feat(locale): CJK calendars, postpositional bounds, and number-word arithmetic fixes

apps/site/public/llms-small.txt

@@ -49,24 +49,33 @@ Distinct from success `alternatives`, failure `candidate`, and issue `suggestion
 import { createLingo } from "@pascal-app/lingo"
 import { es } from "@pascal-app/lingo/locales/es"
 import { fr } from "@pascal-app/lingo/locales/fr"
+import { ja } from "@pascal-app/lingo/locales/ja"
 import { zh } from "@pascal-app/lingo/locales/zh"
 
-const lingo = createLingo({ locales: [es, fr, zh] })
+const lingo = createLingo({ locales: [es, fr, ja, zh] })
 lingo.parseQuantity("dos kg") // auto-detected as es; result.locale === "es"
 lingo.parseRange("entre 5 et 10 kg", { locale: "fr" })
 lingo.parseQuantity("5公斤", { locale: "zh" })
+lingo.parseQuantity("mille cinq cents metres", { locale: "fr" }) // 1500 m
+
+// Comparators that FOLLOW the quantity (CJK); these return ranges, not quantities.
+lingo.parseRange("5キロ未満", { locale: "ja" })  // { min: null, max: 5 kg }
+lingo.parseRange("5公斤以上", { locale: "zh" })  // { min: 5 kg, max: null }
+lingo.parseQuantity("100元", { locale: "zh" })   // 100 CNY (pack currency default)
 ```
 
 Packs: `en`, `en-gb`, `es`, `fr`, `pt`, `zh`, `ja`. Omit `locale` to detect among loaded packs plus English; pass `locale` for known fields. English is built in and does not require a pack. `parseDate(text, { locale, localePacks, now })` can use loaded packs for relative date vocabulary.
 
 Pack notes:
 - `en`: built-in English quantity/date grammar; wave-1 adds spoken decimals and date idioms such as `two point five kg` and `quarter of five`.
 - `en-gb`: English plus day-first numeric dates and UK weekday offsets (`Monday week`, `Tuesday fortnight`).
-- `es`: Spanish number/range words, Romance composition, spoken decimals, approximants, and date clock/edge idioms.
-- `fr`: French number/range words, exhaustive CLDR-style vigesimal compounds, spoken decimals, approximants, and weekday-offset/clock idioms.
-- `pt`: Portuguese number/range words, compound hundreds/tens, spoken decimals, approximants, and clock/day-offset idioms.
-- `zh`: Chinese CJK number walker, scale grouping, post-unit half, duration counters, period edges, and day offsets. CJK spoken clock (`下午3点半`) is deferred.
-- `ja`: Japanese CJK number walker, scale grouping, post-unit half, duration counters, wave-dash ranges, and double-step period modifiers. CJK spoken clock (`午後3時半`) is deferred.
+- `es`: Spanish number/range words, Romance composition, spoken decimals, approximants, and date clock/edge idioms. Long scale: `mil millones` = 1e9, `billón` = 1e12. Ordinal days (`el 1º de marzo`).
+- `fr`: French number/range words, exhaustive CLDR-style vigesimal compounds, spoken decimals, approximants, and weekday-offset/clock idioms. Long scale: `milliard` = 1e9, `billion` = 1e12. Ordinal days (`le 1er mars`).
+- `pt`: Portuguese number/range words, compound hundreds/tens (`cento e vinte`, `mil e quinhentos`), spoken decimals, approximants, clock/day-offset idioms, `-feira` weekdays, and locative date fillers (`na proxima segunda-feira`).
+- `zh`: Chinese CJK number walker, scale grouping, post-unit half, duration counters, period edges, day offsets, postpositional bounds (`5公斤以上`, `5公斤以下`, `5公斤以内`), CNY default plus `元`/`块` aliases, suffix dates (`2026年3月5日`, `3月5日`), weekdays (`星期三`, `周三`), and spoken clocks with day periods (`下午3点半`, `三点一刻`, `明天下午3点`).
+- `ja`: Japanese CJK number walker, scale grouping, post-unit half, duration counters, wave-dash ranges, double-step period modifiers, postpositional bounds (`5キロ未満`, `5キロ以上`, `5キロ超`), JPY default plus `円` aliases, suffix dates (`2026年3月5日`), weekdays (`水曜日`), and spoken clocks with day periods (`午後3時半`, `明日午後3時`).
+
+Note: `以上`/`未満`-style comparators produce an open RANGE, so read them with `parseRange`/`rangeField`, not `parseQuantity`. The shared `￥`/`¥` symbol resolves per locale (CNY under `zh`, JPY under `ja`).
 
 Verified locale examples (fixed `now = 2026-07-03 14:30` local time for date examples):
 

apps/site/src/components/site/parse-playground.tsx

@@ -24,12 +24,31 @@ const localeLingo = createLingo({ locales: [es, fr, pt, zh, ja, enGb] })
 
 const EXAMPLES: Record<LocaleChoice, readonly string[]> = {
   auto: ['72 in to cm', 'dos kg', 'entre 5 et 10 kg', '5公斤', '暑い'],
-  en: ['2 ft', '72 in to cm', 'between 5 and 10 kg', '5 meters', "it's hot"],
-  es: ['dos kg', 'entre 5 y 10 kg', '72 pulgadas a cm', 'dos metros y medio', 'al menos 2 m'],
-  fr: ['deux kg', 'entre 5 et 10 kg', '72 pouces en cm', 'deux metres et demi', 'au moins 2 m'],
-  pt: ['dois kg', 'entre 5 e 10 kg', '72 polegadas em cm', 'pelo menos 2 m'],
-  zh: ['5公斤', '3米', '5到10公斤', '很热'],
-  ja: ['5キロ', '3メートル', '5から10キロ', '暑い'],
+  en: [
+    '2 ft',
+    '72 in to cm',
+    'between 5 and 10 kg',
+    'between one thousand and two thousand meters',
+    "it's hot",
+  ],
+  es: [
+    'dos kg',
+    'entre 5 y 10 kg',
+    '72 pulgadas a cm',
+    'ciento veinte kg',
+    'mil millones de kg',
+    'al menos 2 m',
+  ],
+  fr: [
+    'deux kg',
+    'entre 5 et 10 kg',
+    '72 pouces en cm',
+    'mille cinq cents metres',
+    'deux metres et demi',
+  ],
+  pt: ['dois kg', 'entre 5 e 10 kg', 'cento e vinte metros', 'mil e quinhentos metros'],
+  zh: ['5公斤', '5公斤以上', '5公斤以下', '五公斤左右', '100元', '很热'],
+  ja: ['5キロ', '5キロ未満', '5キロ以上', '5キロぐらい', '1000円', '暑い'],
   'en-gb': ['12 stone', '3 quid', 'roundabout 2 m'],
 }
 

apps/site/src/lib/code-snippets.ts

@@ -164,13 +164,19 @@ humanizeDuration(5400, { style: "natural" })     // "an hour and a half"`
 export const localeSnippet = `import { createLingo } from "@pascal-app/lingo"
 import { es } from "@pascal-app/lingo/locales/es"
 import { fr } from "@pascal-app/lingo/locales/fr"
+import { ja } from "@pascal-app/lingo/locales/ja"
 import { zh } from "@pascal-app/lingo/locales/zh"
 
-const lingo = createLingo({ locales: [es, fr, zh] })
+const lingo = createLingo({ locales: [es, fr, ja, zh] })
 
 lingo.parseQuantity("dos kg")                    // auto-detected as es
 lingo.parseRange("entre 5 et 10 kg", { locale: "fr" })
 lingo.parseQuantity("5公斤", { locale: "zh" })
+lingo.parseQuantity("mille cinq cents metres", { locale: "fr" })  // 1500 m
+
+// Comparators that follow the quantity, as CJK writes them.
+lingo.parseRange("5キロ未満", { locale: "ja" })    // < 5 kg
+lingo.parseRange("5公斤以上", { locale: "zh" })    // >= 5 kg
 
 // Success results expose the resolved profile.
 lingo.parse("dos kg").locale                      // "es"`

apps/site/src/lib/docs.md.ts

@@ -318,7 +318,9 @@ lingo.parse('72 in to cm') // locale: 'en'`,
   '',
   'Published packs: `@pascal-app/lingo/locales/en`, `en-gb`, `es`, `fr`, `pt`, `zh`, and `ja`. Pack-owned CJK unit aliases and fuzzy words are only registered when their pack is loaded. Passing an explicit locale that was not loaded returns `LOCALE_NOT_LOADED` instead of silently falling back to English.',
   '',
-  'Date parsing accepts the same packs directly: `parseDate("mañana", { now, locale: "es", localePacks: [es] })`.',
+  'Each pack reads its language the way people write it, not a transliteration of English. Romance packs cover tens/hundreds/scale words (`ciento veinte`, `mille cinq cents`, `mil millones`), ordinal days (`le 1er mars`), and spoken decimals (`dos coma cinco`). The Chinese and Japanese packs cover grouped numerals (`一百五十万`), comparators that follow the quantity (`5キロ未満`, `5公斤以上`), trailing approximants (`五公斤左右`), and per-locale currency defaults — the shared `￥` resolves to CNY under `zh` and JPY under `ja`.',
+  '',
+  'Date parsing accepts the same packs directly: `parseDate("mañana", { now, locale: "es", localePacks: [es] })`. Localized date grammar includes spoken clocks (`las tres menos cuarto`, `deux heures et quart`), period edges (`a finales de mes`), weekday offsets (`lundi en huit`), and CJK calendar forms — suffix dates (`2026年3月5日`, `3月5日`), weekdays (`星期三`, `水曜日`), clocks with day periods (`午後3時半`, `下午3点`), and date+time written without a space (`明天下午3点`).',
   '',
   '## Catalog',
   '',

packages/lingo/CHANGELOG.md

@@ -7,8 +7,56 @@ change**, even if the API is untouched.
 
 ## [Unreleased]
 
+### Added
+
+- Chinese and Japanese calendar and clock grammar: numeric dates written with
+  suffixes (`2026年3月5日`, `3月5日`, and years spelled digit-by-digit as
+  `二〇二六年`), weekdays (`星期三`, `周三`, `水曜日`), clocks closed by
+  `点`/`時`/`分`/`秒` with day periods (`下午3点`, `午前9時半`, `3点一刻`), and
+  date+time compounds written with no separating space (`明天下午3点`,
+  `明日午後3時`).
+- Postpositional bound phrases for scripts that put the comparator after the
+  quantity: `5キロ未満`, `5キロ以上`, `5公斤以下`, `5公斤以内`, `5キロ超`.
+  Previously `5キロ未満` parsed as a bare `5 kg`, dropping the bound.
+- Chinese and Japanese default currencies, so the shared `¥`/`￥` symbol
+  resolves per locale (`￥100` is CNY under `zh`, JPY under `ja`), plus the
+  `元`/`圆`/`块`/`人民币` and `円`/`えん` aliases.
+- French long-scale number words (`milliard` = 10^9, `billion` = 10^12),
+  Spanish `mil millones`/`billón`, Portuguese `cento`, and Romance ordinal
+  day-of-month forms (`le 1er mars`, `el 1º de marzo`).
+- A per-locale benchmark suite (`bun run bench`) so multi-language throughput is
+  measured rather than assumed.
+- Per-locale corpus contracts gained 96 rows and the English contract 5,
+  covering the grammar above plus the number-word fixes below.
+
+### Fixed
+
+- Number words: hundreds now multiply only the 1..99 group in front of them, so
+  French `mille cinq cents` is 1500 (was 100500). A banked smaller scale
+  multiplies the next one, so Spanish `mil millones` is 10^9 (was 1,001,000) and
+  `dos mil millones` is 2×10^9. Both were silent wrong answers.
+- `between A and B` now keeps the and-word for the range when both sides are
+  spelled scale words: `between one thousand and two thousand meters`,
+  `entre mille et deux mille`, `entre mil y dos mil` (all previously failed to
+  parse). `between five and a half and ten kg` still reads 5.5..10 — the
+  fraction tail binds tighter than the range separator.
+- Portuguese `cento e vinte` (120) and `mil e quinhentos` (1500) parse; the
+  and-word after a bare scale word links it to its remainder.
+- Portuguese locative contractions are date fillers, so
+  `na proxima segunda-feira` resolves.
+- Scripts written without word spaces no longer swallow grammar words glued to a
+  quantity (`三至五天`, `5公斤左右`). Unit aliases stay atomic through the split,
+  so `一時間半` remains 1.5 hours rather than splitting at the range word `間`.
+
 ### Changed
 
+- Resolved language profiles and locale-detection scans are memoized per pack
+  set, so repeated parses with `locale`/auto-detection no longer re-merge packs
+  or re-tokenize for detection on every call.
+- Size budgets recalibrated once for this wave (see D70 in `wiki/decisions.md`
+  for the measured numbers and rationale); `scripts/size.mjs` remains the source
+  of truth. No previously-pinned corpus interpretation changed — the wave is
+  additive.
 - README (npm): repo-relative links (`docs/recipes.md`, `plans/`, `wiki/`) now
   point at absolute GitHub URLs so they resolve on npmjs.com, and the docs
   site plus the agent `llms.txt` index are linked from the top.

packages/lingo/README.md

@@ -559,24 +559,34 @@ spanish.ok && spanish.quantity.to('kg').value   // 35
 const chinese = app.parseQuantity('三十五公斤', { locale: 'zh' })
 chinese.ok && chinese.quantity.to('kg').value   // 35
 
+// comparators that follow the quantity, as CJK writes them
+app.parse('5キロ未満', { locale: 'ja' })          // < 5 kg  (open range)
+app.parse('5公斤以上', { locale: 'zh' })          // ≥ 5 kg
+app.parse('100元', { locale: 'zh' })             // 100 CNY
+
 const clock = parseDate('las tres menos cuarto', {
   locale: 'es',
   localePacks: [es],
   now: new Date(2026, 6, 3, 14, 30),
 })
 clock.ok && [clock.date.getHours(), clock.date.getMinutes()] // [2, 45]
 
+const cjkDate = parseDate('2026年3月5日', { locale: 'zh', localePacks: [zh] })
+cjkDate.ok && [cjkDate.date.getMonth() + 1, cjkDate.date.getDate()] // [3, 5] (local)
+
 // omit `locale` to auto-detect among the loaded packs
 ```
 
 Packs (`en`, `en-gb`, `es`, `fr`, `pt`, `zh`, `ja`) are additive and tree-shakeable;
 they add number words, units, ranges, and relative dates (through
-`@pascal-app/lingo/date`) for their language. The first idiom wave includes
-Romance tens/hundreds/decimal words, CJK scale/grouped numbers and post-unit
-half, localized clock phrases, period edges, weekday offsets, and locale unit
-words in date/duration parsing. Requesting an unloaded locale returns
-`LOCALE_NOT_LOADED` rather than silently parsing as English; `humanizeDate()` output
-stays English for now.
+`@pascal-app/lingo/date`) for their language. Coverage includes Romance
+tens/hundreds/scale words (`ciento veinte`, `mille cinq cents`, `mil millones`),
+CJK scale/grouped numbers and post-unit half, postpositional comparators
+(`5キロ未満`, `5公斤以上`), per-locale currency defaults (`￥` is CNY under `zh`,
+JPY under `ja`), localized clock phrases and period edges, and CJK calendar
+dates and clocks (`2026年3月5日`, `午後3時半`, `明天下午3点`). Requesting an
+unloaded locale returns `LOCALE_NOT_LOADED` rather than silently parsing as
+English; `humanizeDate()` output stays English for now.
 
 The global `lingo()` is itself a `createLingo()` singleton. One code path,
 two tiers of convenience.

packages/lingo/llms.txt

@@ -49,24 +49,33 @@ Distinct from success `alternatives`, failure `candidate`, and issue `suggestion
 import { createLingo } from "@pascal-app/lingo"
 import { es } from "@pascal-app/lingo/locales/es"
 import { fr } from "@pascal-app/lingo/locales/fr"
+import { ja } from "@pascal-app/lingo/locales/ja"
 import { zh } from "@pascal-app/lingo/locales/zh"
 
-const lingo = createLingo({ locales: [es, fr, zh] })
+const lingo = createLingo({ locales: [es, fr, ja, zh] })
 lingo.parseQuantity("dos kg") // auto-detected as es; result.locale === "es"
 lingo.parseRange("entre 5 et 10 kg", { locale: "fr" })
 lingo.parseQuantity("5公斤", { locale: "zh" })
+lingo.parseQuantity("mille cinq cents metres", { locale: "fr" }) // 1500 m
+
+// Comparators that FOLLOW the quantity (CJK); these return ranges, not quantities.
+lingo.parseRange("5キロ未満", { locale: "ja" })  // { min: null, max: 5 kg }
+lingo.parseRange("5公斤以上", { locale: "zh" })  // { min: 5 kg, max: null }
+lingo.parseQuantity("100元", { locale: "zh" })   // 100 CNY (pack currency default)
 ```
 
 Packs: `en`, `en-gb`, `es`, `fr`, `pt`, `zh`, `ja`. Omit `locale` to detect among loaded packs plus English; pass `locale` for known fields. English is built in and does not require a pack. `parseDate(text, { locale, localePacks, now })` can use loaded packs for relative date vocabulary.
 
 Pack notes:
 - `en`: built-in English quantity/date grammar; wave-1 adds spoken decimals and date idioms such as `two point five kg` and `quarter of five`.
 - `en-gb`: English plus day-first numeric dates and UK weekday offsets (`Monday week`, `Tuesday fortnight`).
-- `es`: Spanish number/range words, Romance composition, spoken decimals, approximants, and date clock/edge idioms.
-- `fr`: French number/range words, exhaustive CLDR-style vigesimal compounds, spoken decimals, approximants, and weekday-offset/clock idioms.
-- `pt`: Portuguese number/range words, compound hundreds/tens, spoken decimals, approximants, and clock/day-offset idioms.
-- `zh`: Chinese CJK number walker, scale grouping, post-unit half, duration counters, period edges, and day offsets. CJK spoken clock (`下午3点半`) is deferred.
-- `ja`: Japanese CJK number walker, scale grouping, post-unit half, duration counters, wave-dash ranges, and double-step period modifiers. CJK spoken clock (`午後3時半`) is deferred.
+- `es`: Spanish number/range words, Romance composition, spoken decimals, approximants, and date clock/edge idioms. Long scale: `mil millones` = 1e9, `billón` = 1e12. Ordinal days (`el 1º de marzo`).
+- `fr`: French number/range words, exhaustive CLDR-style vigesimal compounds, spoken decimals, approximants, and weekday-offset/clock idioms. Long scale: `milliard` = 1e9, `billion` = 1e12. Ordinal days (`le 1er mars`).
+- `pt`: Portuguese number/range words, compound hundreds/tens (`cento e vinte`, `mil e quinhentos`), spoken decimals, approximants, clock/day-offset idioms, `-feira` weekdays, and locative date fillers (`na proxima segunda-feira`).
+- `zh`: Chinese CJK number walker, scale grouping, post-unit half, duration counters, period edges, day offsets, postpositional bounds (`5公斤以上`, `5公斤以下`, `5公斤以内`), CNY default plus `元`/`块` aliases, suffix dates (`2026年3月5日`, `3月5日`), weekdays (`星期三`, `周三`), and spoken clocks with day periods (`下午3点半`, `三点一刻`, `明天下午3点`).
+- `ja`: Japanese CJK number walker, scale grouping, post-unit half, duration counters, wave-dash ranges, double-step period modifiers, postpositional bounds (`5キロ未満`, `5キロ以上`, `5キロ超`), JPY default plus `円` aliases, suffix dates (`2026年3月5日`), weekdays (`水曜日`), and spoken clocks with day periods (`午後3時半`, `明日午後3時`).
+
+Note: `以上`/`未満`-style comparators produce an open RANGE, so read them with `parseRange`/`rangeField`, not `parseQuantity`. The shared `￥`/`¥` symbol resolves per locale (CNY under `zh`, JPY under `ja`).
 
 Verified locale examples (fixed `now = 2026-07-03 14:30` local time for date examples):