thaizip 0.7.5 → 0.7.7

package/README.md

@@ -1,176 +1,56 @@
 # thaizip
 
 [![npm version](https://img.shields.io/npm/v/thaizip.svg)](https://www.npmjs.com/package/thaizip)
-[![npm downloads](https://img.shields.io/npm/dm/thaizip.svg)](https://www.npmjs.com/package/thaizip)
 [![license](https://img.shields.io/npm/l/thaizip.svg)](https://github.com/naay99999/thai-zip/blob/main/LICENSE)
 
-Fast fuzzy autocomplete for Thai addresses — subdistrict, district, province, postal code. Thai and English input, zero runtime dependencies, React optional.
+Fast fuzzy autocomplete for Thai addresses — subdistrict, district, province, postal code.
 
-**Docs:** https://naay99999.github.io/thai-zip/ (ไทย / [English](https://naay99999.github.io/thai-zip/en/))
+📚 **[Documentation](https://naay99999.github.io/thai-zip/)** (ไทย / [English](https://naay99999.github.io/thai-zip/en/)) · [Playground](https://naay99999.github.io/thai-zip/en/playground/)
 
-| Thai input | English input (romanization alias) |
-|---|---|
-| ![thaizip demo — typing "ลาดพร้าว" and getting fuzzy Thai suggestions](.github/assets/demo.png) | ![thaizip demo — typing "ladprao" and matching the same Thai suggestions](.github/assets/demo-en.png) |
+![thaizip demo — typing "ลาดพร้าว", "ladprao" and "10230" to get fuzzy Thai address suggestions, then resolving a selection](.github/assets/demo.gif)
+
+## Features
+
+- **Fuzzy search** in Thai or English — typos and common spellings like `ladprao` still match
+- **Postal code lookup** — exact or prefix (`10230`, `102`)
+- **Cascade dropdowns** — province → district → subdistrict
+- **Bundled data** — 77 provinces, 918 districts, 7,385 subdistricts, loaded lazily
+- **Zero runtime dependencies**, works anywhere JS runs; optional React hook
+
+## Install
 
 ```bash
 npm install thaizip
 ```
 
-| import path | contents |
-|---|---|
-| `thaizip` | core functions + types (no React code, 4.4 KB gzip) |
-| `thaizip/react` | `useThaiAddressAutocomplete` (ships `"use client"`) |
-| `thaizip/data` | `loadDefaultIndex`, `clearDefaultIndex`, `getDefaultIndexIfLoaded` (2.3 KB gzip, plus a 114 KB gzip data chunk fetched lazily on the first `loadDefaultIndex()` call) |
-
 ## Quick start
 
 ```ts
 import { loadDefaultIndex } from 'thaizip/data'
-import { searchThaiAddress, formatThaiAddressSuggestion, resolveThaiAddress } from 'thaizip'
+import { searchThaiAddress, resolveThaiAddress } from 'thaizip'
 
-const index = await loadDefaultIndex() // ~40ms of synchronous work, cached after
+const index = await loadDefaultIndex()
 
-searchThaiAddress(index, 'ลาดพร้าว')
-searchThaiAddress(index, 'bang rak')
-searchThaiAddress(index, '10500')
-
-formatThaiAddressSuggestion(record)                  // for the dropdown
-formatThaiAddressSuggestion(record, { locale: 'en' })
-resolveThaiAddress(record)                           // for saving
+const [record] = searchThaiAddress(index, 'ลาดพร้าว')
+resolveThaiAddress(record)
+// { subdistrict: 'ลาดพร้าว', district: 'เขตลาดพร้าว', province: 'กรุงเทพมหานคร', postalCode: '10230', ... }
 ```
 
-Call `loadDefaultIndex()` at mount or route load — deferring it to the user's first keystroke costs ~2 dropped frames.
-
-## Search
-
-```ts
-searchThaiAddress(index, query, {
-  limit: 10,                 // text queries only
-  threshold: 0.4,            // match quality 0–1
-  zipLimit: Infinity,        // all-digit queries — unlimited by default
-  romanizationAliases: true, // expand non-RTGS spellings
-})
-```
-
-Results rank by trigram score, then by how well the query matches the subdistrict's **own** name (exact → prefix → substring), then alphabetically. That second key is what keeps ตำบลลาดพร้าว above the other subdistricts that merely sit inside เขตลาดพร้าว.
-
-> Combined text + zip in one query (`"ลาดพร้าว 10900"`) is not supported — search by name or zip separately.
-
-### Postal codes
-
-One Thai postal code can cover many subdistricts — `45000` covers 33, and 230 of the 953 codes cover more than 10 — so `zipLimit` defaults to unlimited rather than silently truncating at `limit`.
-
-```ts
-import { lookupByZipCode } from 'thaizip'
-
-lookupByZipCode(index, '45000') // exact
-lookupByZipCode(index, '450')   // prefix
-```
-
-Direct ZIP lookup rejects raw inputs longer than 1000 characters before scanning them. A finite `zipLimit` also bounds how many matching records are collected.
-
-### English input
-
-The dataset indexes official RTGS spellings, so `bang rak` and `chatuchak` work directly. 87 common non-RTGS spellings are mapped on first: `lardprao` → `lat phrao`, `ladkrabang` → `lat krabang`, `krungthep` → `bangkok`. It is a curated dictionary, not a transliterator — unlisted spellings still miss. Opt out with `romanizationAliases: false`.
-
-### Cascade dropdowns
-
-```ts
-import { listProvinces, listAmphures, listTambons } from 'thaizip'
-
-listProvinces(index)            // 77, sorted by Thai name
-listAmphures(index, provinceId)
-listTambons(index, amphureId)   // each with its zipCode
-```
-
-Backed by pre-built groupings, not scans over all 7,385 records. Unknown ids return `[]`.
-
-## React
+React:
 
 ```tsx
 import { useThaiAddressAutocomplete } from 'thaizip/react'
 
-const { query, setQuery, setQuerySilent, suggestions, isOpen, selectSuggestion, clear } =
-  useThaiAddressAutocomplete({ index })
-```
-
-Options: `index`, `limit`, `debounce` (200ms), `threshold`, `zipLimit`, `locale`, `initialQuery`, `onSelect`.
-
-| value | description |
-|---|---|
-| `query` / `setQuery` | input value; `setQuery` searches |
-| `setQuerySilent` | set the value **without** searching or reopening the dropdown |
-| `suggestions` | `ThaiAddressSuggestion[]` |
-| `isOpen` | `query` is non-empty and suggestions exist |
-| `selectSuggestion` | `(item) => ResolvedThaiAddress \| null` — clears suggestions, leaves `query` alone; `null` if the item is stale |
-| `clear` | reset query and suggestions |
-
-### Controlled forms
-
-`selectSuggestion` leaves `query` untouched by design. To echo the choice back into the input use `setQuerySilent` — plain `setQuery` would immediately reopen the dropdown:
-
-```tsx
-const handlePick = (s: ThaiAddressSuggestion) => {
-  const address = selectSuggestion(s)
-  if (!address) return
-  setQuerySilent(s.label)
-  form.setValue('address', address)
-}
-```
-
-Editing a saved address? Pass `initialQuery` rather than calling `setQuery` in an effect.
-
-### Accessibility
-
-The hook owns query and suggestion state only. Keyboard navigation and ARIA combobox semantics are yours — `npx react-thaizip add autocomplete` scaffolds a component that already has them.
-
-## Custom data
-
-```ts
-import { buildThaiAddressIndex, validateRawData } from 'thaizip'
-
-const index = buildThaiAddressIndex({ provinces, amphures, tambons }, {
-  onSkip: (tambon) => console.warn('skipped:', tambon.id),
-  validate: true, // default
-})
-```
-
-Input is validated by default, so a stray non-string field fails with `[thaizip] tambon 100404: expected string for name_th, got number` instead of crashing inside the normalizer. It costs nothing measurable on a full-size dataset. `validateRawData(data)` runs the same checks standalone.
-
-`RawData` is meant to be trusted, application-controlled input — your own dataset, not arbitrary end-user uploads. `buildThaiAddressIndex` has no built-in cap on row count or field length; build cost scales linearly with total input size, but an unbounded or maliciously oversized payload (e.g. an admin importer fed directly to this function) can still cost real time and memory with nothing to stop it. If you ever build an index from untrusted input, enforce your own size limits (row count, field length) before calling `buildThaiAddressIndex`.
-
-Types: `RawData`, `RawProvince`, `RawAmphure`, `RawTambon`.
-
-## Types
-
-```ts
-type ThaiAddressSuggestion = {
-  id: string
-  label: string      // follows `locale`, Thai by default
-  labelTh: string
-  labelEn: string
-  tambon: string;    tambonEn: string
-  amphure: string;   amphureEn: string
-  province: string;  provinceEn: string
-  zipCode: string
-}
-
-type ResolvedThaiAddress = {
-  tambon: string;    tambonEn: string     // alias: subdistrict / subdistrictEn
-  amphure: string;   amphureEn: string    // alias: district / districtEn
-  province: string;  provinceEn: string
-  zipCode: string                         // alias: postalCode
-  subdistrict: string; subdistrictEn: string
-  district: string;    districtEn: string
-  postalCode: string
-}
+const { query, setQuery, suggestions, selectSuggestion } = useThaiAddressAutocomplete({ index })
 ```
 
-## Data
+Need a ready-made UI component? [`react-thaizip`](https://github.com/naay99999/react-thai-zip) scaffolds one into your project.
 
-77 provinces, 918 districts, 7,385 subdistricts. Records whose parent district or province is soft-deleted are excluded automatically.
+## Learn more
 
-`searchThaiAddress` is pure and framework-free — it works unchanged in Node, Vue, Svelte, or vanilla JS.
+- [Getting started](https://naay99999.github.io/thai-zip/en/getting-started/)
+- [Guides](https://naay99999.github.io/thai-zip/en/guides/search/) — search, postal codes, cascades, English input, React, custom data
+- [API reference](https://naay99999.github.io/thai-zip/en/reference/search/)
 
 ## License
 

package/dist/data.cjs

@@ -35,9 +35,19 @@ function extractTrigramsNormalized(text) {
 }
 
 // src/core/indexer.ts
+var MAX_TOTAL_ROWS = 5e4;
+var MAX_NAME_LENGTH = 256;
+var MAX_ZIP_LENGTH = 32;
+var MAX_POSTING_BUDGET = 1e6;
 function typeErr(table, id, field, expected, value) {
   return new TypeError(`[thaizip] ${table} ${id}: expected ${expected} for ${field}, got ${typeof value}`);
 }
+function assertId(table, id, field, value) {
+  if (typeof value !== "number") throw typeErr(table, id, field, "number", value);
+  if (!Number.isSafeInteger(value)) {
+    throw new TypeError(`[thaizip] ${table} ${id}: expected safe integer for ${field}, got ${value}`);
+  }
+}
 function typeName(value) {
   return value === null ? "null" : typeof value;
 }
@@ -55,47 +65,79 @@ function validateRawData(data) {
   if (!Array.isArray(tambons)) {
     throw new TypeError(`[thaizip] RawData.tambons must be an array, got ${typeName(tambons)}`);
   }
-  const seenProvinceIds = /* @__PURE__ */ new Set();
+  if (provinces.length + amphures.length + tambons.length > MAX_TOTAL_ROWS) {
+    throw new RangeError(`[thaizip] RawData row limit exceeded (${MAX_TOTAL_ROWS})`);
+  }
+  const provinceById = /* @__PURE__ */ new Map();
   for (let i = 0; i < provinces.length; i++) {
     const p = provinces[i];
     if (p === null || typeof p !== "object") {
       throw new TypeError(`[thaizip] province[${i}]: expected object, got ${typeName(p)}`);
     }
-    if (typeof p.id !== "number") throw typeErr("province", p.id, "id", "number", p.id);
-    if (seenProvinceIds.has(p.id)) throw new TypeError(`[thaizip] duplicate province id: ${p.id}`);
-    seenProvinceIds.add(p.id);
+    assertId("province", p.id, "id", p.id);
+    if (provinceById.has(p.id)) throw new TypeError(`[thaizip] duplicate province id: ${p.id}`);
+    provinceById.set(p.id, p);
     if (typeof p.name_th !== "string") throw typeErr("province", p.id, "name_th", "string", p.name_th);
     if (typeof p.name_en !== "string") throw typeErr("province", p.id, "name_en", "string", p.name_en);
+    if (p.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] province ${p.id}: name_th too long`);
+    if (p.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] province ${p.id}: name_en too long`);
   }
-  const seenAmphureIds = /* @__PURE__ */ new Set();
+  const amphureById = /* @__PURE__ */ new Map();
   for (let i = 0; i < amphures.length; i++) {
     const a = amphures[i];
     if (a === null || typeof a !== "object") {
       throw new TypeError(`[thaizip] amphure[${i}]: expected object, got ${typeName(a)}`);
     }
-    if (typeof a.id !== "number") throw typeErr("amphure", a.id, "id", "number", a.id);
-    if (seenAmphureIds.has(a.id)) throw new TypeError(`[thaizip] duplicate amphure id: ${a.id}`);
-    seenAmphureIds.add(a.id);
+    assertId("amphure", a.id, "id", a.id);
+    if (amphureById.has(a.id)) throw new TypeError(`[thaizip] duplicate amphure id: ${a.id}`);
+    amphureById.set(a.id, a);
     if (typeof a.name_th !== "string") throw typeErr("amphure", a.id, "name_th", "string", a.name_th);
     if (typeof a.name_en !== "string") throw typeErr("amphure", a.id, "name_en", "string", a.name_en);
-    if (typeof a.province_id !== "number") throw typeErr("amphure", a.id, "province_id", "number", a.province_id);
+    assertId("amphure", a.id, "province_id", a.province_id);
+    if (a.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] amphure ${a.id}: name_th too long`);
+    if (a.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] amphure ${a.id}: name_en too long`);
   }
   const seenTambonIds = /* @__PURE__ */ new Set();
+  const provinceTrigramCounts = /* @__PURE__ */ new Map();
+  const amphureTrigramCounts = /* @__PURE__ */ new Map();
+  let postingBudget = 0;
   for (let i = 0; i < tambons.length; i++) {
     const t = tambons[i];
     if (t === null || typeof t !== "object") {
       throw new TypeError(`[thaizip] tambon[${i}]: expected object, got ${typeName(t)}`);
     }
-    if (typeof t.id !== "number") throw typeErr("tambon", t.id, "id", "number", t.id);
+    assertId("tambon", t.id, "id", t.id);
     if (seenTambonIds.has(t.id)) throw new TypeError(`[thaizip] duplicate tambon id: ${t.id}`);
     seenTambonIds.add(t.id);
     if (typeof t.name_th !== "string") throw typeErr("tambon", t.id, "name_th", "string", t.name_th);
     if (typeof t.name_en !== "string") throw typeErr("tambon", t.id, "name_en", "string", t.name_en);
+    if (t.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: name_th too long`);
+    if (t.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: name_en too long`);
     const zipType = typeof t.zip_code;
     if (zipType !== "string" && zipType !== "number") {
       throw typeErr("tambon", t.id, "zip_code", "string or number", t.zip_code);
     }
-    if (typeof t.amphure_id !== "number") throw typeErr("tambon", t.id, "amphure_id", "number", t.amphure_id);
+    assertId("tambon", t.id, "amphure_id", t.amphure_id);
+    const zipLength = String(t.zip_code).length;
+    if (zipLength > MAX_ZIP_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: zip_code too long`);
+    if (t.deleted_at) continue;
+    const amphure = amphureById.get(t.amphure_id);
+    const province = amphure && provinceById.get(amphure.province_id);
+    if (!amphure || amphure.deleted_at || !province || province.deleted_at) continue;
+    let provinceGrams = provinceTrigramCounts.get(province.id);
+    if (provinceGrams === void 0) {
+      provinceGrams = combinedTrigrams(province.name_th, province.name_en).size;
+      provinceTrigramCounts.set(province.id, provinceGrams);
+    }
+    let amphureGrams = amphureTrigramCounts.get(amphure.id);
+    if (amphureGrams === void 0) {
+      amphureGrams = combinedTrigrams(amphure.name_th, amphure.name_en).size;
+      amphureTrigramCounts.set(amphure.id, amphureGrams);
+    }
+    postingBudget += t.name_th.length + t.name_en.length + zipLength + provinceGrams + amphureGrams;
+    if (postingBudget > MAX_POSTING_BUDGET) {
+      throw new RangeError(`[thaizip] estimated trigram posting budget exceeded (${MAX_POSTING_BUDGET})`);
+    }
   }
 }
 function addTrigrams(map, trigrams, idx) {
@@ -126,13 +168,7 @@ function buildThaiAddressIndex(data, options) {
     amphures.filter((a) => !a.deleted_at).map((a) => [a.id, a])
   );
   const provTrigrams = /* @__PURE__ */ new Map();
-  for (const [id, prov] of provMap) {
-    provTrigrams.set(id, combinedTrigrams(prov.name_th, prov.name_en));
-  }
   const ampTrigrams = /* @__PURE__ */ new Map();
-  for (const [id, amp] of ampMap) {
-    ampTrigrams.set(id, combinedTrigrams(amp.name_th, amp.name_en));
-  }
   const records = [];
   const map = /* @__PURE__ */ new Map();
   const zipIndex = /* @__PURE__ */ new Map();
@@ -182,8 +218,18 @@ function buildThaiAddressIndex(data, options) {
     normTambonEn.push(normEn);
     addTrigrams(map, extractTrigramsNormalized(normEn), idx);
     addTrigrams(map, extractTrigrams(record.zipCode), idx);
-    addTrigrams(map, provTrigrams.get(province.id), idx);
-    addTrigrams(map, ampTrigrams.get(amphure.id), idx);
+    let provinceGrams = provTrigrams.get(province.id);
+    if (!provinceGrams) {
+      provinceGrams = combinedTrigrams(province.name_th, province.name_en);
+      provTrigrams.set(province.id, provinceGrams);
+    }
+    let amphureGrams = ampTrigrams.get(amphure.id);
+    if (!amphureGrams) {
+      amphureGrams = combinedTrigrams(amphure.name_th, amphure.name_en);
+      ampTrigrams.set(amphure.id, amphureGrams);
+    }
+    addTrigrams(map, provinceGrams, idx);
+    addTrigrams(map, amphureGrams, idx);
   }
   const sortedZipKeys = [...zipIndex.keys()].sort();
   const sortedZipPostings = sortedZipKeys.map((z) => zipIndex.get(z));
@@ -194,6 +240,40 @@ function buildThaiAddressIndex(data, options) {
 var cached = null;
 var inflightPromise = null;
 var loadGeneration = 0;
+function rejectDefaultIndexMutation() {
+  throw new TypeError("[thaizip] the shared default index is read-only");
+}
+var readOnlyMapPrototype = Object.create(Map.prototype);
+var readOnlySetPrototype = Object.create(Set.prototype);
+for (const [prototype, methods] of [
+  [readOnlyMapPrototype, ["set", "delete", "clear"]],
+  [readOnlySetPrototype, ["add", "delete", "clear"]]
+]) {
+  for (const method of methods) {
+    Object.defineProperty(prototype, method, { value: rejectDefaultIndexMutation });
+  }
+}
+function lockMapValues(map, lockValue) {
+  for (const value of map.values()) lockValue(value);
+  Object.setPrototypeOf(map, readOnlyMapPrototype);
+  Object.freeze(map);
+}
+function lockDefaultIndex(index) {
+  for (const record of index.records) Object.freeze(record);
+  Object.freeze(index.records);
+  Object.freeze(index.normTambon);
+  Object.freeze(index.normTambonEn);
+  if (index.sortedZipKeys) Object.freeze(index.sortedZipKeys);
+  if (index.sortedZipPostings) Object.freeze(index.sortedZipPostings);
+  lockMapValues(index.map, (set) => {
+    Object.setPrototypeOf(set, readOnlySetPrototype);
+    Object.freeze(set);
+  });
+  lockMapValues(index.zipIndex, Object.freeze);
+  lockMapValues(index.byProvince, Object.freeze);
+  lockMapValues(index.byAmphure, Object.freeze);
+  return Object.freeze(index);
+}
 function clearDefaultIndex() {
   cached = null;
   inflightPromise = null;
@@ -207,11 +287,11 @@ async function loadDefaultIndex() {
   if (!inflightPromise) {
     const gen = loadGeneration;
     inflightPromise = import('./defaultData.js').then(({ p, a, t }) => {
-      const index = buildThaiAddressIndex({
+      const index = lockDefaultIndex(buildThaiAddressIndex({
         provinces: p.map(([id, name_th, name_en]) => ({ id, name_th, name_en, geography_id: 0, deleted_at: null })),
         amphures: a.map(([id, name_th, name_en, province_id]) => ({ id, name_th, name_en, province_id, deleted_at: null })),
         tambons: t.map(([id, name_th, name_en, amphure_id, zip_code]) => ({ id, name_th, name_en, amphure_id, zip_code, deleted_at: null }))
-      }, { validate: false });
+      }, { validate: false }));
       if (loadGeneration === gen) {
         cached = index;
         inflightPromise = null;

package/dist/data.js

@@ -33,9 +33,19 @@ function extractTrigramsNormalized(text) {
 }
 
 // src/core/indexer.ts
+var MAX_TOTAL_ROWS = 5e4;
+var MAX_NAME_LENGTH = 256;
+var MAX_ZIP_LENGTH = 32;
+var MAX_POSTING_BUDGET = 1e6;
 function typeErr(table, id, field, expected, value) {
   return new TypeError(`[thaizip] ${table} ${id}: expected ${expected} for ${field}, got ${typeof value}`);
 }
+function assertId(table, id, field, value) {
+  if (typeof value !== "number") throw typeErr(table, id, field, "number", value);
+  if (!Number.isSafeInteger(value)) {
+    throw new TypeError(`[thaizip] ${table} ${id}: expected safe integer for ${field}, got ${value}`);
+  }
+}
 function typeName(value) {
   return value === null ? "null" : typeof value;
 }
@@ -53,47 +63,79 @@ function validateRawData(data) {
   if (!Array.isArray(tambons)) {
     throw new TypeError(`[thaizip] RawData.tambons must be an array, got ${typeName(tambons)}`);
   }
-  const seenProvinceIds = /* @__PURE__ */ new Set();
+  if (provinces.length + amphures.length + tambons.length > MAX_TOTAL_ROWS) {
+    throw new RangeError(`[thaizip] RawData row limit exceeded (${MAX_TOTAL_ROWS})`);
+  }
+  const provinceById = /* @__PURE__ */ new Map();
   for (let i = 0; i < provinces.length; i++) {
     const p = provinces[i];
     if (p === null || typeof p !== "object") {
       throw new TypeError(`[thaizip] province[${i}]: expected object, got ${typeName(p)}`);
     }
-    if (typeof p.id !== "number") throw typeErr("province", p.id, "id", "number", p.id);
-    if (seenProvinceIds.has(p.id)) throw new TypeError(`[thaizip] duplicate province id: ${p.id}`);
-    seenProvinceIds.add(p.id);
+    assertId("province", p.id, "id", p.id);
+    if (provinceById.has(p.id)) throw new TypeError(`[thaizip] duplicate province id: ${p.id}`);
+    provinceById.set(p.id, p);
     if (typeof p.name_th !== "string") throw typeErr("province", p.id, "name_th", "string", p.name_th);
     if (typeof p.name_en !== "string") throw typeErr("province", p.id, "name_en", "string", p.name_en);
+    if (p.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] province ${p.id}: name_th too long`);
+    if (p.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] province ${p.id}: name_en too long`);
   }
-  const seenAmphureIds = /* @__PURE__ */ new Set();
+  const amphureById = /* @__PURE__ */ new Map();
   for (let i = 0; i < amphures.length; i++) {
     const a = amphures[i];
     if (a === null || typeof a !== "object") {
       throw new TypeError(`[thaizip] amphure[${i}]: expected object, got ${typeName(a)}`);
     }
-    if (typeof a.id !== "number") throw typeErr("amphure", a.id, "id", "number", a.id);
-    if (seenAmphureIds.has(a.id)) throw new TypeError(`[thaizip] duplicate amphure id: ${a.id}`);
-    seenAmphureIds.add(a.id);
+    assertId("amphure", a.id, "id", a.id);
+    if (amphureById.has(a.id)) throw new TypeError(`[thaizip] duplicate amphure id: ${a.id}`);
+    amphureById.set(a.id, a);
     if (typeof a.name_th !== "string") throw typeErr("amphure", a.id, "name_th", "string", a.name_th);
     if (typeof a.name_en !== "string") throw typeErr("amphure", a.id, "name_en", "string", a.name_en);
-    if (typeof a.province_id !== "number") throw typeErr("amphure", a.id, "province_id", "number", a.province_id);
+    assertId("amphure", a.id, "province_id", a.province_id);
+    if (a.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] amphure ${a.id}: name_th too long`);
+    if (a.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] amphure ${a.id}: name_en too long`);
   }
   const seenTambonIds = /* @__PURE__ */ new Set();
+  const provinceTrigramCounts = /* @__PURE__ */ new Map();
+  const amphureTrigramCounts = /* @__PURE__ */ new Map();
+  let postingBudget = 0;
   for (let i = 0; i < tambons.length; i++) {
     const t = tambons[i];
     if (t === null || typeof t !== "object") {
       throw new TypeError(`[thaizip] tambon[${i}]: expected object, got ${typeName(t)}`);
     }
-    if (typeof t.id !== "number") throw typeErr("tambon", t.id, "id", "number", t.id);
+    assertId("tambon", t.id, "id", t.id);
     if (seenTambonIds.has(t.id)) throw new TypeError(`[thaizip] duplicate tambon id: ${t.id}`);
     seenTambonIds.add(t.id);
     if (typeof t.name_th !== "string") throw typeErr("tambon", t.id, "name_th", "string", t.name_th);
     if (typeof t.name_en !== "string") throw typeErr("tambon", t.id, "name_en", "string", t.name_en);
+    if (t.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: name_th too long`);
+    if (t.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: name_en too long`);
     const zipType = typeof t.zip_code;
     if (zipType !== "string" && zipType !== "number") {
       throw typeErr("tambon", t.id, "zip_code", "string or number", t.zip_code);
     }
-    if (typeof t.amphure_id !== "number") throw typeErr("tambon", t.id, "amphure_id", "number", t.amphure_id);
+    assertId("tambon", t.id, "amphure_id", t.amphure_id);
+    const zipLength = String(t.zip_code).length;
+    if (zipLength > MAX_ZIP_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: zip_code too long`);
+    if (t.deleted_at) continue;
+    const amphure = amphureById.get(t.amphure_id);
+    const province = amphure && provinceById.get(amphure.province_id);
+    if (!amphure || amphure.deleted_at || !province || province.deleted_at) continue;
+    let provinceGrams = provinceTrigramCounts.get(province.id);
+    if (provinceGrams === void 0) {
+      provinceGrams = combinedTrigrams(province.name_th, province.name_en).size;
+      provinceTrigramCounts.set(province.id, provinceGrams);
+    }
+    let amphureGrams = amphureTrigramCounts.get(amphure.id);
+    if (amphureGrams === void 0) {
+      amphureGrams = combinedTrigrams(amphure.name_th, amphure.name_en).size;
+      amphureTrigramCounts.set(amphure.id, amphureGrams);
+    }
+    postingBudget += t.name_th.length + t.name_en.length + zipLength + provinceGrams + amphureGrams;
+    if (postingBudget > MAX_POSTING_BUDGET) {
+      throw new RangeError(`[thaizip] estimated trigram posting budget exceeded (${MAX_POSTING_BUDGET})`);
+    }
   }
 }
 function addTrigrams(map, trigrams, idx) {
@@ -124,13 +166,7 @@ function buildThaiAddressIndex(data, options) {
     amphures.filter((a) => !a.deleted_at).map((a) => [a.id, a])
   );
   const provTrigrams = /* @__PURE__ */ new Map();
-  for (const [id, prov] of provMap) {
-    provTrigrams.set(id, combinedTrigrams(prov.name_th, prov.name_en));
-  }
   const ampTrigrams = /* @__PURE__ */ new Map();
-  for (const [id, amp] of ampMap) {
-    ampTrigrams.set(id, combinedTrigrams(amp.name_th, amp.name_en));
-  }
   const records = [];
   const map = /* @__PURE__ */ new Map();
   const zipIndex = /* @__PURE__ */ new Map();
@@ -180,8 +216,18 @@ function buildThaiAddressIndex(data, options) {
     normTambonEn.push(normEn);
     addTrigrams(map, extractTrigramsNormalized(normEn), idx);
     addTrigrams(map, extractTrigrams(record.zipCode), idx);
-    addTrigrams(map, provTrigrams.get(province.id), idx);
-    addTrigrams(map, ampTrigrams.get(amphure.id), idx);
+    let provinceGrams = provTrigrams.get(province.id);
+    if (!provinceGrams) {
+      provinceGrams = combinedTrigrams(province.name_th, province.name_en);
+      provTrigrams.set(province.id, provinceGrams);
+    }
+    let amphureGrams = ampTrigrams.get(amphure.id);
+    if (!amphureGrams) {
+      amphureGrams = combinedTrigrams(amphure.name_th, amphure.name_en);
+      ampTrigrams.set(amphure.id, amphureGrams);
+    }
+    addTrigrams(map, provinceGrams, idx);
+    addTrigrams(map, amphureGrams, idx);
   }
   const sortedZipKeys = [...zipIndex.keys()].sort();
   const sortedZipPostings = sortedZipKeys.map((z) => zipIndex.get(z));
@@ -192,6 +238,40 @@ function buildThaiAddressIndex(data, options) {
 var cached = null;
 var inflightPromise = null;
 var loadGeneration = 0;
+function rejectDefaultIndexMutation() {
+  throw new TypeError("[thaizip] the shared default index is read-only");
+}
+var readOnlyMapPrototype = Object.create(Map.prototype);
+var readOnlySetPrototype = Object.create(Set.prototype);
+for (const [prototype, methods] of [
+  [readOnlyMapPrototype, ["set", "delete", "clear"]],
+  [readOnlySetPrototype, ["add", "delete", "clear"]]
+]) {
+  for (const method of methods) {
+    Object.defineProperty(prototype, method, { value: rejectDefaultIndexMutation });
+  }
+}
+function lockMapValues(map, lockValue) {
+  for (const value of map.values()) lockValue(value);
+  Object.setPrototypeOf(map, readOnlyMapPrototype);
+  Object.freeze(map);
+}
+function lockDefaultIndex(index) {
+  for (const record of index.records) Object.freeze(record);
+  Object.freeze(index.records);
+  Object.freeze(index.normTambon);
+  Object.freeze(index.normTambonEn);
+  if (index.sortedZipKeys) Object.freeze(index.sortedZipKeys);
+  if (index.sortedZipPostings) Object.freeze(index.sortedZipPostings);
+  lockMapValues(index.map, (set) => {
+    Object.setPrototypeOf(set, readOnlySetPrototype);
+    Object.freeze(set);
+  });
+  lockMapValues(index.zipIndex, Object.freeze);
+  lockMapValues(index.byProvince, Object.freeze);
+  lockMapValues(index.byAmphure, Object.freeze);
+  return Object.freeze(index);
+}
 function clearDefaultIndex() {
   cached = null;
   inflightPromise = null;
@@ -205,11 +285,11 @@ async function loadDefaultIndex() {
   if (!inflightPromise) {
     const gen = loadGeneration;
     inflightPromise = import('./defaultData.js').then(({ p, a, t }) => {
-      const index = buildThaiAddressIndex({
+      const index = lockDefaultIndex(buildThaiAddressIndex({
         provinces: p.map(([id, name_th, name_en]) => ({ id, name_th, name_en, geography_id: 0, deleted_at: null })),
         amphures: a.map(([id, name_th, name_en, province_id]) => ({ id, name_th, name_en, province_id, deleted_at: null })),
         tambons: t.map(([id, name_th, name_en, amphure_id, zip_code]) => ({ id, name_th, name_en, amphure_id, zip_code, deleted_at: null }))
-      }, { validate: false });
+      }, { validate: false }));
       if (loadGeneration === gen) {
         cached = index;
         inflightPromise = null;

package/dist/index.cjs

@@ -35,9 +35,19 @@ function extractTrigramsNormalized(text) {
 }
 
 // src/core/indexer.ts
+var MAX_TOTAL_ROWS = 5e4;
+var MAX_NAME_LENGTH = 256;
+var MAX_ZIP_LENGTH = 32;
+var MAX_POSTING_BUDGET = 1e6;
 function typeErr(table, id, field, expected, value) {
   return new TypeError(`[thaizip] ${table} ${id}: expected ${expected} for ${field}, got ${typeof value}`);
 }
+function assertId(table, id, field, value) {
+  if (typeof value !== "number") throw typeErr(table, id, field, "number", value);
+  if (!Number.isSafeInteger(value)) {
+    throw new TypeError(`[thaizip] ${table} ${id}: expected safe integer for ${field}, got ${value}`);
+  }
+}
 function typeName(value) {
   return value === null ? "null" : typeof value;
 }
@@ -55,47 +65,79 @@ function validateRawData(data) {
   if (!Array.isArray(tambons)) {
     throw new TypeError(`[thaizip] RawData.tambons must be an array, got ${typeName(tambons)}`);
   }
-  const seenProvinceIds = /* @__PURE__ */ new Set();
+  if (provinces.length + amphures.length + tambons.length > MAX_TOTAL_ROWS) {
+    throw new RangeError(`[thaizip] RawData row limit exceeded (${MAX_TOTAL_ROWS})`);
+  }
+  const provinceById = /* @__PURE__ */ new Map();
   for (let i = 0; i < provinces.length; i++) {
     const p = provinces[i];
     if (p === null || typeof p !== "object") {
       throw new TypeError(`[thaizip] province[${i}]: expected object, got ${typeName(p)}`);
     }
-    if (typeof p.id !== "number") throw typeErr("province", p.id, "id", "number", p.id);
-    if (seenProvinceIds.has(p.id)) throw new TypeError(`[thaizip] duplicate province id: ${p.id}`);
-    seenProvinceIds.add(p.id);
+    assertId("province", p.id, "id", p.id);
+    if (provinceById.has(p.id)) throw new TypeError(`[thaizip] duplicate province id: ${p.id}`);
+    provinceById.set(p.id, p);
     if (typeof p.name_th !== "string") throw typeErr("province", p.id, "name_th", "string", p.name_th);
     if (typeof p.name_en !== "string") throw typeErr("province", p.id, "name_en", "string", p.name_en);
+    if (p.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] province ${p.id}: name_th too long`);
+    if (p.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] province ${p.id}: name_en too long`);
   }
-  const seenAmphureIds = /* @__PURE__ */ new Set();
+  const amphureById = /* @__PURE__ */ new Map();
   for (let i = 0; i < amphures.length; i++) {
     const a = amphures[i];
     if (a === null || typeof a !== "object") {
       throw new TypeError(`[thaizip] amphure[${i}]: expected object, got ${typeName(a)}`);
     }
-    if (typeof a.id !== "number") throw typeErr("amphure", a.id, "id", "number", a.id);
-    if (seenAmphureIds.has(a.id)) throw new TypeError(`[thaizip] duplicate amphure id: ${a.id}`);
-    seenAmphureIds.add(a.id);
+    assertId("amphure", a.id, "id", a.id);
+    if (amphureById.has(a.id)) throw new TypeError(`[thaizip] duplicate amphure id: ${a.id}`);
+    amphureById.set(a.id, a);
     if (typeof a.name_th !== "string") throw typeErr("amphure", a.id, "name_th", "string", a.name_th);
     if (typeof a.name_en !== "string") throw typeErr("amphure", a.id, "name_en", "string", a.name_en);
-    if (typeof a.province_id !== "number") throw typeErr("amphure", a.id, "province_id", "number", a.province_id);
+    assertId("amphure", a.id, "province_id", a.province_id);
+    if (a.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] amphure ${a.id}: name_th too long`);
+    if (a.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] amphure ${a.id}: name_en too long`);
   }
   const seenTambonIds = /* @__PURE__ */ new Set();
+  const provinceTrigramCounts = /* @__PURE__ */ new Map();
+  const amphureTrigramCounts = /* @__PURE__ */ new Map();
+  let postingBudget = 0;
   for (let i = 0; i < tambons.length; i++) {
     const t = tambons[i];
     if (t === null || typeof t !== "object") {
       throw new TypeError(`[thaizip] tambon[${i}]: expected object, got ${typeName(t)}`);
     }
-    if (typeof t.id !== "number") throw typeErr("tambon", t.id, "id", "number", t.id);
+    assertId("tambon", t.id, "id", t.id);
     if (seenTambonIds.has(t.id)) throw new TypeError(`[thaizip] duplicate tambon id: ${t.id}`);
     seenTambonIds.add(t.id);
     if (typeof t.name_th !== "string") throw typeErr("tambon", t.id, "name_th", "string", t.name_th);
     if (typeof t.name_en !== "string") throw typeErr("tambon", t.id, "name_en", "string", t.name_en);
+    if (t.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: name_th too long`);
+    if (t.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: name_en too long`);
     const zipType = typeof t.zip_code;
     if (zipType !== "string" && zipType !== "number") {
       throw typeErr("tambon", t.id, "zip_code", "string or number", t.zip_code);
     }
-    if (typeof t.amphure_id !== "number") throw typeErr("tambon", t.id, "amphure_id", "number", t.amphure_id);
+    assertId("tambon", t.id, "amphure_id", t.amphure_id);
+    const zipLength = String(t.zip_code).length;
+    if (zipLength > MAX_ZIP_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: zip_code too long`);
+    if (t.deleted_at) continue;
+    const amphure = amphureById.get(t.amphure_id);
+    const province = amphure && provinceById.get(amphure.province_id);
+    if (!amphure || amphure.deleted_at || !province || province.deleted_at) continue;
+    let provinceGrams = provinceTrigramCounts.get(province.id);
+    if (provinceGrams === void 0) {
+      provinceGrams = combinedTrigrams(province.name_th, province.name_en).size;
+      provinceTrigramCounts.set(province.id, provinceGrams);
+    }
+    let amphureGrams = amphureTrigramCounts.get(amphure.id);
+    if (amphureGrams === void 0) {
+      amphureGrams = combinedTrigrams(amphure.name_th, amphure.name_en).size;
+      amphureTrigramCounts.set(amphure.id, amphureGrams);
+    }
+    postingBudget += t.name_th.length + t.name_en.length + zipLength + provinceGrams + amphureGrams;
+    if (postingBudget > MAX_POSTING_BUDGET) {
+      throw new RangeError(`[thaizip] estimated trigram posting budget exceeded (${MAX_POSTING_BUDGET})`);
+    }
   }
 }
 function addTrigrams(map, trigrams, idx) {
@@ -126,13 +168,7 @@ function buildThaiAddressIndex(data, options) {
     amphures.filter((a) => !a.deleted_at).map((a) => [a.id, a])
   );
   const provTrigrams = /* @__PURE__ */ new Map();
-  for (const [id, prov] of provMap) {
-    provTrigrams.set(id, combinedTrigrams(prov.name_th, prov.name_en));
-  }
   const ampTrigrams = /* @__PURE__ */ new Map();
-  for (const [id, amp] of ampMap) {
-    ampTrigrams.set(id, combinedTrigrams(amp.name_th, amp.name_en));
-  }
   const records = [];
   const map = /* @__PURE__ */ new Map();
   const zipIndex = /* @__PURE__ */ new Map();
@@ -182,8 +218,18 @@ function buildThaiAddressIndex(data, options) {
     normTambonEn.push(normEn);
     addTrigrams(map, extractTrigramsNormalized(normEn), idx);
     addTrigrams(map, extractTrigrams(record.zipCode), idx);
-    addTrigrams(map, provTrigrams.get(province.id), idx);
-    addTrigrams(map, ampTrigrams.get(amphure.id), idx);
+    let provinceGrams = provTrigrams.get(province.id);
+    if (!provinceGrams) {
+      provinceGrams = combinedTrigrams(province.name_th, province.name_en);
+      provTrigrams.set(province.id, provinceGrams);
+    }
+    let amphureGrams = ampTrigrams.get(amphure.id);
+    if (!amphureGrams) {
+      amphureGrams = combinedTrigrams(amphure.name_th, amphure.name_en);
+      ampTrigrams.set(amphure.id, amphureGrams);
+    }
+    addTrigrams(map, provinceGrams, idx);
+    addTrigrams(map, amphureGrams, idx);
   }
   const sortedZipKeys = [...zipIndex.keys()].sort();
   const sortedZipPostings = sortedZipKeys.map((z) => zipIndex.get(z));
@@ -381,7 +427,7 @@ function searchThaiAddress(index, query, options) {
   if (ZIP_CODE_RE.test(normalized)) {
     if (normalized.length < 2) return [];
     return lookupByZipCode(index, normalized, {
-      zipLimit: options?.zipLimit ?? Infinity
+      zipLimit: options?.zipLimit ?? (normalized.length >= 5 ? Infinity : limit)
     });
   }
   if (normalized.length < 3) return [];
@@ -543,5 +589,3 @@ exports.normalizeThaiAddressText = normalizeThaiAddressText;
 exports.resolveThaiAddress = resolveThaiAddress;
 exports.searchThaiAddress = searchThaiAddress;
 exports.validateRawData = validateRawData;
-//# sourceMappingURL=index.cjs.map
-//# sourceMappingURL=index.cjs.map

package/dist/index.d.cts

@@ -101,9 +101,11 @@ type SearchOptions = {
     limit?: number;
     threshold?: number;
     /**
-     * Maximum results for all-digit (zip code) queries. A zip code maps to at most
-     * a few dozen tambons, so this defaults to `Infinity` — capping zip results at
-     * the text-autocomplete `limit` silently hides valid subdistricts.
+     * Maximum results for all-digit (zip code) queries. When omitted, a complete
+     * 5-digit zip returns every matching tambon (at most a few dozen — capping
+     * those at the text-autocomplete `limit` would silently hide valid
+     * subdistricts), while a partial prefix (2–4 digits, which can match hundreds
+     * of records) is capped at `limit`. Pass `Infinity` to return every prefix match.
      */
     zipLimit?: number;
     /**
@@ -145,6 +147,10 @@ type BuildIndexOptions = {
      * Validate that every consumed field has the expected runtime type and throw a
      * descriptive `TypeError` if not. Defaults to `true`; set to `false` to skip
      * the checks when the data source is already trusted (e.g. the bundled index).
+     *
+     * `false` also skips the row, name-length and posting-budget limits that bound
+     * memory and CPU use, so never disable it for data you do not control
+     * (user uploads, third-party APIs, a CMS editable by others).
      */
     validate?: boolean;
 };
@@ -161,10 +167,9 @@ type RawData = {
  * private datasets) who want to fail fast with a descriptive error instead of an
  * opaque crash deep inside the normalizer, or silent `"undefined"` labels.
  *
- * Throws a `TypeError` (message prefixed `[thaizip]`) naming the table, the
- * offending record's `id`, the field, and the actual `typeof` on the first bad
- * field found. Does not filter soft-deleted rows first — validates every row as
- * provided.
+ * Throws a `TypeError` for bad shapes or duplicate IDs and a `RangeError` for
+ * resource limits (all prefixed `[thaizip]`). Does not filter soft-deleted rows
+ * before shape checks — validates every row as provided.
  *
  * IDs must be unique within each table (duplicates throw, they are not silently
  * discarded — two suggestions sharing an id would resolve to the wrong record).

package/dist/index.d.ts

@@ -101,9 +101,11 @@ type SearchOptions = {
     limit?: number;
     threshold?: number;
     /**
-     * Maximum results for all-digit (zip code) queries. A zip code maps to at most
-     * a few dozen tambons, so this defaults to `Infinity` — capping zip results at
-     * the text-autocomplete `limit` silently hides valid subdistricts.
+     * Maximum results for all-digit (zip code) queries. When omitted, a complete
+     * 5-digit zip returns every matching tambon (at most a few dozen — capping
+     * those at the text-autocomplete `limit` would silently hide valid
+     * subdistricts), while a partial prefix (2–4 digits, which can match hundreds
+     * of records) is capped at `limit`. Pass `Infinity` to return every prefix match.
      */
     zipLimit?: number;
     /**
@@ -145,6 +147,10 @@ type BuildIndexOptions = {
      * Validate that every consumed field has the expected runtime type and throw a
      * descriptive `TypeError` if not. Defaults to `true`; set to `false` to skip
      * the checks when the data source is already trusted (e.g. the bundled index).
+     *
+     * `false` also skips the row, name-length and posting-budget limits that bound
+     * memory and CPU use, so never disable it for data you do not control
+     * (user uploads, third-party APIs, a CMS editable by others).
      */
     validate?: boolean;
 };
@@ -161,10 +167,9 @@ type RawData = {
  * private datasets) who want to fail fast with a descriptive error instead of an
  * opaque crash deep inside the normalizer, or silent `"undefined"` labels.
  *
- * Throws a `TypeError` (message prefixed `[thaizip]`) naming the table, the
- * offending record's `id`, the field, and the actual `typeof` on the first bad
- * field found. Does not filter soft-deleted rows first — validates every row as
- * provided.
+ * Throws a `TypeError` for bad shapes or duplicate IDs and a `RangeError` for
+ * resource limits (all prefixed `[thaizip]`). Does not filter soft-deleted rows
+ * before shape checks — validates every row as provided.
  *
  * IDs must be unique within each table (duplicates throw, they are not silently
  * discarded — two suggestions sharing an id would resolve to the wrong record).

package/dist/index.js

@@ -33,9 +33,19 @@ function extractTrigramsNormalized(text) {
 }
 
 // src/core/indexer.ts
+var MAX_TOTAL_ROWS = 5e4;
+var MAX_NAME_LENGTH = 256;
+var MAX_ZIP_LENGTH = 32;
+var MAX_POSTING_BUDGET = 1e6;
 function typeErr(table, id, field, expected, value) {
   return new TypeError(`[thaizip] ${table} ${id}: expected ${expected} for ${field}, got ${typeof value}`);
 }
+function assertId(table, id, field, value) {
+  if (typeof value !== "number") throw typeErr(table, id, field, "number", value);
+  if (!Number.isSafeInteger(value)) {
+    throw new TypeError(`[thaizip] ${table} ${id}: expected safe integer for ${field}, got ${value}`);
+  }
+}
 function typeName(value) {
   return value === null ? "null" : typeof value;
 }
@@ -53,47 +63,79 @@ function validateRawData(data) {
   if (!Array.isArray(tambons)) {
     throw new TypeError(`[thaizip] RawData.tambons must be an array, got ${typeName(tambons)}`);
   }
-  const seenProvinceIds = /* @__PURE__ */ new Set();
+  if (provinces.length + amphures.length + tambons.length > MAX_TOTAL_ROWS) {
+    throw new RangeError(`[thaizip] RawData row limit exceeded (${MAX_TOTAL_ROWS})`);
+  }
+  const provinceById = /* @__PURE__ */ new Map();
   for (let i = 0; i < provinces.length; i++) {
     const p = provinces[i];
     if (p === null || typeof p !== "object") {
       throw new TypeError(`[thaizip] province[${i}]: expected object, got ${typeName(p)}`);
     }
-    if (typeof p.id !== "number") throw typeErr("province", p.id, "id", "number", p.id);
-    if (seenProvinceIds.has(p.id)) throw new TypeError(`[thaizip] duplicate province id: ${p.id}`);
-    seenProvinceIds.add(p.id);
+    assertId("province", p.id, "id", p.id);
+    if (provinceById.has(p.id)) throw new TypeError(`[thaizip] duplicate province id: ${p.id}`);
+    provinceById.set(p.id, p);
     if (typeof p.name_th !== "string") throw typeErr("province", p.id, "name_th", "string", p.name_th);
     if (typeof p.name_en !== "string") throw typeErr("province", p.id, "name_en", "string", p.name_en);
+    if (p.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] province ${p.id}: name_th too long`);
+    if (p.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] province ${p.id}: name_en too long`);
   }
-  const seenAmphureIds = /* @__PURE__ */ new Set();
+  const amphureById = /* @__PURE__ */ new Map();
   for (let i = 0; i < amphures.length; i++) {
     const a = amphures[i];
     if (a === null || typeof a !== "object") {
       throw new TypeError(`[thaizip] amphure[${i}]: expected object, got ${typeName(a)}`);
     }
-    if (typeof a.id !== "number") throw typeErr("amphure", a.id, "id", "number", a.id);
-    if (seenAmphureIds.has(a.id)) throw new TypeError(`[thaizip] duplicate amphure id: ${a.id}`);
-    seenAmphureIds.add(a.id);
+    assertId("amphure", a.id, "id", a.id);
+    if (amphureById.has(a.id)) throw new TypeError(`[thaizip] duplicate amphure id: ${a.id}`);
+    amphureById.set(a.id, a);
     if (typeof a.name_th !== "string") throw typeErr("amphure", a.id, "name_th", "string", a.name_th);
     if (typeof a.name_en !== "string") throw typeErr("amphure", a.id, "name_en", "string", a.name_en);
-    if (typeof a.province_id !== "number") throw typeErr("amphure", a.id, "province_id", "number", a.province_id);
+    assertId("amphure", a.id, "province_id", a.province_id);
+    if (a.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] amphure ${a.id}: name_th too long`);
+    if (a.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] amphure ${a.id}: name_en too long`);
   }
   const seenTambonIds = /* @__PURE__ */ new Set();
+  const provinceTrigramCounts = /* @__PURE__ */ new Map();
+  const amphureTrigramCounts = /* @__PURE__ */ new Map();
+  let postingBudget = 0;
   for (let i = 0; i < tambons.length; i++) {
     const t = tambons[i];
     if (t === null || typeof t !== "object") {
       throw new TypeError(`[thaizip] tambon[${i}]: expected object, got ${typeName(t)}`);
     }
-    if (typeof t.id !== "number") throw typeErr("tambon", t.id, "id", "number", t.id);
+    assertId("tambon", t.id, "id", t.id);
     if (seenTambonIds.has(t.id)) throw new TypeError(`[thaizip] duplicate tambon id: ${t.id}`);
     seenTambonIds.add(t.id);
     if (typeof t.name_th !== "string") throw typeErr("tambon", t.id, "name_th", "string", t.name_th);
     if (typeof t.name_en !== "string") throw typeErr("tambon", t.id, "name_en", "string", t.name_en);
+    if (t.name_th.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: name_th too long`);
+    if (t.name_en.length > MAX_NAME_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: name_en too long`);
     const zipType = typeof t.zip_code;
     if (zipType !== "string" && zipType !== "number") {
       throw typeErr("tambon", t.id, "zip_code", "string or number", t.zip_code);
     }
-    if (typeof t.amphure_id !== "number") throw typeErr("tambon", t.id, "amphure_id", "number", t.amphure_id);
+    assertId("tambon", t.id, "amphure_id", t.amphure_id);
+    const zipLength = String(t.zip_code).length;
+    if (zipLength > MAX_ZIP_LENGTH) throw new RangeError(`[thaizip] tambon ${t.id}: zip_code too long`);
+    if (t.deleted_at) continue;
+    const amphure = amphureById.get(t.amphure_id);
+    const province = amphure && provinceById.get(amphure.province_id);
+    if (!amphure || amphure.deleted_at || !province || province.deleted_at) continue;
+    let provinceGrams = provinceTrigramCounts.get(province.id);
+    if (provinceGrams === void 0) {
+      provinceGrams = combinedTrigrams(province.name_th, province.name_en).size;
+      provinceTrigramCounts.set(province.id, provinceGrams);
+    }
+    let amphureGrams = amphureTrigramCounts.get(amphure.id);
+    if (amphureGrams === void 0) {
+      amphureGrams = combinedTrigrams(amphure.name_th, amphure.name_en).size;
+      amphureTrigramCounts.set(amphure.id, amphureGrams);
+    }
+    postingBudget += t.name_th.length + t.name_en.length + zipLength + provinceGrams + amphureGrams;
+    if (postingBudget > MAX_POSTING_BUDGET) {
+      throw new RangeError(`[thaizip] estimated trigram posting budget exceeded (${MAX_POSTING_BUDGET})`);
+    }
   }
 }
 function addTrigrams(map, trigrams, idx) {
@@ -124,13 +166,7 @@ function buildThaiAddressIndex(data, options) {
     amphures.filter((a) => !a.deleted_at).map((a) => [a.id, a])
   );
   const provTrigrams = /* @__PURE__ */ new Map();
-  for (const [id, prov] of provMap) {
-    provTrigrams.set(id, combinedTrigrams(prov.name_th, prov.name_en));
-  }
   const ampTrigrams = /* @__PURE__ */ new Map();
-  for (const [id, amp] of ampMap) {
-    ampTrigrams.set(id, combinedTrigrams(amp.name_th, amp.name_en));
-  }
   const records = [];
   const map = /* @__PURE__ */ new Map();
   const zipIndex = /* @__PURE__ */ new Map();
@@ -180,8 +216,18 @@ function buildThaiAddressIndex(data, options) {
     normTambonEn.push(normEn);
     addTrigrams(map, extractTrigramsNormalized(normEn), idx);
     addTrigrams(map, extractTrigrams(record.zipCode), idx);
-    addTrigrams(map, provTrigrams.get(province.id), idx);
-    addTrigrams(map, ampTrigrams.get(amphure.id), idx);
+    let provinceGrams = provTrigrams.get(province.id);
+    if (!provinceGrams) {
+      provinceGrams = combinedTrigrams(province.name_th, province.name_en);
+      provTrigrams.set(province.id, provinceGrams);
+    }
+    let amphureGrams = ampTrigrams.get(amphure.id);
+    if (!amphureGrams) {
+      amphureGrams = combinedTrigrams(amphure.name_th, amphure.name_en);
+      ampTrigrams.set(amphure.id, amphureGrams);
+    }
+    addTrigrams(map, provinceGrams, idx);
+    addTrigrams(map, amphureGrams, idx);
   }
   const sortedZipKeys = [...zipIndex.keys()].sort();
   const sortedZipPostings = sortedZipKeys.map((z) => zipIndex.get(z));
@@ -379,7 +425,7 @@ function searchThaiAddress(index, query, options) {
   if (ZIP_CODE_RE.test(normalized)) {
     if (normalized.length < 2) return [];
     return lookupByZipCode(index, normalized, {
-      zipLimit: options?.zipLimit ?? Infinity
+      zipLimit: options?.zipLimit ?? (normalized.length >= 5 ? Infinity : limit)
     });
   }
   if (normalized.length < 3) return [];
@@ -531,5 +577,3 @@ function resolveThaiAddress(record) {
 }
 
 export { applyRomanizationAliases, buildThaiAddressIndex, formatThaiAddressSuggestion, listAmphures, listProvinces, listTambons, lookupByZipCode, normalizeThaiAddressText, resolveThaiAddress, searchThaiAddress, validateRawData };
-//# sourceMappingURL=index.js.map
-//# sourceMappingURL=index.js.map

package/dist/react.cjs

@@ -225,7 +225,7 @@ function searchThaiAddress(index, query, options) {
   if (ZIP_CODE_RE.test(normalized)) {
     if (normalized.length < 2) return [];
     return lookupByZipCode(index, normalized, {
-      zipLimit: options?.zipLimit ?? Infinity
+      zipLimit: options?.zipLimit ?? (normalized.length >= 5 ? Infinity : limit)
     });
   }
   if (normalized.length < 3) return [];
@@ -402,5 +402,3 @@ function useThaiAddressAutocomplete(options) {
 }
 
 exports.useThaiAddressAutocomplete = useThaiAddressAutocomplete;
-//# sourceMappingURL=react.cjs.map
-//# sourceMappingURL=react.cjs.map

package/dist/react.d.cts

@@ -79,6 +79,7 @@ type UseThaiAddressAutocompleteOptions = {
     limit?: number;
     debounce?: number;
     threshold?: number;
+    /** Maximum results for zip code queries; see `SearchOptions.zipLimit` for the default. */
     zipLimit?: number;
     /** Seed value for `query` on first render. Defaults to `''`. */
     initialQuery?: string;

package/dist/react.d.ts

@@ -79,6 +79,7 @@ type UseThaiAddressAutocompleteOptions = {
     limit?: number;
     debounce?: number;
     threshold?: number;
+    /** Maximum results for zip code queries; see `SearchOptions.zipLimit` for the default. */
     zipLimit?: number;
     /** Seed value for `query` on first render. Defaults to `''`. */
     initialQuery?: string;

package/dist/react.js

@@ -223,7 +223,7 @@ function searchThaiAddress(index, query, options) {
   if (ZIP_CODE_RE.test(normalized)) {
     if (normalized.length < 2) return [];
     return lookupByZipCode(index, normalized, {
-      zipLimit: options?.zipLimit ?? Infinity
+      zipLimit: options?.zipLimit ?? (normalized.length >= 5 ? Infinity : limit)
     });
   }
   if (normalized.length < 3) return [];
@@ -400,5 +400,3 @@ function useThaiAddressAutocomplete(options) {
 }
 
 export { useThaiAddressAutocomplete };
-//# sourceMappingURL=react.js.map
-//# sourceMappingURL=react.js.map

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "thaizip",
-  "version": "0.7.5",
+  "version": "0.7.7",
   "description": "Fast, fuzzy Thai address autocomplete library (headless + React)",
   "keywords": [
     "thai",