(source, target, src, tgt, tmId)
+    this.touch()
+    this.persist()
+    return {
+      entry: {
+        id: result.entry.id,
+        tmId: result.entry.tmId,
+        source: result.entry.source,
+        target: result.entry.target,
+        sourceLang: result.entry.sourceLang,
+        targetLang: result.entry.targetLang,
+      },
+      created: result.created,
+      count: this.tm.length,
+    }
+  }
+,
+  importTM({ pairs, sourceLang, targetLang, tmId } = {}) {
+    if (!Array.isArray(pairs)) throw new CatError('INVALID_TM_IMPORT', '记忆库导入数据格式不正确。')
+    if (pairs.length > MAX_SEGMENTS) throw new CatError('TOO_MANY_TM', `一次最多导入 ${MAX_SEGMENTS} 条记忆。`)
+
+    const src = cleanLang(sourceLang || this.settings.sourceLang, this.settings.sourceLang)
+    const tgt = cleanLang(targetLang || this.settings.targetLang, this.settings.targetLang)
+    const libraryId = this.requireTmLibrary(tmId)
+    let added = 0
+    let updated = 0
+
+    for (const pair of pairs) {
+      if (!pair || !String(pair.source ?? '').trim() || !String(pair.target ?? '').trim()) continue
+      const result = this.upsertTM(
+        pair.source,
+        pair.target,
+        pair.sourceLang || src,
+        pair.targetLang || tgt,
+        libraryId,
+      )
+      if (result.created) added += 1
+      else updated += 1
+    }
+
+    if (added || updated) this.touch()
+    this.persist()
+    return { added, updated, total: this.tm.length }
+  }
+,
+  clearTM({ tmIds, all = false } = {}) {
+    let targets
+    let scope
+    if (all === true) {
+      targets = new Set(this.tmLibraries.map((library) => library.tmId))
+      scope = 'all'
+    } else if (Array.isArray(tmIds) && tmIds.length) {
+      targets = new Set(tmIds.map((id) => this.requireMountedTmLibrary(id)))
+      scope = 'explicit'
+    } else {
+      targets = new Set(this.mounts.tmIds)
+      scope = 'mounted'
+    }
+
+    const before = this.tm.length
+    this.tm = this.tm.filter((entry) => !targets.has(entry.tmId || DEFAULT_TM_ID))
+    const removed = before - this.tm.length
+    this.touch()
+    this.persist()
+    return { removed, scope, tmIds: [...targets] }
+  }
+
+}

package/lib/tmx.js

@@ -1,85 +1,85 @@
-/** TMX 1.4 subset: one `<tu>` per translation unit, one `<tuv>` per language. */
-
-import { decodeXml, escapeXml } from './xml.js'
-import { CAIT_VERSION } from './store-utils.js'
-
-function attr(attrs, name) {
-  const match = String(attrs ?? '').match(
-    new RegExp(`(?:^|\\s)(?:xml:)?${name}\\s*=\\s*("([^"]*)"|'([^']*)')`, 'i'),
-  )
-  return match ? decodeXml(match[2] ?? match[3] ?? '') : ''
-}
-
-function segText(raw) {
-  const cdata = String(raw ?? '').replace(/<!\[CDATA\[([\s\S]*?)\]\]>/g, '$1')
-  return decodeXml(cdata.replace(/<[^>]*>/g, ''))
-    .replace(/\s+/g, ' ')
-    .trim()
-}
-
-export function parseTmx(xml) {
-  const text = String(xml ?? '')
-  const header = text.match(/<header\b([^>]*)>/i)
-  const units = []
-  let inlineMarkup = 0
-  const tuPattern = /<tu\b[^>]*>([\s\S]*?)<\/tu>/gi
-  let tu
-
-  while ((tu = tuPattern.exec(text)) !== null) {
-    const variants = []
-    const tuvPattern = /<tuv\b([^>]*)>([\s\S]*?)<\/tuv>/gi
-    let tuv
-    while ((tuv = tuvPattern.exec(tu[1])) !== null) {
-      const seg = tuv[2].match(/<seg\b[^>]*>([\s\S]*?)<\/seg>/i)
-      if (!seg) continue
-      // `<bpt>/<ept>/<ph>` encode inline formatting and placeholders in TMX.
-      // This subset intentionally imports plain text, so count the opening tags
-      // and warn instead of silently pretending the formatting survived.
-      inlineMarkup += (seg[1].match(/<(?:bpt|ept|ph)\b[^>]*>/gi) || []).length
-      const lang = attr(tuv[1], 'lang')
-      const value = segText(seg[1])
-      if (value) variants.push({ lang: lang || '', text: value })
-    }
-    if (variants.length) units.push(variants)
-  }
-
-  const sourceLang = attr(header?.[1], 'srclang') || ''
-  const languages = [...new Set(units.flat().map((variant) => variant.lang).filter(Boolean))]
-  const warnings = inlineMarkup
-    ? [
-        {
-          code: 'TMX_INLINE_MARKUP_LOST',
-          count: inlineMarkup,
-          message: `TMX 含 ${inlineMarkup} 个内联占位符/格式标签（bpt/ept/ph），导入时已按纯文本处理，格式信息丢失。`,
-        },
-      ]
-    : []
-
-  return {
-    sourceLang,
-    targetLang: languages.find((lang) => lang !== sourceLang) || '',
-    units,
-    warnings,
-  }
-}
-
-export function buildTmx(doc) {
-  const lines = [
-    '<?xml version="1.0" encoding="UTF-8"?>',
-    '<tmx version="1.4">',
-    `  <header creationtool="CAIT" creationtoolversion="${CAIT_VERSION}" datatype="plaintext" segtype="sentence" adminlang="en" srclang="${escapeXml(doc.sourceLang)}" o-tmf="CAIT" creationdate="${new Date().toISOString()}"/>`,
-    '  <body>',
-  ]
-
-  for (const segment of doc.segments) {
-    lines.push('    <tu>')
-    lines.push(`      <tuv xml:lang="${escapeXml(doc.sourceLang)}"><seg>${escapeXml(segment.source)}</seg></tuv>`)
-    if (segment.target) {
-      lines.push(`      <tuv xml:lang="${escapeXml(doc.targetLang)}"><seg>${escapeXml(segment.target)}</seg></tuv>`)
-    }
-    lines.push('    </tu>')
-  }
-
-  lines.push('  </body>', '</tmx>', '')
-  return lines.join('\n')
-}
+/** TMX 1.4 subset: one `<tu>` per translation unit, one `<tuv>` per language. */
+
+import { decodeXml, escapeXml } from './xml.js'
+import { CAIT_VERSION } from './store-utils.js'
+
+function attr(attrs, name) {
+  const match = String(attrs ?? '').match(
+    new RegExp(`(?:^|\\s)(?:xml:)?${name}\\s*=\\s*("([^"]*)"|'([^']*)')`, 'i'),
+  )
+  return match ? decodeXml(match[2] ?? match[3] ?? '') : ''
+}
+
+function segText(raw) {
+  const cdata = String(raw ?? '').replace(/<!\[CDATA\[([\s\S]*?)\]\]>/g, '$1')
+  return decodeXml(cdata.replace(/<[^>]*>/g, ''))
+    .replace(/\s+/g, ' ')
+    .trim()
+}
+
+export function parseTmx(xml) {
+  const text = String(xml ?? '')
+  const header = text.match(/<header\b([^>]*)>/i)
+  const units = []
+  let inlineMarkup = 0
+  const tuPattern = /<tu\b[^>]*>([\s\S]*?)<\/tu>/gi
+  let tu
+
+  while ((tu = tuPattern.exec(text)) !== null) {
+    const variants = []
+    const tuvPattern = /<tuv\b([^>]*)>([\s\S]*?)<\/tuv>/gi
+    let tuv
+    while ((tuv = tuvPattern.exec(tu[1])) !== null) {
+      const seg = tuv[2].match(/<seg\b[^>]*>([\s\S]*?)<\/seg>/i)
+      if (!seg) continue
+      // `<bpt>/<ept>/<ph>` encode inline formatting and placeholders in TMX.
+      // This subset intentionally imports plain text, so count the opening tags
+      // and warn instead of silently pretending the formatting survived.
+      inlineMarkup += (seg[1].match(/<(?:bpt|ept|ph)\b[^>]*>/gi) || []).length
+      const lang = attr(tuv[1], 'lang')
+      const value = segText(seg[1])
+      if (value) variants.push({ lang: lang || '', text: value })
+    }
+    if (variants.length) units.push(variants)
+  }
+
+  const sourceLang = attr(header?.[1], 'srclang') || ''
+  const languages = [...new Set(units.flat().map((variant) => variant.lang).filter(Boolean))]
+  const warnings = inlineMarkup
+    ? [
+        {
+          code: 'TMX_INLINE_MARKUP_LOST',
+          count: inlineMarkup,
+          message: `TMX 含 ${inlineMarkup} 个内联占位符/格式标签（bpt/ept/ph），导入时已按纯文本处理，格式信息丢失。`,
+        },
+      ]
+    : []
+
+  return {
+    sourceLang,
+    targetLang: languages.find((lang) => lang !== sourceLang) || '',
+    units,
+    warnings,
+  }
+}
+
+export function buildTmx(doc) {
+  const lines = [
+    '<?xml version="1.0" encoding="UTF-8"?>',
+    '<tmx version="1.4">',
+    `  <header creationtool="CAIT" creationtoolversion="${CAIT_VERSION}" datatype="plaintext" segtype="sentence" adminlang="en" srclang="${escapeXml(doc.sourceLang)}" o-tmf="CAIT" creationdate="${new Date().toISOString()}"/>`,
+    '  <body>',
+  ]
+
+  for (const segment of doc.segments) {
+    lines.push('    <tu>')
+    lines.push(`      <tuv xml:lang="${escapeXml(doc.sourceLang)}"><seg>${escapeXml(segment.source)}</seg></tuv>`)
+    if (segment.target) {
+      lines.push(`      <tuv xml:lang="${escapeXml(doc.targetLang)}"><seg>${escapeXml(segment.target)}</seg></tuv>`)
+    }
+    lines.push('    </tu>')
+  }
+
+  lines.push('  </body>', '</tmx>', '')
+  return lines.join('\n')
+}