gs-search 0.1.7 → 0.1.8

package/lib/core.cjs

@@ -181,21 +181,129 @@ function murmur3_32(str, h = 305419896) {
   const remainder = len & 3;
   return remainder > 0 && (remainder >= 3 && (k1 ^= (str.charCodeAt(i + 2) & 255) << 16), remainder >= 2 && (k1 ^= (str.charCodeAt(i + 1) & 255) << 8), remainder >= 1 && (k1 ^= str.charCodeAt(i) & 255), k1 = (k1 & 65535) * 3432918353 + (((k1 >>> 16) * 3432918353 & 65535) << 16) & 4294967295, k1 = k1 << 15 | k1 >>> 17, k1 = (k1 & 65535) * 461845907 + (((k1 >>> 16) * 461845907 & 65535) << 16) & 4294967295, h ^= k1), h ^= len, h ^= h >>> 16, h = (h & 65535) * 2246822507 + (((h >>> 16) * 2246822507 & 65535) << 16) & 4294967295, h ^= h >>> 13, h = (h & 65535) * 3266489909 + (((h >>> 16) * 3266489909 & 65535) << 16) & 4294967295, h ^= h >>> 16, h >>> 0;
 }
+function murmur3_64(str, seed = 305419896) {
+  const len = str.length, nBlocks = len >> 3;
+  let h1 = BigInt(seed), h2 = BigInt(seed);
+  const c1 = 0x87c37b91114253d5n, c2 = 0x4cf5ad432745937fn;
+  let i = 0;
+  for (; i < nBlocks; ) {
+    let k12 = BigInt(str.charCodeAt(i) & 255) | BigInt(str.charCodeAt(++i) & 255) << 8n | BigInt(str.charCodeAt(++i) & 255) << 16n | BigInt(str.charCodeAt(++i) & 255) << 24n | BigInt(str.charCodeAt(++i) & 255) << 32n | BigInt(str.charCodeAt(++i) & 255) << 40n | BigInt(str.charCodeAt(++i) & 255) << 48n | BigInt(str.charCodeAt(++i) & 255) << 56n;
+    ++i;
+    let k22 = BigInt(str.charCodeAt(i) & 255) | BigInt(str.charCodeAt(++i) & 255) << 8n | BigInt(str.charCodeAt(++i) & 255) << 16n | BigInt(str.charCodeAt(++i) & 255) << 24n | BigInt(str.charCodeAt(++i) & 255) << 32n | BigInt(str.charCodeAt(++i) & 255) << 40n | BigInt(str.charCodeAt(++i) & 255) << 48n | BigInt(str.charCodeAt(++i) & 255) << 56n;
+    ++i, k12 = k12 * c1 % 2n ** 64n, k12 = (k12 << 31n | k12 >> 33n) % 2n ** 64n, k12 = k12 * c2 % 2n ** 64n, h1 ^= k12, h1 = (h1 << 27n | h1 >> 37n) % 2n ** 64n, h1 = (h1 + h2) % 2n ** 64n, h1 = (h1 * 5n + 0x52dce729n) % 2n ** 64n, k22 = k22 * c2 % 2n ** 64n, k22 = (k22 << 33n | k22 >> 31n) % 2n ** 64n, k22 = k22 * c1 % 2n ** 64n, h2 ^= k22, h2 = (h2 << 31n | h2 >> 33n) % 2n ** 64n, h2 = (h2 + h1) % 2n ** 64n, h2 = (h2 * 5n + 0x38495ab5n) % 2n ** 64n;
+  }
+  let k1 = 0n, k2 = 0n;
+  const remainder = len & 7;
+  return remainder > 0 && (remainder >= 8 && (k2 ^= BigInt(str.charCodeAt(i + 7) & 255) << 56n), remainder >= 7 && (k2 ^= BigInt(str.charCodeAt(i + 6) & 255) << 48n), remainder >= 6 && (k2 ^= BigInt(str.charCodeAt(i + 5) & 255) << 40n), remainder >= 5 && (k2 ^= BigInt(str.charCodeAt(i + 4) & 255) << 32n), remainder >= 4 && (k1 ^= BigInt(str.charCodeAt(i + 3) & 255) << 24n), remainder >= 3 && (k1 ^= BigInt(str.charCodeAt(i + 2) & 255) << 16n), remainder >= 2 && (k1 ^= BigInt(str.charCodeAt(i + 1) & 255) << 8n), remainder >= 1 && (k1 ^= BigInt(str.charCodeAt(i) & 255)), k1 = k1 * c1 % 2n ** 64n, k1 = (k1 << 31n | k1 >> 33n) % 2n ** 64n, k1 = k1 * c2 % 2n ** 64n, h1 ^= k1, k2 = k2 * c2 % 2n ** 64n, k2 = (k2 << 33n | k2 >> 31n) % 2n ** 64n, k2 = k2 * c1 % 2n ** 64n, h2 ^= k2), h1 ^= BigInt(len), h2 ^= BigInt(len), h1 = (h1 + h2) % 2n ** 64n, h2 = (h2 + h1) % 2n ** 64n, h1 = (h1 ^ h1 >> 33n) * 0xff51afd7ed558ccdn, h1 = (h1 ^ h1 >> 33n) * 0xc4ceb9fe1a85ec53n, h1 = h1 ^ h1 >> 33n, h2 = (h2 ^ h2 >> 33n) * 0xff51afd7ed558ccdn, h2 = (h2 ^ h2 >> 33n) * 0xc4ceb9fe1a85ec53n, h2 = h2 ^ h2 >> 33n, (h2 << 64n | h1) & 0xffffffffffffffffn;
+}
+class Murmur3_32 {
+  seed;
+  /**
+   * 构造函数
+   * @param seed 种子值，默认为0x12345678
+   */
+  constructor(seed = 305419896) {
+    this.seed = seed;
+  }
+  /**
+   * 计算字符串的32位哈希值
+   * @param str 输入字符串
+   * @returns 32位哈希值（number类型）
+   */
+  hash(str) {
+    return murmur3_32(str, this.seed);
+  }
+}
+class Murmur3_64 {
+  seed;
+  /**
+   * 构造函数
+   * @param seed 种子值，默认为0x12345678
+   */
+  constructor(seed = 305419896) {
+    this.seed = seed;
+  }
+  /**
+   * 计算字符串的64位哈希值
+   * @param str 输入字符串
+   * @returns 64位哈希值（bigint类型）
+   */
+  hash(str) {
+    return murmur3_64(str, this.seed);
+  }
+}
+class Murmur3_128 {
+  seed;
+  /**
+   * 构造函数
+   * @param seed 种子值，默认为0x12345678
+   */
+  constructor(seed = 305419896) {
+    this.seed = seed;
+  }
+  /**
+   * 计算字符串的128位哈希值
+   * @param str 输入字符串
+   * @returns 128位哈希值（bigint类型）
+   */
+  hash(str) {
+    const hash1 = murmur3_64(str, this.seed);
+    return murmur3_64(str + str, this.seed ^ 1515870810) << 64n | hash1;
+  }
+}
+class Murmur3HashFactory {
+  /**
+   * 创建32位哈希算法实例
+   * @returns 32位哈希算法实例
+   */
+  create32() {
+    return new Murmur3_32();
+  }
+  /**
+   * 创建64位哈希算法实例
+   * @returns 64位哈希算法实例
+   */
+  create64() {
+    return new Murmur3_64();
+  }
+  /**
+   * 创建128位哈希算法实例
+   * @returns 128位哈希算法实例
+   */
+  create128() {
+    return new Murmur3_128();
+  }
+}
+const defaultHashFactory = new Murmur3HashFactory();
 class IndexSegment {
   #filename;
   #storage;
   #buffer = null;
   #view = null;
+  #hashAlgorithm;
   /**
-   * 使用MurmurHash3计算字符串哈希值
+   * 构造函数
+   * @param filename 文件名
+   * @param storage 存储接口
+   * @param hashAlgorithm 哈希算法实例，默认为Murmur3_32
+   */
+  constructor(filename, storage, hashAlgorithm = new Murmur3_32()) {
+    this.#filename = filename, this.#storage = storage, this.#hashAlgorithm = hashAlgorithm;
+  }
+  /**
+   * 使用当前哈希算法计算字符串哈希值
    * @param str 要哈希的字符串
    * @returns 32位无符号哈希值
    */
-  static hash(str) {
-    return murmur3_32(str);
+  hash(str) {
+    return this.#hashAlgorithm.hash(str);
   }
-  constructor(filename, storage) {
-    this.#filename = filename, this.#storage = storage;
+  /**
+   * 设置哈希算法
+   * @param hashAlgorithm 新的哈希算法实例
+   */
+  setHashAlgorithm(hashAlgorithm) {
+    this.#hashAlgorithm = hashAlgorithm;
   }
   async loadIndex() {
     return this.#buffer ? !0 : (this.#buffer = await this.#storage.read(this.#filename), this.#buffer ? (this.#view = new DataView(this.#buffer), !0) : !1);
@@ -206,7 +314,7 @@ class IndexSegment {
       const uniqueTokens = /* @__PURE__ */ new Map();
       for (const token of doc.tokens)
         uniqueTokens.has(token) || (uniqueTokens.set(token, !0), tokenMap.has(token) || tokenMap.set(token, {
-          hash: IndexSegment.hash(token),
+          hash: this.hash(token),
           postings: []
         }), tokenMap.get(token).postings.push(doc.id));
     }
@@ -222,8 +330,8 @@ class IndexSegment {
     const headerSize = 12, dictSize = entries.length * 20, postingsSize = totalPostings * 4, tokensOffset = headerSize + dictSize + postingsSize, totalSize = tokensOffset + totalTokensSize, buffer = new ArrayBuffer(totalSize), view = new DataView(buffer);
     view.setUint32(0, 1229866072), view.setUint32(4, entries.length, !0), view.setUint32(8, tokensOffset, !0);
     let currentDictOffset = headerSize, currentPostingsOffset = headerSize + dictSize, currentTokenOffset = tokensOffset;
-    for (const [token, { hash: hash2, postings }] of entries) {
-      view.setUint32(currentDictOffset, hash2, !0);
+    for (const [token, { hash, postings }] of entries) {
+      view.setUint32(currentDictOffset, hash, !0);
       const tokenBytes = encoder.encode(token);
       view.setUint32(currentDictOffset + 4, tokenBytes.length, !0), view.setUint32(currentDictOffset + 8, currentTokenOffset, !0), view.setUint32(currentDictOffset + 12, currentPostingsOffset, !0), view.setUint32(currentDictOffset + 16, postings.length, !0), currentDictOffset += 20;
       for (let i = 0; i < postings.length; i++)
@@ -236,7 +344,7 @@ class IndexSegment {
   }
   search(term) {
     if (!this.#view || !this.#buffer) return [];
-    const h = IndexSegment.hash(term), count = this.#view.getUint32(4, !0);
+    const h = this.hash(term), count = this.#view.getUint32(4, !0);
     let left = 0, right = count - 1;
     const headerSize = 12, entrySize = 20, decoder = new TextDecoder();
     for (; left <= right; ) {
@@ -277,6 +385,116 @@ class IndexSegment {
     return [];
   }
 }
+class IndexSegment64 {
+  #filename;
+  #storage;
+  #buffer = null;
+  #view = null;
+  #hashAlgorithm;
+  /**
+   * 构造函数
+   * @param filename 文件名
+   * @param storage 存储接口
+   * @param hashAlgorithm 哈希算法实例，默认为Murmur3_64
+   */
+  constructor(filename, storage, hashAlgorithm = new Murmur3_64()) {
+    this.#filename = filename, this.#storage = storage, this.#hashAlgorithm = hashAlgorithm;
+  }
+  /**
+   * 使用当前哈希算法计算字符串哈希值
+   * @param str 要哈希的字符串
+   * @returns 64位无符号哈希值
+   */
+  hash(str) {
+    return this.#hashAlgorithm.hash(str);
+  }
+  /**
+   * 设置哈希算法
+   * @param hashAlgorithm 新的哈希算法实例
+   */
+  setHashAlgorithm(hashAlgorithm) {
+    this.#hashAlgorithm = hashAlgorithm;
+  }
+  async loadIndex() {
+    return this.#buffer ? !0 : (this.#buffer = await this.#storage.read(this.#filename), this.#buffer ? (this.#view = new DataView(this.#buffer), !0) : !1);
+  }
+  async buildAndSave(docs) {
+    const tokenMap = /* @__PURE__ */ new Map();
+    for (const doc of docs) {
+      const uniqueTokens = /* @__PURE__ */ new Map();
+      for (const token of doc.tokens)
+        uniqueTokens.has(token) || (uniqueTokens.set(token, !0), tokenMap.has(token) || tokenMap.set(token, {
+          hash: this.hash(token),
+          postings: []
+        }), tokenMap.get(token).postings.push(doc.id));
+    }
+    const entries = Array.from(tokenMap.entries());
+    entries.sort(([a, ah], [b, bh]) => ah.hash !== bh.hash ? ah.hash > bh.hash ? 1 : -1 : a.localeCompare(b));
+    const encoder = new TextEncoder();
+    let totalPostings = 0, totalTokensSize = 0;
+    for (const [token, { postings }] of entries) {
+      totalPostings += postings.length;
+      const bytes = encoder.encode(token);
+      totalTokensSize += bytes.length + 1;
+    }
+    const headerSize = 16, dictSize = entries.length * 28, postingsSize = totalPostings * 4, tokensOffset = headerSize + dictSize + postingsSize, totalSize = tokensOffset + totalTokensSize, buffer = new ArrayBuffer(totalSize), view = new DataView(buffer);
+    view.setUint32(0, 1229866072), view.setUint32(4, entries.length, !0), view.setUint32(8, tokensOffset, !0), view.setUint32(12, 64, !0);
+    let currentDictOffset = headerSize, currentPostingsOffset = headerSize + dictSize, currentTokenOffset = tokensOffset;
+    for (const [token, { hash, postings }] of entries) {
+      view.setBigUint64(currentDictOffset, hash, !0);
+      const tokenBytes = encoder.encode(token);
+      view.setUint32(currentDictOffset + 8, tokenBytes.length, !0), view.setUint32(currentDictOffset + 12, currentTokenOffset, !0), view.setUint32(currentDictOffset + 16, currentPostingsOffset, !0), view.setUint32(currentDictOffset + 20, postings.length, !0), currentDictOffset += 28;
+      for (let i = 0; i < postings.length; i++)
+        view.setUint32(currentPostingsOffset, postings[i], !0), currentPostingsOffset += 4;
+      for (let i = 0; i < tokenBytes.length; i++)
+        view.setUint8(currentTokenOffset++, tokenBytes[i]);
+      view.setUint8(currentTokenOffset++, 0);
+    }
+    await this.#storage.write(this.#filename, buffer), this.#buffer = buffer, this.#view = view;
+  }
+  search(term) {
+    if (!this.#view || !this.#buffer) return [];
+    const h = this.hash(term), count = this.#view.getUint32(4, !0);
+    let left = 0, right = count - 1;
+    const headerSize = 16, entrySize = 28, decoder = new TextDecoder();
+    for (; left <= right; ) {
+      const mid = left + right >>> 1, entryPos = headerSize + mid * entrySize, entryHash = this.#view.getBigUint64(entryPos, !0);
+      if (entryHash < h)
+        left = mid + 1;
+      else if (entryHash > h)
+        right = mid - 1;
+      else {
+        if (!(mid > 0 && this.#view.getBigUint64(headerSize + (mid - 1) * entrySize, !0) === h || mid < count - 1 && this.#view.getBigUint64(headerSize + (mid + 1) * entrySize, !0) === h)) {
+          const postingsOffset = this.#view.getUint32(headerSize + mid * entrySize + 16, !0), postingsLen = this.#view.getUint32(headerSize + mid * entrySize + 20, !0), result = [];
+          for (let j = 0; j < postingsLen; j++)
+            result.push(this.#view.getUint32(postingsOffset + j * 4, !0));
+          return result;
+        }
+        let firstMatch = mid;
+        for (; firstMatch > 0; ) {
+          const prevPos = headerSize + (firstMatch - 1) * entrySize;
+          if (this.#view.getBigUint64(prevPos, !0) === h)
+            firstMatch--;
+          else
+            break;
+        }
+        for (let i = firstMatch; i < count; i++) {
+          const checkPos = headerSize + i * entrySize;
+          if (this.#view.getBigUint64(checkPos, !0) !== h) break;
+          const tokenLen = this.#view.getUint32(checkPos + 8, !0), tokenOffset = this.#view.getUint32(checkPos + 12, !0), tokenBuffer = new Uint8Array(this.#buffer, tokenOffset, tokenLen);
+          if (decoder.decode(tokenBuffer) === term) {
+            const postingsOffset = this.#view.getUint32(checkPos + 16, !0), postingsLen = this.#view.getUint32(checkPos + 20, !0), result = [];
+            for (let j = 0; j < postingsLen; j++)
+              result.push(this.#view.getUint32(postingsOffset + j * 4, !0));
+            return result;
+          }
+        }
+        return [];
+      }
+    }
+    return [];
+  }
+}
 const defaultTokenize = ({ text }) => {
   try {
     if (typeof Intl < "u" && typeof Intl.Segmenter == "function" && typeof Array.from == "function") {
@@ -295,6 +513,9 @@ class SearchEngine {
   #segments;
   #initialized = !1;
   #config;
+  #isHash64Bit = !0;
+  #hashAlgorithm32;
+  #hashAlgorithm64;
   // 批处理状态
   #inBatch = !1;
   #pendingTokenCounts = { word: 0, char: 0 };
@@ -306,12 +527,19 @@ class SearchEngine {
       minCharTokenSave: 0,
       indexingTokenizer: config.indexingTokenizer || defaultTokenize,
       ...config
-    }, (this.#config.minWordTokenSave || 0) >= (this.#config.wordSegmentTokenThreshold || 1e5))
+    }, this.#processHashAlgorithmConfig(), (this.#config.minWordTokenSave || 0) >= (this.#config.wordSegmentTokenThreshold || 1e5))
       throw new Error("minWordTokenSave must be less than wordSegmentTokenThreshold");
     if ((this.#config.minCharTokenSave || 0) >= (this.#config.charSegmentTokenThreshold || 5e5))
       throw new Error("minCharTokenSave must be less than charSegmentTokenThreshold");
     this.#storage = config.storage, this.#meta = new MetaManager(this.#storage), this.#cache = new IntermediateCache(this.#storage), this.#segments = /* @__PURE__ */ new Map();
   }
+  /**
+   * 处理哈希算法配置
+   */
+  #processHashAlgorithmConfig() {
+    const hashConfig = this.#config.hashAlgorithm;
+    hashConfig === 64 ? (this.#isHash64Bit = !0, this.#hashAlgorithm64 = new Murmur3_64()) : hashConfig === 32 ? (this.#isHash64Bit = !1, this.#hashAlgorithm32 = new Murmur3_32()) : hashConfig && typeof hashConfig.hash == "function" ? typeof hashConfig.hash("test") == "bigint" ? (this.#isHash64Bit = !0, this.#hashAlgorithm64 = hashConfig) : (this.#isHash64Bit = !1, this.#hashAlgorithm32 = hashConfig) : (this.#isHash64Bit = !0, this.#hashAlgorithm64 = new Murmur3_64());
+  }
   /**
    * 开启批处理
    * 批处理期间 addDocuments 只写入缓存，不触发索引段构建
@@ -397,7 +625,7 @@ class SearchEngine {
       const segmentsMeta = this.#meta.getSegments(type);
       for (const meta of segmentsMeta) {
         const filename = meta.filename;
-        !this.#segments.has(filename) && !segmentsToLoad.has(filename) && segmentsToLoad.set(filename, new IndexSegment(filename, this.#storage));
+        !this.#segments.has(filename) && !segmentsToLoad.has(filename) && (this.#isHash64Bit ? segmentsToLoad.set(filename, new IndexSegment64(filename, this.#storage, this.#hashAlgorithm64)) : segmentsToLoad.set(filename, new IndexSegment(filename, this.#storage, this.#hashAlgorithm32)));
       }
     };
     collectSegments("word"), collectSegments("char"), await Promise.all(
@@ -466,7 +694,7 @@ class SearchEngine {
       ...this.#meta.getSegments("char")
     ];
     for (const seg of allSegments)
-      this.#segments.has(seg.filename) || this.#segments.set(seg.filename, new IndexSegment(seg.filename, this.#storage)), await this.#segments.get(seg.filename).loadIndex();
+      this.#segments.has(seg.filename) || (this.#isHash64Bit ? this.#segments.set(seg.filename, new IndexSegment64(seg.filename, this.#storage, this.#hashAlgorithm64)) : this.#segments.set(seg.filename, new IndexSegment(seg.filename, this.#storage, this.#hashAlgorithm32))), await this.#segments.get(seg.filename).loadIndex();
     this.#initialized = !0;
   }
   #getIndexingTokens(doc) {
@@ -498,7 +726,7 @@ class SearchEngine {
     }
     const docsToBuild = await this.#cache.readRange(cacheFilename, startOffset, currentCacheSize);
     let segment = this.#segments.get(targetSegmentName);
-    segment || (segment = new IndexSegment(targetSegmentName, this.#storage), this.#segments.set(targetSegmentName, segment)), await segment.buildAndSave(docsToBuild), this.#meta.updateSegment(type, targetSegmentName, startOffset, currentCacheSize, newTokenCountTotal, isNew);
+    segment || (this.#isHash64Bit ? segment = new IndexSegment64(targetSegmentName, this.#storage, this.#hashAlgorithm64) : segment = new IndexSegment(targetSegmentName, this.#storage, this.#hashAlgorithm32), this.#segments.set(targetSegmentName, segment)), await segment.buildAndSave(docsToBuild), this.#meta.updateSegment(type, targetSegmentName, startOffset, currentCacheSize, newTokenCountTotal, isNew);
   }
 }
-exports.SearchEngine = SearchEngine, exports.hash = murmur3_32, exports.murmur3_32 = murmur3_32;
+exports.Murmur3HashFactory = Murmur3HashFactory, exports.Murmur3_128 = Murmur3_128, exports.Murmur3_32 = Murmur3_32, exports.Murmur3_64 = Murmur3_64, exports.SearchEngine = SearchEngine, exports.defaultHashFactory = defaultHashFactory, exports.murmur3_32 = murmur3_32, exports.murmur3_64 = murmur3_64;

package/lib/core.d.ts

@@ -1,4 +1,4 @@
-import { ISearchEngine, ISearchEngineOption, IDocument, IDocumentBase, IResult, ISearchEngineStatus } from './type';
+import { ISearchEngine, ISearchEngineOption, IDocument, IDocumentBase, IResult, ISearchEngineStatus, IHashAlgorithm32, IHashAlgorithm64, IHashAlgorithm128, IHashAlgorithmFactory } from './type';
 
 /**
  * 核心搜索引擎类 (多实例支持)
@@ -52,4 +52,90 @@ declare class SearchEngine implements ISearchEngine {
  */
 declare function murmur3_32(str: string, h?: number): number;
 
-export { SearchEngine, murmur3_32 as hash, murmur3_32 };
+/**
+ * MurmurHash3 64位实现
+ * 高效的非加密哈希函数，适用于哈希表等数据结构
+ */
+/**
+ * 计算字符串的64位MurmurHash3哈希值
+ * @param str 要哈希的字符串
+ * @param seed 种子值，默认为0x12345678
+ * @returns 64位无符号哈希值（BigInt类型）
+ */
+declare function murmur3_64(str: string, seed?: number): bigint;
+
+/**
+ * Murmur3 32位哈希算法实现类
+ */
+declare class Murmur3_32 implements IHashAlgorithm32 {
+    private seed;
+    /**
+     * 构造函数
+     * @param seed 种子值，默认为0x12345678
+     */
+    constructor(seed?: number);
+    /**
+     * 计算字符串的32位哈希值
+     * @param str 输入字符串
+     * @returns 32位哈希值（number类型）
+     */
+    hash(str: string): number;
+}
+/**
+ * Murmur3 64位哈希算法实现类
+ */
+declare class Murmur3_64 implements IHashAlgorithm64 {
+    private seed;
+    /**
+     * 构造函数
+     * @param seed 种子值，默认为0x12345678
+     */
+    constructor(seed?: number);
+    /**
+     * 计算字符串的64位哈希值
+     * @param str 输入字符串
+     * @returns 64位哈希值（bigint类型）
+     */
+    hash(str: string): bigint;
+}
+/**
+ * Murmur3 128位哈希算法实现类
+ * 注意：当前项目中没有实际的128位实现，这里只是一个占位符
+ */
+declare class Murmur3_128 implements IHashAlgorithm128 {
+    private seed;
+    /**
+     * 构造函数
+     * @param seed 种子值，默认为0x12345678
+     */
+    constructor(seed?: number);
+    /**
+     * 计算字符串的128位哈希值
+     * @param str 输入字符串
+     * @returns 128位哈希值（bigint类型）
+     */
+    hash(str: string): bigint;
+}
+/**
+ * Murmur3哈希算法工厂类
+ */
+declare class Murmur3HashFactory implements IHashAlgorithmFactory {
+    /**
+     * 创建32位哈希算法实例
+     * @returns 32位哈希算法实例
+     */
+    create32(): IHashAlgorithm32;
+    /**
+     * 创建64位哈希算法实例
+     * @returns 64位哈希算法实例
+     */
+    create64(): IHashAlgorithm64;
+    /**
+     * 创建128位哈希算法实例
+     * @returns 128位哈希算法实例
+     */
+    create128(): IHashAlgorithm128;
+}
+declare const defaultHashFactory: Murmur3HashFactory;
+
+export { Murmur3HashFactory, Murmur3_128, Murmur3_32, Murmur3_64, SearchEngine, defaultHashFactory, murmur3_32, murmur3_64 };

package/lib/core.js

@@ -180,21 +180,129 @@ function murmur3_32(str, h = 305419896) {
   const remainder = len & 3;
   return remainder > 0 && (remainder >= 3 && (k1 ^= (str.charCodeAt(i + 2) & 255) << 16), remainder >= 2 && (k1 ^= (str.charCodeAt(i + 1) & 255) << 8), remainder >= 1 && (k1 ^= str.charCodeAt(i) & 255), k1 = (k1 & 65535) * 3432918353 + (((k1 >>> 16) * 3432918353 & 65535) << 16) & 4294967295, k1 = k1 << 15 | k1 >>> 17, k1 = (k1 & 65535) * 461845907 + (((k1 >>> 16) * 461845907 & 65535) << 16) & 4294967295, h ^= k1), h ^= len, h ^= h >>> 16, h = (h & 65535) * 2246822507 + (((h >>> 16) * 2246822507 & 65535) << 16) & 4294967295, h ^= h >>> 13, h = (h & 65535) * 3266489909 + (((h >>> 16) * 3266489909 & 65535) << 16) & 4294967295, h ^= h >>> 16, h >>> 0;
 }
+function murmur3_64(str, seed = 305419896) {
+  const len = str.length, nBlocks = len >> 3;
+  let h1 = BigInt(seed), h2 = BigInt(seed);
+  const c1 = 0x87c37b91114253d5n, c2 = 0x4cf5ad432745937fn;
+  let i = 0;
+  for (; i < nBlocks; ) {
+    let k12 = BigInt(str.charCodeAt(i) & 255) | BigInt(str.charCodeAt(++i) & 255) << 8n | BigInt(str.charCodeAt(++i) & 255) << 16n | BigInt(str.charCodeAt(++i) & 255) << 24n | BigInt(str.charCodeAt(++i) & 255) << 32n | BigInt(str.charCodeAt(++i) & 255) << 40n | BigInt(str.charCodeAt(++i) & 255) << 48n | BigInt(str.charCodeAt(++i) & 255) << 56n;
+    ++i;
+    let k22 = BigInt(str.charCodeAt(i) & 255) | BigInt(str.charCodeAt(++i) & 255) << 8n | BigInt(str.charCodeAt(++i) & 255) << 16n | BigInt(str.charCodeAt(++i) & 255) << 24n | BigInt(str.charCodeAt(++i) & 255) << 32n | BigInt(str.charCodeAt(++i) & 255) << 40n | BigInt(str.charCodeAt(++i) & 255) << 48n | BigInt(str.charCodeAt(++i) & 255) << 56n;
+    ++i, k12 = k12 * c1 % 2n ** 64n, k12 = (k12 << 31n | k12 >> 33n) % 2n ** 64n, k12 = k12 * c2 % 2n ** 64n, h1 ^= k12, h1 = (h1 << 27n | h1 >> 37n) % 2n ** 64n, h1 = (h1 + h2) % 2n ** 64n, h1 = (h1 * 5n + 0x52dce729n) % 2n ** 64n, k22 = k22 * c2 % 2n ** 64n, k22 = (k22 << 33n | k22 >> 31n) % 2n ** 64n, k22 = k22 * c1 % 2n ** 64n, h2 ^= k22, h2 = (h2 << 31n | h2 >> 33n) % 2n ** 64n, h2 = (h2 + h1) % 2n ** 64n, h2 = (h2 * 5n + 0x38495ab5n) % 2n ** 64n;
+  }
+  let k1 = 0n, k2 = 0n;
+  const remainder = len & 7;
+  return remainder > 0 && (remainder >= 8 && (k2 ^= BigInt(str.charCodeAt(i + 7) & 255) << 56n), remainder >= 7 && (k2 ^= BigInt(str.charCodeAt(i + 6) & 255) << 48n), remainder >= 6 && (k2 ^= BigInt(str.charCodeAt(i + 5) & 255) << 40n), remainder >= 5 && (k2 ^= BigInt(str.charCodeAt(i + 4) & 255) << 32n), remainder >= 4 && (k1 ^= BigInt(str.charCodeAt(i + 3) & 255) << 24n), remainder >= 3 && (k1 ^= BigInt(str.charCodeAt(i + 2) & 255) << 16n), remainder >= 2 && (k1 ^= BigInt(str.charCodeAt(i + 1) & 255) << 8n), remainder >= 1 && (k1 ^= BigInt(str.charCodeAt(i) & 255)), k1 = k1 * c1 % 2n ** 64n, k1 = (k1 << 31n | k1 >> 33n) % 2n ** 64n, k1 = k1 * c2 % 2n ** 64n, h1 ^= k1, k2 = k2 * c2 % 2n ** 64n, k2 = (k2 << 33n | k2 >> 31n) % 2n ** 64n, k2 = k2 * c1 % 2n ** 64n, h2 ^= k2), h1 ^= BigInt(len), h2 ^= BigInt(len), h1 = (h1 + h2) % 2n ** 64n, h2 = (h2 + h1) % 2n ** 64n, h1 = (h1 ^ h1 >> 33n) * 0xff51afd7ed558ccdn, h1 = (h1 ^ h1 >> 33n) * 0xc4ceb9fe1a85ec53n, h1 = h1 ^ h1 >> 33n, h2 = (h2 ^ h2 >> 33n) * 0xff51afd7ed558ccdn, h2 = (h2 ^ h2 >> 33n) * 0xc4ceb9fe1a85ec53n, h2 = h2 ^ h2 >> 33n, (h2 << 64n | h1) & 0xffffffffffffffffn;
+}
+class Murmur3_32 {
+  seed;
+  /**
+   * 构造函数
+   * @param seed 种子值，默认为0x12345678
+   */
+  constructor(seed = 305419896) {
+    this.seed = seed;
+  }
+  /**
+   * 计算字符串的32位哈希值
+   * @param str 输入字符串
+   * @returns 32位哈希值（number类型）
+   */
+  hash(str) {
+    return murmur3_32(str, this.seed);
+  }
+}
+class Murmur3_64 {
+  seed;
+  /**
+   * 构造函数
+   * @param seed 种子值，默认为0x12345678
+   */
+  constructor(seed = 305419896) {
+    this.seed = seed;
+  }
+  /**
+   * 计算字符串的64位哈希值
+   * @param str 输入字符串
+   * @returns 64位哈希值（bigint类型）
+   */
+  hash(str) {
+    return murmur3_64(str, this.seed);
+  }
+}
+class Murmur3_128 {
+  seed;
+  /**
+   * 构造函数
+   * @param seed 种子值，默认为0x12345678
+   */
+  constructor(seed = 305419896) {
+    this.seed = seed;
+  }
+  /**
+   * 计算字符串的128位哈希值
+   * @param str 输入字符串
+   * @returns 128位哈希值（bigint类型）
+   */
+  hash(str) {
+    const hash1 = murmur3_64(str, this.seed);
+    return murmur3_64(str + str, this.seed ^ 1515870810) << 64n | hash1;
+  }
+}
+class Murmur3HashFactory {
+  /**
+   * 创建32位哈希算法实例
+   * @returns 32位哈希算法实例
+   */
+  create32() {
+    return new Murmur3_32();
+  }
+  /**
+   * 创建64位哈希算法实例
+   * @returns 64位哈希算法实例
+   */
+  create64() {
+    return new Murmur3_64();
+  }
+  /**
+   * 创建128位哈希算法实例
+   * @returns 128位哈希算法实例
+   */
+  create128() {
+    return new Murmur3_128();
+  }
+}
+const defaultHashFactory = new Murmur3HashFactory();
 class IndexSegment {
   #filename;
   #storage;
   #buffer = null;
   #view = null;
+  #hashAlgorithm;
   /**
-   * 使用MurmurHash3计算字符串哈希值
+   * 构造函数
+   * @param filename 文件名
+   * @param storage 存储接口
+   * @param hashAlgorithm 哈希算法实例，默认为Murmur3_32
+   */
+  constructor(filename, storage, hashAlgorithm = new Murmur3_32()) {
+    this.#filename = filename, this.#storage = storage, this.#hashAlgorithm = hashAlgorithm;
+  }
+  /**
+   * 使用当前哈希算法计算字符串哈希值
    * @param str 要哈希的字符串
    * @returns 32位无符号哈希值
    */
-  static hash(str) {
-    return murmur3_32(str);
+  hash(str) {
+    return this.#hashAlgorithm.hash(str);
   }
-  constructor(filename, storage) {
-    this.#filename = filename, this.#storage = storage;
+  /**
+   * 设置哈希算法
+   * @param hashAlgorithm 新的哈希算法实例
+   */
+  setHashAlgorithm(hashAlgorithm) {
+    this.#hashAlgorithm = hashAlgorithm;
   }
   async loadIndex() {
     return this.#buffer ? !0 : (this.#buffer = await this.#storage.read(this.#filename), this.#buffer ? (this.#view = new DataView(this.#buffer), !0) : !1);
@@ -205,7 +313,7 @@ class IndexSegment {
       const uniqueTokens = /* @__PURE__ */ new Map();
       for (const token of doc.tokens)
         uniqueTokens.has(token) || (uniqueTokens.set(token, !0), tokenMap.has(token) || tokenMap.set(token, {
-          hash: IndexSegment.hash(token),
+          hash: this.hash(token),
           postings: []
         }), tokenMap.get(token).postings.push(doc.id));
     }
@@ -221,8 +329,8 @@ class IndexSegment {
     const headerSize = 12, dictSize = entries.length * 20, postingsSize = totalPostings * 4, tokensOffset = headerSize + dictSize + postingsSize, totalSize = tokensOffset + totalTokensSize, buffer = new ArrayBuffer(totalSize), view = new DataView(buffer);
     view.setUint32(0, 1229866072), view.setUint32(4, entries.length, !0), view.setUint32(8, tokensOffset, !0);
     let currentDictOffset = headerSize, currentPostingsOffset = headerSize + dictSize, currentTokenOffset = tokensOffset;
-    for (const [token, { hash: hash2, postings }] of entries) {
-      view.setUint32(currentDictOffset, hash2, !0);
+    for (const [token, { hash, postings }] of entries) {
+      view.setUint32(currentDictOffset, hash, !0);
       const tokenBytes = encoder.encode(token);
       view.setUint32(currentDictOffset + 4, tokenBytes.length, !0), view.setUint32(currentDictOffset + 8, currentTokenOffset, !0), view.setUint32(currentDictOffset + 12, currentPostingsOffset, !0), view.setUint32(currentDictOffset + 16, postings.length, !0), currentDictOffset += 20;
       for (let i = 0; i < postings.length; i++)
@@ -235,7 +343,7 @@ class IndexSegment {
   }
   search(term) {
     if (!this.#view || !this.#buffer) return [];
-    const h = IndexSegment.hash(term), count = this.#view.getUint32(4, !0);
+    const h = this.hash(term), count = this.#view.getUint32(4, !0);
     let left = 0, right = count - 1;
     const headerSize = 12, entrySize = 20, decoder = new TextDecoder();
     for (; left <= right; ) {
@@ -276,6 +384,116 @@ class IndexSegment {
     return [];
   }
 }
+class IndexSegment64 {
+  #filename;
+  #storage;
+  #buffer = null;
+  #view = null;
+  #hashAlgorithm;
+  /**
+   * 构造函数
+   * @param filename 文件名
+   * @param storage 存储接口
+   * @param hashAlgorithm 哈希算法实例，默认为Murmur3_64
+   */
+  constructor(filename, storage, hashAlgorithm = new Murmur3_64()) {
+    this.#filename = filename, this.#storage = storage, this.#hashAlgorithm = hashAlgorithm;
+  }
+  /**
+   * 使用当前哈希算法计算字符串哈希值
+   * @param str 要哈希的字符串
+   * @returns 64位无符号哈希值
+   */
+  hash(str) {
+    return this.#hashAlgorithm.hash(str);
+  }
+  /**
+   * 设置哈希算法
+   * @param hashAlgorithm 新的哈希算法实例
+   */
+  setHashAlgorithm(hashAlgorithm) {
+    this.#hashAlgorithm = hashAlgorithm;
+  }
+  async loadIndex() {
+    return this.#buffer ? !0 : (this.#buffer = await this.#storage.read(this.#filename), this.#buffer ? (this.#view = new DataView(this.#buffer), !0) : !1);
+  }
+  async buildAndSave(docs) {
+    const tokenMap = /* @__PURE__ */ new Map();
+    for (const doc of docs) {
+      const uniqueTokens = /* @__PURE__ */ new Map();
+      for (const token of doc.tokens)
+        uniqueTokens.has(token) || (uniqueTokens.set(token, !0), tokenMap.has(token) || tokenMap.set(token, {
+          hash: this.hash(token),
+          postings: []
+        }), tokenMap.get(token).postings.push(doc.id));
+    }
+    const entries = Array.from(tokenMap.entries());
+    entries.sort(([a, ah], [b, bh]) => ah.hash !== bh.hash ? ah.hash > bh.hash ? 1 : -1 : a.localeCompare(b));
+    const encoder = new TextEncoder();
+    let totalPostings = 0, totalTokensSize = 0;
+    for (const [token, { postings }] of entries) {
+      totalPostings += postings.length;
+      const bytes = encoder.encode(token);
+      totalTokensSize += bytes.length + 1;
+    }
+    const headerSize = 16, dictSize = entries.length * 28, postingsSize = totalPostings * 4, tokensOffset = headerSize + dictSize + postingsSize, totalSize = tokensOffset + totalTokensSize, buffer = new ArrayBuffer(totalSize), view = new DataView(buffer);
+    view.setUint32(0, 1229866072), view.setUint32(4, entries.length, !0), view.setUint32(8, tokensOffset, !0), view.setUint32(12, 64, !0);
+    let currentDictOffset = headerSize, currentPostingsOffset = headerSize + dictSize, currentTokenOffset = tokensOffset;
+    for (const [token, { hash, postings }] of entries) {
+      view.setBigUint64(currentDictOffset, hash, !0);
+      const tokenBytes = encoder.encode(token);
+      view.setUint32(currentDictOffset + 8, tokenBytes.length, !0), view.setUint32(currentDictOffset + 12, currentTokenOffset, !0), view.setUint32(currentDictOffset + 16, currentPostingsOffset, !0), view.setUint32(currentDictOffset + 20, postings.length, !0), currentDictOffset += 28;
+      for (let i = 0; i < postings.length; i++)
+        view.setUint32(currentPostingsOffset, postings[i], !0), currentPostingsOffset += 4;
+      for (let i = 0; i < tokenBytes.length; i++)
+        view.setUint8(currentTokenOffset++, tokenBytes[i]);
+      view.setUint8(currentTokenOffset++, 0);
+    }
+    await this.#storage.write(this.#filename, buffer), this.#buffer = buffer, this.#view = view;
+  }
+  search(term) {
+    if (!this.#view || !this.#buffer) return [];
+    const h = this.hash(term), count = this.#view.getUint32(4, !0);
+    let left = 0, right = count - 1;
+    const headerSize = 16, entrySize = 28, decoder = new TextDecoder();
+    for (; left <= right; ) {
+      const mid = left + right >>> 1, entryPos = headerSize + mid * entrySize, entryHash = this.#view.getBigUint64(entryPos, !0);
+      if (entryHash < h)
+        left = mid + 1;
+      else if (entryHash > h)
+        right = mid - 1;
+      else {
+        if (!(mid > 0 && this.#view.getBigUint64(headerSize + (mid - 1) * entrySize, !0) === h || mid < count - 1 && this.#view.getBigUint64(headerSize + (mid + 1) * entrySize, !0) === h)) {
+          const postingsOffset = this.#view.getUint32(headerSize + mid * entrySize + 16, !0), postingsLen = this.#view.getUint32(headerSize + mid * entrySize + 20, !0), result = [];
+          for (let j = 0; j < postingsLen; j++)
+            result.push(this.#view.getUint32(postingsOffset + j * 4, !0));
+          return result;
+        }
+        let firstMatch = mid;
+        for (; firstMatch > 0; ) {
+          const prevPos = headerSize + (firstMatch - 1) * entrySize;
+          if (this.#view.getBigUint64(prevPos, !0) === h)
+            firstMatch--;
+          else
+            break;
+        }
+        for (let i = firstMatch; i < count; i++) {
+          const checkPos = headerSize + i * entrySize;
+          if (this.#view.getBigUint64(checkPos, !0) !== h) break;
+          const tokenLen = this.#view.getUint32(checkPos + 8, !0), tokenOffset = this.#view.getUint32(checkPos + 12, !0), tokenBuffer = new Uint8Array(this.#buffer, tokenOffset, tokenLen);
+          if (decoder.decode(tokenBuffer) === term) {
+            const postingsOffset = this.#view.getUint32(checkPos + 16, !0), postingsLen = this.#view.getUint32(checkPos + 20, !0), result = [];
+            for (let j = 0; j < postingsLen; j++)
+              result.push(this.#view.getUint32(postingsOffset + j * 4, !0));
+            return result;
+          }
+        }
+        return [];
+      }
+    }
+    return [];
+  }
+}
 const defaultTokenize = ({ text }) => {
   try {
     if (typeof Intl < "u" && typeof Intl.Segmenter == "function" && typeof Array.from == "function") {
@@ -294,6 +512,9 @@ class SearchEngine {
   #segments;
   #initialized = !1;
   #config;
+  #isHash64Bit = !0;
+  #hashAlgorithm32;
+  #hashAlgorithm64;
   // 批处理状态
   #inBatch = !1;
   #pendingTokenCounts = { word: 0, char: 0 };
@@ -305,12 +526,19 @@ class SearchEngine {
       minCharTokenSave: 0,
       indexingTokenizer: config.indexingTokenizer || defaultTokenize,
       ...config
-    }, (this.#config.minWordTokenSave || 0) >= (this.#config.wordSegmentTokenThreshold || 1e5))
+    }, this.#processHashAlgorithmConfig(), (this.#config.minWordTokenSave || 0) >= (this.#config.wordSegmentTokenThreshold || 1e5))
       throw new Error("minWordTokenSave must be less than wordSegmentTokenThreshold");
     if ((this.#config.minCharTokenSave || 0) >= (this.#config.charSegmentTokenThreshold || 5e5))
       throw new Error("minCharTokenSave must be less than charSegmentTokenThreshold");
     this.#storage = config.storage, this.#meta = new MetaManager(this.#storage), this.#cache = new IntermediateCache(this.#storage), this.#segments = /* @__PURE__ */ new Map();
   }
+  /**
+   * 处理哈希算法配置
+   */
+  #processHashAlgorithmConfig() {
+    const hashConfig = this.#config.hashAlgorithm;
+    hashConfig === 64 ? (this.#isHash64Bit = !0, this.#hashAlgorithm64 = new Murmur3_64()) : hashConfig === 32 ? (this.#isHash64Bit = !1, this.#hashAlgorithm32 = new Murmur3_32()) : hashConfig && typeof hashConfig.hash == "function" ? typeof hashConfig.hash("test") == "bigint" ? (this.#isHash64Bit = !0, this.#hashAlgorithm64 = hashConfig) : (this.#isHash64Bit = !1, this.#hashAlgorithm32 = hashConfig) : (this.#isHash64Bit = !0, this.#hashAlgorithm64 = new Murmur3_64());
+  }
   /**
    * 开启批处理
    * 批处理期间 addDocuments 只写入缓存，不触发索引段构建
@@ -396,7 +624,7 @@ class SearchEngine {
       const segmentsMeta = this.#meta.getSegments(type);
       for (const meta of segmentsMeta) {
         const filename = meta.filename;
-        !this.#segments.has(filename) && !segmentsToLoad.has(filename) && segmentsToLoad.set(filename, new IndexSegment(filename, this.#storage));
+        !this.#segments.has(filename) && !segmentsToLoad.has(filename) && (this.#isHash64Bit ? segmentsToLoad.set(filename, new IndexSegment64(filename, this.#storage, this.#hashAlgorithm64)) : segmentsToLoad.set(filename, new IndexSegment(filename, this.#storage, this.#hashAlgorithm32)));
       }
     };
     collectSegments("word"), collectSegments("char"), await Promise.all(
@@ -465,7 +693,7 @@ class SearchEngine {
       ...this.#meta.getSegments("char")
     ];
     for (const seg of allSegments)
-      this.#segments.has(seg.filename) || this.#segments.set(seg.filename, new IndexSegment(seg.filename, this.#storage)), await this.#segments.get(seg.filename).loadIndex();
+      this.#segments.has(seg.filename) || (this.#isHash64Bit ? this.#segments.set(seg.filename, new IndexSegment64(seg.filename, this.#storage, this.#hashAlgorithm64)) : this.#segments.set(seg.filename, new IndexSegment(seg.filename, this.#storage, this.#hashAlgorithm32))), await this.#segments.get(seg.filename).loadIndex();
     this.#initialized = !0;
   }
   #getIndexingTokens(doc) {
@@ -497,11 +725,16 @@ class SearchEngine {
     }
     const docsToBuild = await this.#cache.readRange(cacheFilename, startOffset, currentCacheSize);
     let segment = this.#segments.get(targetSegmentName);
-    segment || (segment = new IndexSegment(targetSegmentName, this.#storage), this.#segments.set(targetSegmentName, segment)), await segment.buildAndSave(docsToBuild), this.#meta.updateSegment(type, targetSegmentName, startOffset, currentCacheSize, newTokenCountTotal, isNew);
+    segment || (this.#isHash64Bit ? segment = new IndexSegment64(targetSegmentName, this.#storage, this.#hashAlgorithm64) : segment = new IndexSegment(targetSegmentName, this.#storage, this.#hashAlgorithm32), this.#segments.set(targetSegmentName, segment)), await segment.buildAndSave(docsToBuild), this.#meta.updateSegment(type, targetSegmentName, startOffset, currentCacheSize, newTokenCountTotal, isNew);
   }
 }
 export {
+  Murmur3HashFactory,
+  Murmur3_128,
+  Murmur3_32,
+  Murmur3_64,
   SearchEngine,
-  murmur3_32 as hash,
-  murmur3_32
+  defaultHashFactory,
+  murmur3_32,
+  murmur3_64
 };

package/lib/type.d.ts

@@ -61,6 +61,35 @@ interface IStorage {
     getFileSize(filename: string): Promise<number>;
 }
 
+interface IHashAlgorithm<T> {
+    /**
+     * 计算字符串的哈希值
+     * @param str 输入字符串
+     * @returns 哈希值
+     */
+    hash(str: string): T;
+}
+interface IHashAlgorithm32 extends IHashAlgorithm<number> {
+}
+interface IHashAlgorithm64 extends IHashAlgorithm<bigint> {
+}
+interface IHashAlgorithm128 extends IHashAlgorithm<bigint> {
+}
+interface IHashAlgorithmFactory {
+    /**
+     * 创建32位哈希算法实例
+     */
+    create32(): IHashAlgorithm32;
+    /**
+     * 创建64位哈希算法实例
+     */
+    create64(): IHashAlgorithm64;
+    /**
+     * 创建128位哈希算法实例
+     */
+    create128(): IHashAlgorithm128;
+}
+
 /**
  * 索引类型
  */
@@ -93,6 +122,15 @@ interface ISearchEngineOption {
      * - 影响: 直接决定搜索匹配的范围和结果的相关性
      */
     searchTokenizer?: SearchTokenizer;
+    /**
+     * 哈希算法配置 (可选)
+     * - 32: 使用默认32位哈希算法
+     * - 64: 使用默认64位哈希算法
+     * - IHashAlgorithm32: 使用自定义32位哈希算法
+     * - IHashAlgorithm64: 使用自定义64位哈希算法
+     * - undefined: 默认使用32位哈希算法
+     */
+    hashAlgorithm?: 32 | 64 | IHashAlgorithm32 | IHashAlgorithm64;
     /**
      * 词索引分段阈值 (Token数) - 分段算法配置
      * - 作用: 控制词索引文件的大小，超过阈值时创建新的索引段
@@ -173,4 +211,38 @@ interface ISearchEngine {
     hasDocument(id: number): Promise<boolean>;
 }
 
-export type { IDocument, IDocumentBase, IIndexMeta, IResult, ISearchEngine, ISearchEngineOption, ISearchEngineStatus, ISegmentMeta, IStorage, ITokenizedDoc, IndexType, IndexingTokenizer, SearchTokenizer };
+/**
+ * 索引段接口，定义了IndexSegment和IndexSegment64的共同方法
+ */
+
+interface IIndexSegment {
+    /**
+     * 使用当前哈希算法计算字符串哈希值
+     * @param str 要哈希的字符串
+     * @returns 哈希值（number | bigint）
+     */
+    hash(str: string): number | bigint;
+    /**
+     * 设置哈希算法
+     * @param hashAlgorithm 新的哈希算法实例
+     */
+    setHashAlgorithm(hashAlgorithm: IHashAlgorithm32 | IHashAlgorithm64): void;
+    /**
+     * 加载索引
+     * @returns 是否成功加载索引
+     */
+    loadIndex(): Promise<boolean>;
+    /**
+     * 构建并保存索引
+     * @param docs 要索引的文档
+     */
+    buildAndSave(docs: ITokenizedDoc[]): Promise<void>;
+    /**
+     * 搜索索引
+     * @param term 搜索词
+     * @returns 匹配的文档ID数组
+     */
+    search(term: string): number[];
+}
+
+export type { IDocument, IDocumentBase, IHashAlgorithm, IHashAlgorithm128, IHashAlgorithm32, IHashAlgorithm64, IHashAlgorithmFactory, IIndexMeta, IIndexSegment, IResult, ISearchEngine, ISearchEngineOption, ISearchEngineStatus, ISegmentMeta, IStorage, ITokenizedDoc, IndexType, IndexingTokenizer, SearchTokenizer };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "gs-search",
-  "version": "0.1.7",
+  "version": "0.1.8",
   "type": "module",
   "main": "lib/index.cjs",
   "module": "lib/index.js",