thaizip 0.5.1 → 0.7.2

package/dist/data.js

@@ -25,8 +25,8 @@ var init_defaultData = __esm({
 });
 
 // src/core/normalizer.ts
-var TONE_MARKS = /[\u0E48-\u0E4B\u0E47\u0E4C]/g;
-var ADDRESS_PREFIXES = /^(จังหวัด|อำเภอ|ตำบล|แขวง|เขต)/;
+var TONE_MARKS = /[่-๋็์]/g;
+var ADDRESS_PREFIXES = /^(จังหวัด|อำเภอ|ตำบล|แขวง|เขต|จ\.|อ\.|ต\.|ข\.)/;
 function normalizeThaiAddressText(input) {
   if (!input) return "";
   let text = input.trim();
@@ -54,7 +54,56 @@ function extractTrigramsNormalized(text) {
 }
 
 // src/core/indexer.ts
-function buildThaiAddressIndex(data) {
+function typeErr(table, id, field, expected, value) {
+  return new TypeError(`[thaizip] ${table} ${id}: expected ${expected} for ${field}, got ${typeof value}`);
+}
+function validateRawData(data) {
+  const { provinces, amphures, tambons } = data;
+  for (let i = 0; i < provinces.length; i++) {
+    const p2 = provinces[i];
+    if (typeof p2.id !== "number") throw typeErr("province", p2.id, "id", "number", p2.id);
+    if (typeof p2.name_th !== "string") throw typeErr("province", p2.id, "name_th", "string", p2.name_th);
+    if (typeof p2.name_en !== "string") throw typeErr("province", p2.id, "name_en", "string", p2.name_en);
+  }
+  for (let i = 0; i < amphures.length; i++) {
+    const a2 = amphures[i];
+    if (typeof a2.id !== "number") throw typeErr("amphure", a2.id, "id", "number", a2.id);
+    if (typeof a2.name_th !== "string") throw typeErr("amphure", a2.id, "name_th", "string", a2.name_th);
+    if (typeof a2.name_en !== "string") throw typeErr("amphure", a2.id, "name_en", "string", a2.name_en);
+    if (typeof a2.province_id !== "number") throw typeErr("amphure", a2.id, "province_id", "number", a2.province_id);
+  }
+  for (let i = 0; i < tambons.length; i++) {
+    const t2 = tambons[i];
+    if (typeof t2.id !== "number") throw typeErr("tambon", t2.id, "id", "number", t2.id);
+    if (typeof t2.name_th !== "string") throw typeErr("tambon", t2.id, "name_th", "string", t2.name_th);
+    if (typeof t2.name_en !== "string") throw typeErr("tambon", t2.id, "name_en", "string", t2.name_en);
+    const zipType = typeof t2.zip_code;
+    if (zipType !== "string" && zipType !== "number") {
+      throw typeErr("tambon", t2.id, "zip_code", "string or number", t2.zip_code);
+    }
+    if (typeof t2.amphure_id !== "number") throw typeErr("tambon", t2.id, "amphure_id", "number", t2.amphure_id);
+  }
+}
+function addTrigrams(map, trigrams, idx) {
+  for (const trigram of trigrams) {
+    let set = map.get(trigram);
+    if (!set) {
+      set = /* @__PURE__ */ new Set();
+      map.set(trigram, set);
+    }
+    set.add(idx);
+  }
+}
+function combinedTrigrams(nameTh, nameEn) {
+  const s = /* @__PURE__ */ new Set();
+  for (const t2 of extractTrigrams(nameTh)) s.add(t2);
+  for (const t2 of extractTrigrams(nameEn)) s.add(t2);
+  return s;
+}
+function buildThaiAddressIndex(data, options) {
+  if (options?.validate !== false) {
+    validateRawData(data);
+  }
   const { provinces, amphures, tambons } = data;
   const provMap = new Map(
     provinces.filter((p2) => !p2.deleted_at).map((p2) => [p2.id, p2])
@@ -62,20 +111,33 @@ function buildThaiAddressIndex(data) {
   const ampMap = new Map(
     amphures.filter((a2) => !a2.deleted_at).map((a2) => [a2.id, a2])
   );
+  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();
+  const normTambon = [];
+  const normTambonEn = [];
+  const byProvince = /* @__PURE__ */ new Map();
+  const byAmphure = /* @__PURE__ */ new Map();
   for (const tambon of tambons) {
     if (tambon.deleted_at) continue;
     const amphure = ampMap.get(tambon.amphure_id);
     if (!amphure) {
-      console.warn(
-        `[thaizip] skipping tambon id=${tambon.id} (${tambon.name_th}): amphure_id=${tambon.amphure_id} not found or deleted`
-      );
+      if (options?.onSkip) options.onSkip(tambon);
       continue;
     }
     const province = provMap.get(amphure.province_id);
-    if (!province) continue;
+    if (!province) {
+      if (options?.onSkip) options.onSkip(tambon);
+      continue;
+    }
     const record = {
       provinceId: province.id,
       provinceNameTh: province.name_th,
@@ -93,43 +155,55 @@ function buildThaiAddressIndex(data) {
     const existing = zipIndex.get(record.zipCode);
     if (existing) existing.push(idx);
     else zipIndex.set(record.zipCode, [idx]);
-    const fields = [
-      record.tambonNameTh,
-      record.tambonNameEn,
-      record.amphureNameTh,
-      record.amphureNameEn,
-      record.provinceNameTh,
-      record.provinceNameEn,
-      record.zipCode
-    ];
-    for (const field of fields) {
-      for (const trigram of extractTrigrams(field)) {
-        let set = map.get(trigram);
-        if (!set) {
-          set = /* @__PURE__ */ new Set();
-          map.set(trigram, set);
-        }
-        set.add(idx);
-      }
-    }
+    const provList = byProvince.get(province.id);
+    if (provList) provList.push(idx);
+    else byProvince.set(province.id, [idx]);
+    const ampList = byAmphure.get(amphure.id);
+    if (ampList) ampList.push(idx);
+    else byAmphure.set(amphure.id, [idx]);
+    const normTh = normalizeThaiAddressText(record.tambonNameTh);
+    normTambon.push(normTh);
+    addTrigrams(map, extractTrigramsNormalized(normTh), idx);
+    const normEn = normalizeThaiAddressText(record.tambonNameEn);
+    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);
   }
-  return { map, records, zipIndex };
+  return { map, records, zipIndex, normTambon, normTambonEn, byProvince, byAmphure };
 }
 
 // src/data/loader.ts
 var cached = null;
+var inflightPromise = null;
+var loadGeneration = 0;
 function clearDefaultIndex() {
   cached = null;
+  inflightPromise = null;
+  loadGeneration++;
 }
 async function loadDefaultIndex() {
   if (cached) return cached;
-  const { p: p2, a: a2, t: t2 } = await Promise.resolve().then(() => (init_defaultData(), defaultData_exports));
-  cached = buildThaiAddressIndex({
-    provinces: p2.map(([id, name_th, name_en]) => ({ id, name_th, name_en, geography_id: 0, deleted_at: null })),
-    amphures: a2.map(([id, name_th, name_en, province_id]) => ({ id, name_th, name_en, province_id, deleted_at: null })),
-    tambons: t2.map(([id, name_th, name_en, amphure_id, zip_code]) => ({ id, name_th, name_en, amphure_id, zip_code: Number(zip_code), deleted_at: null }))
-  });
-  return cached;
+  if (!inflightPromise) {
+    const gen = loadGeneration;
+    inflightPromise = Promise.resolve().then(() => (init_defaultData(), defaultData_exports)).then(({ p: p2, a: a2, t: t2 }) => {
+      const index = buildThaiAddressIndex({
+        provinces: p2.map(([id, name_th, name_en]) => ({ id, name_th, name_en, geography_id: 0, deleted_at: null })),
+        amphures: a2.map(([id, name_th, name_en, province_id]) => ({ id, name_th, name_en, province_id, deleted_at: null })),
+        tambons: t2.map(([id, name_th, name_en, amphure_id, zip_code]) => ({ id, name_th, name_en, amphure_id, zip_code, deleted_at: null }))
+      }, { validate: false });
+      if (loadGeneration === gen) {
+        cached = index;
+        inflightPromise = null;
+      }
+      return index;
+    }).catch((err) => {
+      if (loadGeneration === gen) inflightPromise = null;
+      throw err;
+    });
+  }
+  return inflightPromise;
 }
 
 export { clearDefaultIndex, loadDefaultIndex };