@noy-db/hub 0.2.0-pre.18 → 0.2.0-pre.19

package/dist/aggregate/index.cjs

@@ -342,6 +342,18 @@ var GroupCardinalityError = class extends NoydbError {
     this.maxGroups = maxGroups;
   }
 };
+var LocaleNotSpecifiedError = class extends NoydbError {
+  /** The field name that required a locale. */
+  field;
+  constructor(field, message) {
+    super(
+      "LOCALE_NOT_SPECIFIED",
+      message ?? `Cannot read i18nText field "${field}" without a locale. Pass { locale } to get()/list()/query() or set a default via openVault(name, { locale }).`
+    );
+    this.name = "LocaleNotSpecifiedError";
+    this.field = field;
+  }
+};
 
 // src/money/descriptor.ts
 var MoneyUnsupportedError = class extends NoydbError {
@@ -568,6 +580,198 @@ function wrapMoneyReducers(spec, moneyFields) {
   return changed ? out : spec;
 }
 
+// src/i18n/policy.ts
+function resolvePolicy(onMissing, layer) {
+  const explicit = onMissing && typeof onMissing === "object" ? onMissing[layer] : void 0;
+  const scalar = typeof onMissing === "string" ? onMissing : void 0;
+  const layerDefault = layer === "guard" ? "substitute" : void 0;
+  return explicit ?? layerDefault ?? scalar ?? "throw";
+}
+
+// src/i18n/script.ts
+var LATIN_BASE = /* @__PURE__ */ new Set([
+  "en",
+  "fr",
+  "de",
+  "es",
+  "it",
+  "pt",
+  "nl",
+  "sv",
+  "no",
+  "da",
+  "fi",
+  "is",
+  "pl",
+  "cs",
+  "sk",
+  "hu",
+  "ro",
+  "hr",
+  "sl",
+  "et",
+  "lv",
+  "lt",
+  "tr",
+  "vi",
+  "id",
+  "ms",
+  "tl",
+  "sw",
+  "af",
+  "ca",
+  "gl",
+  "eu",
+  "cy",
+  "ga"
+]);
+var SCRIPT_TABLE = {
+  th: ["Thai"],
+  ko: ["Hangul", "Han"],
+  ja: ["Han", "Hiragana", "Katakana"],
+  zh: ["Han"],
+  ar: ["Arabic"],
+  fa: ["Arabic"],
+  ur: ["Arabic"],
+  ru: ["Cyrillic"],
+  uk: ["Cyrillic"],
+  bg: ["Cyrillic"],
+  sr: ["Cyrillic"],
+  he: ["Hebrew"],
+  el: ["Greek"],
+  hi: ["Devanagari"],
+  ta: ["Tamil"],
+  km: ["Khmer"],
+  lo: ["Lao"],
+  my: ["Myanmar"]
+};
+var SUBTAG_SCRIPTS = {
+  Latn: ["Latin"],
+  Cyrl: ["Cyrillic", "Latin"],
+  Hans: ["Han", "Latin"],
+  Hant: ["Han", "Latin"],
+  Thai: ["Thai", "Latin"],
+  Arab: ["Arabic", "Latin"]
+};
+function inferScripts(locale) {
+  const parts = locale.split("-");
+  const subtag = parts.find((t) => /^[A-Z][a-z]{3}$/.test(t));
+  if (subtag && SUBTAG_SCRIPTS[subtag]) return SUBTAG_SCRIPTS[subtag];
+  const base = (parts[0] ?? "").toLowerCase();
+  if (LATIN_BASE.has(base)) return ["Latin"];
+  const primary = SCRIPT_TABLE[base];
+  if (primary) return [...primary, "Latin"];
+  return ["Latin"];
+}
+var BASELINE = String.raw`\p{White_Space}\p{Script=Common}\p{Script=Inherited}\p{Mark}`;
+
+// src/i18n/core.ts
+function toChain(fallback) {
+  return Array.isArray(fallback) ? fallback : fallback ? [fallback] : [];
+}
+function pickFromChain(value, chain) {
+  for (const fb of chain) {
+    if (fb === "any") {
+      const any = Object.values(value).find((v) => v !== "");
+      if (any !== void 0) return any;
+    } else if (value[fb] !== void 0 && value[fb] !== "") {
+      return value[fb];
+    }
+  }
+  return void 0;
+}
+function resolveI18nText(value, locale, fallback, field, opts) {
+  if (locale === "raw") {
+    return value;
+  }
+  if (!locale) {
+    throw new LocaleNotSpecifiedError(field ?? "<unknown>");
+  }
+  if (value[locale] !== void 0 && value[locale] !== "") {
+    return value[locale];
+  }
+  const policy = opts?.policy ?? "throw";
+  const callerChain = toChain(fallback);
+  const callerHit = pickFromChain(value, callerChain);
+  if (callerHit !== void 0) return callerHit;
+  if (policy === "substitute") {
+    const subHit = pickFromChain(value, toChain(opts?.substitute));
+    if (subHit !== void 0) return subHit;
+    if (opts?.smartSubstitute) {
+      const smartHit = pickNearestScript(value, locale);
+      if (smartHit !== void 0) return smartHit;
+    }
+  }
+  if (policy === "throw") {
+    throw new LocaleNotSpecifiedError(
+      field ?? "<unknown>",
+      `No translation available for locale "${locale}"` + (callerChain.length > 0 ? ` or fallback chain [${callerChain.join(", ")}]` : "") + "."
+    );
+  }
+  return null;
+}
+function pickNearestScript(value, target) {
+  const targetScript = inferScripts(target)[0] ?? "Latin";
+  let best;
+  for (const [loc, v] of Object.entries(value)) {
+    if (typeof v !== "string" || v === "") continue;
+    const s = inferScripts(loc)[0] ?? "Latin";
+    const score = s === targetScript ? 0 : s === "Latin" ? 1 : 2;
+    if (best === void 0 || score < best.score) best = { score, v };
+    if (score === 0) break;
+  }
+  return best?.v;
+}
+function applyAtPath(obj, path, locale, fallback, opts) {
+  const arrayIdx = path.indexOf("[].");
+  if (arrayIdx !== -1) {
+    const arrayKey = path.slice(0, arrayIdx);
+    const restPath = path.slice(arrayIdx + 3);
+    const arr = obj[arrayKey];
+    if (!Array.isArray(arr)) return obj;
+    return {
+      ...obj,
+      [arrayKey]: arr.map((item) => {
+        if (!item || typeof item !== "object" || Array.isArray(item)) return item;
+        return applyAtPath(item, restPath, locale, fallback, opts);
+      })
+    };
+  }
+  const dotIdx = path.indexOf(".");
+  if (dotIdx !== -1) {
+    const head = path.slice(0, dotIdx);
+    const rest = path.slice(dotIdx + 1);
+    const nested = obj[head];
+    if (!nested || typeof nested !== "object" || Array.isArray(nested)) return obj;
+    return {
+      ...obj,
+      [head]: applyAtPath(nested, rest, locale, fallback, opts)
+    };
+  }
+  const raw = obj[path];
+  if (raw === void 0 || raw === null) return obj;
+  if (typeof raw !== "object" || Array.isArray(raw)) return obj;
+  return {
+    ...obj,
+    [path]: resolveI18nText(raw, locale, fallback, path, opts)
+  };
+}
+function applyI18nLocale(record, i18nFields, locale, fallback, layer = "read") {
+  const fieldNames = Object.keys(i18nFields);
+  if (fieldNames.length === 0) return record;
+  let result = record;
+  for (const [field, descriptor] of Object.entries(i18nFields)) {
+    const { onMissing, substitute, smartSubstitute } = descriptor.options;
+    const opts = {
+      policy: resolvePolicy(onMissing, layer),
+      ...substitute !== void 0 ? { substitute } : {},
+      ...smartSubstitute ? { smartSubstitute } : {}
+    };
+    result = applyAtPath(result, field, locale, fallback, opts);
+  }
+  return result;
+}
+
 // src/aggregate/groupby.ts
 var GROUPBY_WARN_CARDINALITY = 1e4;
 var GROUPBY_MAX_CARDINALITY = 1e5;
@@ -705,9 +909,29 @@ var GroupedAggregation = class {
   upstreams;
   dictLabelResolver;
   fields;
-  /** Execute the query, group, reduce, and return an array of rows. */
-  run() {
-    return groupAndReduce(this.executeRecords(), this.fields, this.spec);
+  /**
+   * Execute the query, group, reduce, and return an array of rows.
+   *
+   * `opts` (#285 query-form MV grouping): when a `locale` + `i18nFields` are
+   * given, the declared group-key `i18nText` fields are resolved to that locale
+   * at the `mv` layer BEFORE bucketing — so an i18n group key is a stable string
+   * instead of a raw `{locale}` map. The MV executor passes the MV's
+   * `i18nLocale`/`i18nFields`; ordinary `.run()` callers pass nothing and are
+   * unaffected.
+   */
+  run(opts) {
+    let records = this.executeRecords();
+    if (opts?.locale !== void 0 && opts.i18nFields !== void 0) {
+      const groupI18n = {};
+      for (const f of this.fields) {
+        const d = opts.i18nFields[f];
+        if (d !== void 0) groupI18n[f] = d;
+      }
+      if (Object.keys(groupI18n).length > 0) {
+        records = records.map((r) => applyI18nLocale(r, groupI18n, opts.locale, void 0, "mv"));
+      }
+    }
+    return groupAndReduce(records, this.fields, this.spec);
   }
   /**
    * Execute the query, group, reduce, and resolve `<field>Label` for