gs-search 0.1.2 → 0.1.4

package/lib/browser.cjs

@@ -0,0 +1 @@
+"use strict";exports.BrowserStorage=class{#a;constructor(a){this.#a=a}async#t(){return await(await navigator.storage.getDirectory()).getDirectoryHandle(this.#a,{create:!0})}async write(a,t){const e=await(await(await this.#t()).getFileHandle(a,{create:!0})).createWritable();await e.write(t),await e.close()}async append(a,t){const e=await this.#t();let i;try{i=await e.getFileHandle(a,{create:!0})}catch{i=await e.getFileHandle(a,{create:!0})}const r=await i.getFile(),s=await i.createWritable({keepExistingData:!0});await s.seek(r.size),await s.write(t),await s.close()}async read(a){const t=await this.#t();try{return await(await(await t.getFileHandle(a)).getFile()).arrayBuffer()}catch{return null}}async readRange(a,t,e){const i=await this.#t();try{return await(await(await i.getFileHandle(a)).getFile()).slice(t,e).arrayBuffer()}catch{return null}}async remove(a){const t=await this.#t();try{await t.removeEntry(a)}catch{}}async listFiles(){const a=await this.#t(),t=[];for await(const e of a.keys())t.push(e);return t}async clearAll(){const a=await this.#t();for await(const t of a.keys())await a.removeEntry(t,{recursive:!0})}async getFileSize(a){const t=await this.#t();try{return(await(await t.getFileHandle(a)).getFile()).size}catch{return 0}}};

package/lib/browser.d.ts

@@ -0,0 +1,21 @@
+import { IStorage } from './type';
+
+/**
+ * 浏览器实现 (OPFS - 子目录隔离)
+ * 支持 Main Thread 和 Web Worker
+ */
+
+declare class BrowserStorage implements IStorage {
+    #private;
+    constructor(baseDir: string);
+    write(filename: string, data: ArrayBuffer): Promise<void>;
+    append(filename: string, data: ArrayBuffer): Promise<void>;
+    read(filename: string): Promise<ArrayBuffer | null>;
+    readRange(filename: string, start: number, end: number): Promise<ArrayBuffer | null>;
+    remove(filename: string): Promise<void>;
+    listFiles(): Promise<string[]>;
+    clearAll(): Promise<void>;
+    getFileSize(filename: string): Promise<number>;
+}
+
+export { BrowserStorage };

package/lib/browser.js

@@ -0,0 +1 @@
+class a{#a;constructor(a){this.#a=a}async#t(){return await(await navigator.storage.getDirectory()).getDirectoryHandle(this.#a,{create:!0})}async write(a,t){const e=await(await(await this.#t()).getFileHandle(a,{create:!0})).createWritable();await e.write(t),await e.close()}async append(a,t){const e=await this.#t();let i;try{i=await e.getFileHandle(a,{create:!0})}catch{i=await e.getFileHandle(a,{create:!0})}const r=await i.getFile(),n=await i.createWritable({keepExistingData:!0});await n.seek(r.size),await n.write(t),await n.close()}async read(a){const t=await this.#t();try{return await(await(await t.getFileHandle(a)).getFile()).arrayBuffer()}catch{return null}}async readRange(a,t,e){const i=await this.#t();try{return await(await(await i.getFileHandle(a)).getFile()).slice(t,e).arrayBuffer()}catch{return null}}async remove(a){const t=await this.#t();try{await t.removeEntry(a)}catch{}}async listFiles(){const a=await this.#t(),t=[];for await(const e of a.keys())t.push(e);return t}async clearAll(){const a=await this.#t();for await(const t of a.keys())await a.removeEntry(t,{recursive:!0})}async getFileSize(a){const t=await this.#t();try{return(await(await t.getFileHandle(a)).getFile()).size}catch{return 0}}}export{a as BrowserStorage};

package/lib/core.cjs

@@ -0,0 +1 @@
+"use strict";var e=require('./browser'),t=require('./node');const s="search_meta.json",n="deleted_ids.bin",i="added_ids.bin";class o{#e;#t={wordSegments:[],charSegments:[]};#s=new Set;#n=new Set;constructor(e){this.#e=e}async load(){const e=await this.#e.read(s);if(e){const t=(new TextDecoder).decode(e);this.#t=JSON.parse(t)}else this.#t={wordSegments:[],charSegments:[]};const t=await this.#e.read(n);if(t){const e=new DataView(t);let s=0;const n=t.byteLength;for(;s<n&&!(s+4>n);){const t=e.getUint32(s,!0);this.#s.add(t),s+=4,s<n&&30===e.getUint8(s)&&(s+=1)}}const o=await this.#e.read(i);if(o){const e=new DataView(o);let t=0;const s=o.byteLength;for(;t<s&&!(t+4>s);){const n=e.getUint32(t,!0);this.#n.add(n),t+=4,t<s&&30===e.getUint8(t)&&(t+=1)}}}async save(){const e=JSON.stringify(this.#t);if(await this.#e.write(s,(new TextEncoder).encode(e).buffer),0===this.#s.size)await this.#e.remove(n);else{const e=4*this.#s.size+this.#s.size,t=new ArrayBuffer(e),s=new DataView(t);let i=0;for(const e of this.#s)s.setUint32(i,e,!0),i+=4,s.setUint8(i,30),i+=1;await this.#e.write(n,t)}if(0===this.#n.size)await this.#e.remove(i);else{const e=4*this.#n.size+this.#n.size,t=new ArrayBuffer(e),s=new DataView(t);let n=0;for(const e of this.#n)s.setUint32(n,e,!0),n+=4,s.setUint8(n,30),n+=1;await this.#e.write(i,t)}}getSegments(e){return"word"===e?this.#t.wordSegments:this.#t.charSegments}getDeletedIds(){return this.#s}addDeletedId(e){this.#s.add(e)}isDeleted(e){return this.#s.has(e)}addAddedId(e){this.#n.add(e)}removeAddedId(e){this.#n.delete(e)}isAdded(e){return this.#n.has(e)}getAddedIds(){return this.#n}hasDocument(e){return this.#n.has(e)||this.#s.has(e)}getLastSegmentInfo(e){const t=this.getSegments(e);return 0===t.length?null:t[t.length-1]}updateSegment(e,t,s,n,i,o){const a="word"===e?this.#t.wordSegments:this.#t.charSegments;if(o)a.push({filename:t,start:s,end:n,tokenCount:i});else{const e=a[a.length-1];e&&e.filename===t&&(e.end=n,e.tokenCount=i)}}reset(){this.#t={wordSegments:[],charSegments:[]},this.#s.clear(),this.#n.clear()}}class a{static SEPARATOR=30;#e;constructor(e){this.#e=e}async appendBatch(e,t){if(0===t.length)return await this.#e.getFileSize(e);const s=new TextEncoder;let n=0;for(const e of t){n+=8;for(const t of e.tokens){n+=2+Math.min(s.encode(t).byteLength,65535)}n+=1}const i=new Uint8Array(n);let o=0;for(const e of t){const t=[];for(const n of e.tokens){const e=s.encode(n),i=e.byteLength>65535?e.slice(0,65535):e;t.push(i)}const n=new DataView(i.buffer,o);n.setUint32(0,e.id,!0),n.setUint32(4,t.length,!0),o+=8;for(const e of t)new DataView(i.buffer,o).setUint16(0,e.byteLength,!0),o+=2,i.set(e,o),o+=e.byteLength;i[o++]=a.SEPARATOR}return await this.#e.append(e,i.buffer),await this.#e.getFileSize(e)}async readRange(e,t,s){const n=await this.#e.readRange(e,t,s);if(!n||0===n.byteLength)return[];const i=new DataView(n),o=new Uint8Array(n),r=new TextDecoder,h=[];let d=0;const c=n.byteLength;for(;d<c&&!(d+8>c);){const e=i.getUint32(d,!0);d+=4;const t=i.getUint32(d,!0);d+=4;const s=[];for(let e=0;e<t&&!(d+2>c);e++){const e=i.getUint16(d,!0);if(d+=2,d+e>c)break;const t=new Uint8Array(n,d,e);s.push(r.decode(t)),d+=e}d<c&&o[d]===a.SEPARATOR&&(d+=1),h.push({id:e,tokens:s})}return h}async getCurrentSize(e){return await this.#e.getFileSize(e)}}function r(e,t=305419896){const s=e.length,n=s>>2;let i=0;for(;i<n;){let s=255&e.charCodeAt(i)|(255&e.charCodeAt(++i))<<8|(255&e.charCodeAt(++i))<<16|(255&e.charCodeAt(++i))<<24;++i,s=3432918353*(65535&s)+((3432918353*(s>>>16)&65535)<<16)&4294967295,s=s<<15|s>>>17,s=461845907*(65535&s)+((461845907*(s>>>16)&65535)<<16)&4294967295,t=27492+(65535&(t=5*(65535&(t=(t^=s)<<13|t>>>19))+((5*(t>>>16)&65535)<<16)&4294967295))+(((t>>>16)+58964&65535)<<16)}let o=0;const a=3&s;return a>0&&(a>=3&&(o^=(255&e.charCodeAt(i+2))<<16),a>=2&&(o^=(255&e.charCodeAt(i+1))<<8),a>=1&&(o^=255&e.charCodeAt(i)),o=3432918353*(65535&o)+((3432918353*(o>>>16)&65535)<<16)&4294967295,o=o<<15|o>>>17,o=461845907*(65535&o)+((461845907*(o>>>16)&65535)<<16)&4294967295,t^=o),t^=s,t=2246822507*(65535&(t^=t>>>16))+((2246822507*(t>>>16)&65535)<<16)&4294967295,t=3266489909*(65535&(t^=t>>>13))+((3266489909*(t>>>16)&65535)<<16)&4294967295,(t^=t>>>16)>>>0}class h{#i;#e;#o=null;#a=null;static hash(e){return r(e)}constructor(e,t){this.#i=e,this.#e=t}async loadIndex(){return!!this.#o||(this.#o=await this.#e.read(this.#i),!!this.#o&&(this.#a=new DataView(this.#o),!0))}async buildAndSave(e){const t=new Map;for(const s of e){const e=new Map;for(const n of s.tokens)e.has(n)||(e.set(n,!0),t.has(n)||t.set(n,{hash:h.hash(n),postings:[]}),t.get(n).postings.push(s.id))}const s=Array.from(t.entries());s.sort(([e,{hash:t}],[s,{hash:n}])=>t!==n?t-n:e.localeCompare(s));let n=0,i=0;for(const[e,{postings:t}]of s)n+=t.length,i+=e.length+1;const o=20*s.length,a=12+o+4*n,r=new ArrayBuffer(a+i),d=new DataView(r);d.setUint32(0,1229866072),d.setUint32(4,s.length),d.setUint32(8,a);let c=12,g=12+o,f=a;for(const[e,{hash:t,postings:n}]of s){d.setUint32(c,t),d.setUint32(c+4,e.length),d.setUint32(c+8,f),d.setUint32(c+12,g),d.setUint32(c+16,n.length),c+=20;for(let e=0;e<n.length;e++)d.setUint32(g,n[e],!0),g+=4;const s=(new TextEncoder).encode(e);for(let e=0;e<s.length;e++)d.setUint8(f++,s[e]);d.setUint8(f++,0)}await this.#e.write(this.#i,r),this.#o=r,this.#a=d}search(e){if(!this.#a||!this.#o)return[];const t=h.hash(e),s=this.#a.getUint32(4);let n=0,i=s-1;const o=12,a=20,r=new TextDecoder;for(;n<=i;){const h=n+i>>>1,d=o+h*a,c=this.#a.getUint32(d);if(c<t)n=h+1;else{if(!(c>t)){if(!(h>0&&this.#a.getUint32(o+(h-1)*a)===t||h<s-1&&this.#a.getUint32(o+(h+1)*a)===t)){const e=this.#a.getUint32(o+h*a+12),t=this.#a.getUint32(o+h*a+16),s=[];for(let n=0;n<t;n++)s.push(this.#a.getUint32(e+4*n,!0));return s}let n=h;for(;n>0;){const e=o+(n-1)*a;if(this.#a.getUint32(e)!==t)break;n--}for(let i=n;i<s;i++){const s=o+i*a;if(this.#a.getUint32(s)!==t)break;const n=this.#a.getUint32(s+4),h=this.#a.getUint32(s+8),d=new Uint8Array(this.#o,h,n);if(r.decode(d)===e){const e=this.#a.getUint32(s+12),t=this.#a.getUint32(s+16),n=[];for(let s=0;s<t;s++)n.push(this.#a.getUint32(e+4*s,!0));return n}}return[]}i=h-1}}return[]}}const d="word_cache.bin",c="char_cache.bin";exports.SearchEngine=class{#e;#t;#r;#h;#d=!1;#c;#g=!1;#f={word:0,char:0};constructor(s){if(!s.baseDir)throw new Error("SearchEngine requires 'baseDir' in config.");if(this.#c={wordSegmentTokenThreshold:1e5,charSegmentTokenThreshold:5e5,minWordTokenSave:0,minCharTokenSave:0,...s},(this.#c.minWordTokenSave||0)>=(this.#c.wordSegmentTokenThreshold||1e5))throw new Error("minWordTokenSave must be less than wordSegmentTokenThreshold");if((this.#c.minCharTokenSave||0)>=(this.#c.charSegmentTokenThreshold||5e5))throw new Error("minCharTokenSave must be less than charSegmentTokenThreshold");let n=null;if(this.#c.storage&&("object"==typeof this.#c.storage?n=this.#c.storage:"browser"===this.#c.storage?n=new e.BrowserStorage(this.#c.baseDir):"node"===this.#c.storage&&(n=new t.NodeStorage(this.#c.baseDir))),!n){const s=typeof navigator<"u"&&navigator?.storage?.getDirectory instanceof Function,i=typeof process<"u"&&null!=process.versions&&null!=process.versions.node;s?n=new e.BrowserStorage(this.#c.baseDir):i&&(n=new t.NodeStorage(this.#c.baseDir))}if(!n)throw new Error('Storage initialization failed. Please configure "storage" explicitly or ensure you are in a supported environment (Browser/Node).');this.#e=n,this.#t=new o(this.#e),this.#r=new a(this.#e),this.#h=new Map}async init(){if(this.#d)return;await this.#t.load();const e=[...this.#t.getSegments("word"),...this.#t.getSegments("char")];for(const t of e)this.#h.has(t.filename)||this.#h.set(t.filename,new h(t.filename,this.#e)),await this.#h.get(t.filename).loadIndex();this.#d=!0}startBatch(){this.#g=!0,this.#f={word:0,char:0}}async endBatch(){this.#g=!1,this.#f.word>0&&await this.#l("word",this.#f.word),this.#f.char>0&&await this.#l("char",this.#f.char),this.#f={word:0,char:0},await this.#t.save()}#w(e){try{if(typeof Intl<"u"&&"function"==typeof Intl.Segmenter&&"function"==typeof Array.from){const t=new Intl.Segmenter([],{granularity:"word"}).segment(e);if("object"==typeof t&&null!==t)return Array.from(t).filter(e=>e?.isWordLike).map(e=>e?.segment?.toLowerCase()||"")}}catch{}return e.toLowerCase().split(/[^a-z0-9\u4e00-\u9fa5]+/g).filter(e=>e.length>0)}#m(e){return this.#c.indexingTokenizer?this.#c.indexingTokenizer(e):this.#w(e.text)}#u(e){return this.#c.searchTokenizer?this.#c.searchTokenizer(e):this.#m(e)}async addDocument(e){return this.addDocuments([e])}async addDocumentIfMissing(e){return this.addDocumentsIfMissing([e])}async addDocumentsIfMissing(e){if(this.#d||await this.init(),0===e.length)return;const t=this.#t.getDeletedIds(),s=[],n=[],i=[];for(const o of e){if(t.has(o.id)||this.#t.isAdded(o.id))continue;const e=this.#m(o),a=[],r=[];for(const t of e)t.length>1?a.push(t):1===t.length&&r.push(t);a.length>0&&s.push({id:o.id,tokens:a}),r.length>0&&n.push({id:o.id,tokens:r}),i.push(o)}if(0===i.length)return;let o=0,a=0;if(s.length>0){await this.#r.appendBatch(d,s);for(const e of s)o+=e.tokens.length}if(n.length>0){await this.#r.appendBatch(c,n);for(const e of n)a+=e.tokens.length}for(const e of i)this.#t.addAddedId(e.id);this.#g?(this.#f.word+=o,this.#f.char+=a):(o>0&&await this.#l("word",o),a>0&&await this.#l("char",a),await this.#t.save())}async addDocuments(e){if(this.#d||await this.init(),0===e.length)return;const t=this.#t.getDeletedIds(),s=[],n=[];for(const i of e){if(t.has(i.id))throw new Error(`Document ID ${i.id} has been deleted and cannot be re-added.`);if(this.#t.isAdded(i.id))throw new Error(`Document ID ${i.id} already exists.`);const e=this.#m(i),o=[],a=[];for(const t of e)t.length>1?o.push(t):1===t.length&&a.push(t);o.length>0&&s.push({id:i.id,tokens:o}),a.length>0&&n.push({id:i.id,tokens:a})}let i=0,o=0;if(s.length>0){await this.#r.appendBatch(d,s);for(const e of s)i+=e.tokens.length}if(n.length>0){await this.#r.appendBatch(c,n);for(const e of n)o+=e.tokens.length}for(const t of e)this.#t.addAddedId(t.id);this.#g?(this.#f.word+=i,this.#f.char+=o):(i>0&&await this.#l("word",i),o>0&&await this.#l("char",o),await this.#t.save())}async#l(e,t){const s="word"===e?d:c,n=await this.#r.getCurrentSize(s),i="word"===e?this.#c.wordSegmentTokenThreshold||1e5:this.#c.charSegmentTokenThreshold||5e5,o="word"===e?this.#c.minWordTokenSave||0:this.#c.minCharTokenSave||0,a=this.#t.getLastSegmentInfo(e);let r,g,f,l;const w=()=>{const t=this.#t.getSegments(e).length+1;return`${e}_seg_${t}.bin`};if(a){const e=a.tokenCount;e>=i||e+t>=i?(r=w(),f=!0,g=a.end,l=t):(r=a.filename,f=!1,g=a.start,l=e+t)}else r=w(),f=!0,g=0,l=t;if(l<o)return void this.#t.updateSegment(e,r,g,n,l,f);const m=await this.#r.readRange(s,g,n);let u=this.#h.get(r);u||(u=new h(r,this.#e),this.#h.set(r,u)),await u.buildAndSave(m),this.#t.updateSegment(e,r,g,n,l,f)}async search(e,t){this.#d||await this.init();const s="string"==typeof e?{text:e}:e,n=this.#u(s),i=n.filter(e=>e.length>1),o=n.filter(e=>1===e.length),a=this.#t.getDeletedIds(),r=new Map,d=new Map,c=e=>{const t=this.#t.getSegments(e);for(const e of t){const t=e.filename;!this.#h.has(t)&&!d.has(t)&&d.set(t,new h(t,this.#e))}};c("word"),c("char"),await Promise.all(Array.from(d.entries()).map(([e,t])=>t.loadIndex().then(s=>{s&&this.#h.set(e,t)})));const g=async(e,t)=>{if(0===t.length)return;const s=this.#t.getSegments(e);for(const e of s){const s=e.filename,n=this.#h.get(s);if(n)for(const e of t){const t=n.search(e),s=1+.1*e.length;for(const n of t)if(!a.has(n))if(r.has(n)){const t=r.get(n);t.score+=s,t.tokens.add(e)}else r.set(n,{score:0,tokens:new Set([e])})}}};await g("word",i),await g("char",o);const f=[];return r.forEach((e,t)=>{f.push({id:t,score:e.score,tokens:Array.from(e.tokens)})}),f.sort((e,t)=>t.score-e.score),"number"==typeof t&&t>0?f.slice(0,t):f}async removeDocument(e){this.#d||await this.init(),this.#t.addDeletedId(e),this.#t.removeAddedId(e),await this.#t.save()}async clearAll(){await this.#e.clearAll(),this.#h.clear(),this.#t.reset(),this.#d=!1,this.#g=!1,this.#f={word:0,char:0}}async getStatus(){return this.#d||await this.init(),{wordSegments:this.#t.getSegments("word").length,charSegments:this.#t.getSegments("char").length,deleted:this.#t.getDeletedIds().size,wordCacheSize:await this.#r.getCurrentSize(d),charCacheSize:await this.#r.getCurrentSize(c),inBatch:this.#g}}async hasDocument(e){return this.#d||await this.init(),this.#t.hasDocument(e)}},exports.hash=r,exports.murmur3_32=r;

package/lib/core.d.ts

@@ -0,0 +1,63 @@
+import { ISearchEngineConfig, IDocument, IDocumentBase, IResult } from './type';
+
+/**
+ * 核心搜索引擎类 (多实例支持)
+ */
+declare class SearchEngine {
+    #private;
+    constructor(config: ISearchEngineConfig);
+    init(): Promise<void>;
+    /**
+     * 开启批处理
+     * 批处理期间 addDocuments 只写入缓存，不触发索引段构建
+     */
+    startBatch(): void;
+    /**
+     * 结束批处理
+     * 触发索引构建检查并保存元数据
+     */
+    endBatch(): Promise<void>;
+    addDocument<T extends IDocument = IDocument>(doc: T): Promise<void>;
+    /**
+     * 添加单个文档，如果文档ID已存在则跳过
+     * 用于在批量添加中途出错后的恢复添加行为，也可直接用于单个文档添加
+     */
+    addDocumentIfMissing<T extends IDocument = IDocument>(doc: T): Promise<void>;
+    /**
+     * 添加多个文档，跳过已存在的文档ID
+     * 用于在批量添加中途出错后的恢复添加行为，也可直接用于批量添加
+     */
+    addDocumentsIfMissing<T extends IDocument = IDocument>(docs: T[]): Promise<void>;
+    addDocuments<T extends IDocument = IDocument>(docs: T[]): Promise<void>;
+    search<T extends IDocumentBase = IDocumentBase>(query: T | string, limit?: number): Promise<IResult[]>;
+    removeDocument(id: number): Promise<void>;
+    clearAll(): Promise<void>;
+    getStatus(): Promise<{
+        wordSegments: number;
+        charSegments: number;
+        deleted: number;
+        wordCacheSize: number;
+        charCacheSize: number;
+        inBatch: boolean;
+    }>;
+    /**
+     * 检查文档ID是否曾经添加过（包括已删除的）
+     * @param id 文档ID
+     * @returns 文档是否曾经添加过的布尔值
+     */
+    hasDocument(id: number): Promise<boolean>;
+}
+
+/**
+ * MurmurHash3 32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32位MurmurHash3哈希值
+ * @param str 要哈希的字符串
+ * @param h
+ * @returns 32位无符号哈希值
+ */
+declare function murmur3_32(str: string, h?: number): number;
+
+export { SearchEngine, murmur3_32 as hash, murmur3_32 };

package/lib/core.js

@@ -0,0 +1 @@
+import{BrowserStorage as e}from'./browser';import{NodeStorage as t}from'./node';const s="search_meta.json",n="deleted_ids.bin",i="added_ids.bin";class o{#e;#t={wordSegments:[],charSegments:[]};#s=new Set;#n=new Set;constructor(e){this.#e=e}async load(){const e=await this.#e.read(s);if(e){const t=(new TextDecoder).decode(e);this.#t=JSON.parse(t)}else this.#t={wordSegments:[],charSegments:[]};const t=await this.#e.read(n);if(t){const e=new DataView(t);let s=0;const n=t.byteLength;for(;s<n&&!(s+4>n);){const t=e.getUint32(s,!0);this.#s.add(t),s+=4,s<n&&30===e.getUint8(s)&&(s+=1)}}const o=await this.#e.read(i);if(o){const e=new DataView(o);let t=0;const s=o.byteLength;for(;t<s&&!(t+4>s);){const n=e.getUint32(t,!0);this.#n.add(n),t+=4,t<s&&30===e.getUint8(t)&&(t+=1)}}}async save(){const e=JSON.stringify(this.#t);if(await this.#e.write(s,(new TextEncoder).encode(e).buffer),0===this.#s.size)await this.#e.remove(n);else{const e=4*this.#s.size+this.#s.size,t=new ArrayBuffer(e),s=new DataView(t);let i=0;for(const e of this.#s)s.setUint32(i,e,!0),i+=4,s.setUint8(i,30),i+=1;await this.#e.write(n,t)}if(0===this.#n.size)await this.#e.remove(i);else{const e=4*this.#n.size+this.#n.size,t=new ArrayBuffer(e),s=new DataView(t);let n=0;for(const e of this.#n)s.setUint32(n,e,!0),n+=4,s.setUint8(n,30),n+=1;await this.#e.write(i,t)}}getSegments(e){return"word"===e?this.#t.wordSegments:this.#t.charSegments}getDeletedIds(){return this.#s}addDeletedId(e){this.#s.add(e)}isDeleted(e){return this.#s.has(e)}addAddedId(e){this.#n.add(e)}removeAddedId(e){this.#n.delete(e)}isAdded(e){return this.#n.has(e)}getAddedIds(){return this.#n}hasDocument(e){return this.#n.has(e)||this.#s.has(e)}getLastSegmentInfo(e){const t=this.getSegments(e);return 0===t.length?null:t[t.length-1]}updateSegment(e,t,s,n,i,o){const a="word"===e?this.#t.wordSegments:this.#t.charSegments;if(o)a.push({filename:t,start:s,end:n,tokenCount:i});else{const e=a[a.length-1];e&&e.filename===t&&(e.end=n,e.tokenCount=i)}}reset(){this.#t={wordSegments:[],charSegments:[]},this.#s.clear(),this.#n.clear()}}class a{static SEPARATOR=30;#e;constructor(e){this.#e=e}async appendBatch(e,t){if(0===t.length)return await this.#e.getFileSize(e);const s=new TextEncoder;let n=0;for(const e of t){n+=8;for(const t of e.tokens){n+=2+Math.min(s.encode(t).byteLength,65535)}n+=1}const i=new Uint8Array(n);let o=0;for(const e of t){const t=[];for(const n of e.tokens){const e=s.encode(n),i=e.byteLength>65535?e.slice(0,65535):e;t.push(i)}const n=new DataView(i.buffer,o);n.setUint32(0,e.id,!0),n.setUint32(4,t.length,!0),o+=8;for(const e of t)new DataView(i.buffer,o).setUint16(0,e.byteLength,!0),o+=2,i.set(e,o),o+=e.byteLength;i[o++]=a.SEPARATOR}return await this.#e.append(e,i.buffer),await this.#e.getFileSize(e)}async readRange(e,t,s){const n=await this.#e.readRange(e,t,s);if(!n||0===n.byteLength)return[];const i=new DataView(n),o=new Uint8Array(n),r=new TextDecoder,h=[];let d=0;const c=n.byteLength;for(;d<c&&!(d+8>c);){const e=i.getUint32(d,!0);d+=4;const t=i.getUint32(d,!0);d+=4;const s=[];for(let e=0;e<t&&!(d+2>c);e++){const e=i.getUint16(d,!0);if(d+=2,d+e>c)break;const t=new Uint8Array(n,d,e);s.push(r.decode(t)),d+=e}d<c&&o[d]===a.SEPARATOR&&(d+=1),h.push({id:e,tokens:s})}return h}async getCurrentSize(e){return await this.#e.getFileSize(e)}}function r(e,t=305419896){const s=e.length,n=s>>2;let i=0;for(;i<n;){let s=255&e.charCodeAt(i)|(255&e.charCodeAt(++i))<<8|(255&e.charCodeAt(++i))<<16|(255&e.charCodeAt(++i))<<24;++i,s=3432918353*(65535&s)+((3432918353*(s>>>16)&65535)<<16)&4294967295,s=s<<15|s>>>17,s=461845907*(65535&s)+((461845907*(s>>>16)&65535)<<16)&4294967295,t=27492+(65535&(t=5*(65535&(t=(t^=s)<<13|t>>>19))+((5*(t>>>16)&65535)<<16)&4294967295))+(((t>>>16)+58964&65535)<<16)}let o=0;const a=3&s;return a>0&&(a>=3&&(o^=(255&e.charCodeAt(i+2))<<16),a>=2&&(o^=(255&e.charCodeAt(i+1))<<8),a>=1&&(o^=255&e.charCodeAt(i)),o=3432918353*(65535&o)+((3432918353*(o>>>16)&65535)<<16)&4294967295,o=o<<15|o>>>17,o=461845907*(65535&o)+((461845907*(o>>>16)&65535)<<16)&4294967295,t^=o),t^=s,t=2246822507*(65535&(t^=t>>>16))+((2246822507*(t>>>16)&65535)<<16)&4294967295,t=3266489909*(65535&(t^=t>>>13))+((3266489909*(t>>>16)&65535)<<16)&4294967295,(t^=t>>>16)>>>0}class h{#i;#e;#o=null;#a=null;static hash(e){return r(e)}constructor(e,t){this.#i=e,this.#e=t}async loadIndex(){return!!this.#o||(this.#o=await this.#e.read(this.#i),!!this.#o&&(this.#a=new DataView(this.#o),!0))}async buildAndSave(e){const t=new Map;for(const s of e){const e=new Map;for(const n of s.tokens)e.has(n)||(e.set(n,!0),t.has(n)||t.set(n,{hash:h.hash(n),postings:[]}),t.get(n).postings.push(s.id))}const s=Array.from(t.entries());s.sort(([e,{hash:t}],[s,{hash:n}])=>t!==n?t-n:e.localeCompare(s));let n=0,i=0;for(const[e,{postings:t}]of s)n+=t.length,i+=e.length+1;const o=20*s.length,a=12+o+4*n,r=new ArrayBuffer(a+i),d=new DataView(r);d.setUint32(0,1229866072),d.setUint32(4,s.length),d.setUint32(8,a);let c=12,g=12+o,f=a;for(const[e,{hash:t,postings:n}]of s){d.setUint32(c,t),d.setUint32(c+4,e.length),d.setUint32(c+8,f),d.setUint32(c+12,g),d.setUint32(c+16,n.length),c+=20;for(let e=0;e<n.length;e++)d.setUint32(g,n[e],!0),g+=4;const s=(new TextEncoder).encode(e);for(let e=0;e<s.length;e++)d.setUint8(f++,s[e]);d.setUint8(f++,0)}await this.#e.write(this.#i,r),this.#o=r,this.#a=d}search(e){if(!this.#a||!this.#o)return[];const t=h.hash(e),s=this.#a.getUint32(4);let n=0,i=s-1;const o=12,a=20,r=new TextDecoder;for(;n<=i;){const h=n+i>>>1,d=o+h*a,c=this.#a.getUint32(d);if(c<t)n=h+1;else{if(!(c>t)){if(!(h>0&&this.#a.getUint32(o+(h-1)*a)===t||h<s-1&&this.#a.getUint32(o+(h+1)*a)===t)){const e=this.#a.getUint32(o+h*a+12),t=this.#a.getUint32(o+h*a+16),s=[];for(let n=0;n<t;n++)s.push(this.#a.getUint32(e+4*n,!0));return s}let n=h;for(;n>0;){const e=o+(n-1)*a;if(this.#a.getUint32(e)!==t)break;n--}for(let i=n;i<s;i++){const s=o+i*a;if(this.#a.getUint32(s)!==t)break;const n=this.#a.getUint32(s+4),h=this.#a.getUint32(s+8),d=new Uint8Array(this.#o,h,n);if(r.decode(d)===e){const e=this.#a.getUint32(s+12),t=this.#a.getUint32(s+16),n=[];for(let s=0;s<t;s++)n.push(this.#a.getUint32(e+4*s,!0));return n}}return[]}i=h-1}}return[]}}const d="word_cache.bin",c="char_cache.bin";class g{#e;#t;#r;#h;#d=!1;#c;#g=!1;#f={word:0,char:0};constructor(s){if(!s.baseDir)throw new Error("SearchEngine requires 'baseDir' in config.");if(this.#c={wordSegmentTokenThreshold:1e5,charSegmentTokenThreshold:5e5,minWordTokenSave:0,minCharTokenSave:0,...s},(this.#c.minWordTokenSave||0)>=(this.#c.wordSegmentTokenThreshold||1e5))throw new Error("minWordTokenSave must be less than wordSegmentTokenThreshold");if((this.#c.minCharTokenSave||0)>=(this.#c.charSegmentTokenThreshold||5e5))throw new Error("minCharTokenSave must be less than charSegmentTokenThreshold");let n=null;if(this.#c.storage&&("object"==typeof this.#c.storage?n=this.#c.storage:"browser"===this.#c.storage?n=new e(this.#c.baseDir):"node"===this.#c.storage&&(n=new t(this.#c.baseDir))),!n){const s=typeof navigator<"u"&&navigator?.storage?.getDirectory instanceof Function,i=typeof process<"u"&&null!=process.versions&&null!=process.versions.node;s?n=new e(this.#c.baseDir):i&&(n=new t(this.#c.baseDir))}if(!n)throw new Error('Storage initialization failed. Please configure "storage" explicitly or ensure you are in a supported environment (Browser/Node).');this.#e=n,this.#t=new o(this.#e),this.#r=new a(this.#e),this.#h=new Map}async init(){if(this.#d)return;await this.#t.load();const e=[...this.#t.getSegments("word"),...this.#t.getSegments("char")];for(const t of e)this.#h.has(t.filename)||this.#h.set(t.filename,new h(t.filename,this.#e)),await this.#h.get(t.filename).loadIndex();this.#d=!0}startBatch(){this.#g=!0,this.#f={word:0,char:0}}async endBatch(){this.#g=!1,this.#f.word>0&&await this.#l("word",this.#f.word),this.#f.char>0&&await this.#l("char",this.#f.char),this.#f={word:0,char:0},await this.#t.save()}#w(e){try{if(typeof Intl<"u"&&"function"==typeof Intl.Segmenter&&"function"==typeof Array.from){const t=new Intl.Segmenter([],{granularity:"word"}).segment(e);if("object"==typeof t&&null!==t)return Array.from(t).filter(e=>e?.isWordLike).map(e=>e?.segment?.toLowerCase()||"")}}catch{}return e.toLowerCase().split(/[^a-z0-9\u4e00-\u9fa5]+/g).filter(e=>e.length>0)}#m(e){return this.#c.indexingTokenizer?this.#c.indexingTokenizer(e):this.#w(e.text)}#u(e){return this.#c.searchTokenizer?this.#c.searchTokenizer(e):this.#m(e)}async addDocument(e){return this.addDocuments([e])}async addDocumentIfMissing(e){return this.addDocumentsIfMissing([e])}async addDocumentsIfMissing(e){if(this.#d||await this.init(),0===e.length)return;const t=this.#t.getDeletedIds(),s=[],n=[],i=[];for(const o of e){if(t.has(o.id)||this.#t.isAdded(o.id))continue;const e=this.#m(o),a=[],r=[];for(const t of e)t.length>1?a.push(t):1===t.length&&r.push(t);a.length>0&&s.push({id:o.id,tokens:a}),r.length>0&&n.push({id:o.id,tokens:r}),i.push(o)}if(0===i.length)return;let o=0,a=0;if(s.length>0){await this.#r.appendBatch(d,s);for(const e of s)o+=e.tokens.length}if(n.length>0){await this.#r.appendBatch(c,n);for(const e of n)a+=e.tokens.length}for(const e of i)this.#t.addAddedId(e.id);this.#g?(this.#f.word+=o,this.#f.char+=a):(o>0&&await this.#l("word",o),a>0&&await this.#l("char",a),await this.#t.save())}async addDocuments(e){if(this.#d||await this.init(),0===e.length)return;const t=this.#t.getDeletedIds(),s=[],n=[];for(const i of e){if(t.has(i.id))throw new Error(`Document ID ${i.id} has been deleted and cannot be re-added.`);if(this.#t.isAdded(i.id))throw new Error(`Document ID ${i.id} already exists.`);const e=this.#m(i),o=[],a=[];for(const t of e)t.length>1?o.push(t):1===t.length&&a.push(t);o.length>0&&s.push({id:i.id,tokens:o}),a.length>0&&n.push({id:i.id,tokens:a})}let i=0,o=0;if(s.length>0){await this.#r.appendBatch(d,s);for(const e of s)i+=e.tokens.length}if(n.length>0){await this.#r.appendBatch(c,n);for(const e of n)o+=e.tokens.length}for(const t of e)this.#t.addAddedId(t.id);this.#g?(this.#f.word+=i,this.#f.char+=o):(i>0&&await this.#l("word",i),o>0&&await this.#l("char",o),await this.#t.save())}async#l(e,t){const s="word"===e?d:c,n=await this.#r.getCurrentSize(s),i="word"===e?this.#c.wordSegmentTokenThreshold||1e5:this.#c.charSegmentTokenThreshold||5e5,o="word"===e?this.#c.minWordTokenSave||0:this.#c.minCharTokenSave||0,a=this.#t.getLastSegmentInfo(e);let r,g,f,l;const w=()=>{const t=this.#t.getSegments(e).length+1;return`${e}_seg_${t}.bin`};if(a){const e=a.tokenCount;e>=i||e+t>=i?(r=w(),f=!0,g=a.end,l=t):(r=a.filename,f=!1,g=a.start,l=e+t)}else r=w(),f=!0,g=0,l=t;if(l<o)return void this.#t.updateSegment(e,r,g,n,l,f);const m=await this.#r.readRange(s,g,n);let u=this.#h.get(r);u||(u=new h(r,this.#e),this.#h.set(r,u)),await u.buildAndSave(m),this.#t.updateSegment(e,r,g,n,l,f)}async search(e,t){this.#d||await this.init();const s="string"==typeof e?{text:e}:e,n=this.#u(s),i=n.filter(e=>e.length>1),o=n.filter(e=>1===e.length),a=this.#t.getDeletedIds(),r=new Map,d=new Map,c=e=>{const t=this.#t.getSegments(e);for(const e of t){const t=e.filename;!this.#h.has(t)&&!d.has(t)&&d.set(t,new h(t,this.#e))}};c("word"),c("char"),await Promise.all(Array.from(d.entries()).map(([e,t])=>t.loadIndex().then(s=>{s&&this.#h.set(e,t)})));const g=async(e,t)=>{if(0===t.length)return;const s=this.#t.getSegments(e);for(const e of s){const s=e.filename,n=this.#h.get(s);if(n)for(const e of t){const t=n.search(e),s=1+.1*e.length;for(const n of t)if(!a.has(n))if(r.has(n)){const t=r.get(n);t.score+=s,t.tokens.add(e)}else r.set(n,{score:0,tokens:new Set([e])})}}};await g("word",i),await g("char",o);const f=[];return r.forEach((e,t)=>{f.push({id:t,score:e.score,tokens:Array.from(e.tokens)})}),f.sort((e,t)=>t.score-e.score),"number"==typeof t&&t>0?f.slice(0,t):f}async removeDocument(e){this.#d||await this.init(),this.#t.addDeletedId(e),this.#t.removeAddedId(e),await this.#t.save()}async clearAll(){await this.#e.clearAll(),this.#h.clear(),this.#t.reset(),this.#d=!1,this.#g=!1,this.#f={word:0,char:0}}async getStatus(){return this.#d||await this.init(),{wordSegments:this.#t.getSegments("word").length,charSegments:this.#t.getSegments("char").length,deleted:this.#t.getDeletedIds().size,wordCacheSize:await this.#r.getCurrentSize(d),charCacheSize:await this.#r.getCurrentSize(c),inBatch:this.#g}}async hasDocument(e){return this.#d||await this.init(),this.#t.hasDocument(e)}}export{g as SearchEngine,r as hash,r as murmur3_32};

package/lib/index.cjs

@@ -1 +1 @@
-"use strict";Object.create,Object.defineProperty,Object.getOwnPropertyDescriptor,Object.getOwnPropertyNames,Object.getPrototypeOf,Object.prototype.hasOwnProperty;class t{#t;constructor(t){this.#t=t}async#e(){return await(await navigator.storage.getDirectory()).getDirectoryHandle(this.#t,{create:!0})}async write(t,e){const s=await(await(await this.#e()).getFileHandle(t,{create:!0})).createWritable();await s.write(e),await s.close()}async append(t,e){const s=await this.#e();let i;try{i=await s.getFileHandle(t,{create:!0})}catch{i=await s.getFileHandle(t,{create:!0})}const n=await i.getFile(),a=await i.createWritable({keepExistingData:!0});await a.seek(n.size),await a.write(e),await a.close()}async read(t){const e=await this.#e();try{return await(await(await e.getFileHandle(t)).getFile()).arrayBuffer()}catch{return null}}async readRange(t,e,s){const i=await this.#e();try{return await(await(await i.getFileHandle(t)).getFile()).slice(e,s).arrayBuffer()}catch{return null}}async remove(t){const e=await this.#e();try{await e.removeEntry(t)}catch{}}async listFiles(){const t=await this.#e(),e=[];for await(const s of t.keys())e.push(s);return e}async clearAll(){const t=await this.#e();for await(const e of t.keys())await t.removeEntry(e,{recursive:!0})}async getFileSize(t){const e=await this.#e();try{return(await(await e.getFileHandle(t)).getFile()).size}catch{return 0}}}class e{#s=null;#i=null;#t;#n="";constructor(t){this.#t=t}async#a(){if(this.#s)return;const t=await import("node:fs"),e=await import("node:path");this.#s=t.promises,this.#i=e.default||e,this.#n=this.#i.join(process.cwd(),this.#t);try{await this.#s.access(this.#n)}catch{await this.#s.mkdir(this.#n,{recursive:!0})}}#r(t){return this.#i.join(this.#n,t)}async write(t,e){await this.#a(),await this.#s.writeFile(this.#r(t),Buffer.from(e))}async append(t,e){await this.#a(),await this.#s.appendFile(this.#r(t),Buffer.from(e))}async read(t){await this.#a();try{const e=await this.#s.readFile(this.#r(t));return e.buffer.slice(e.byteOffset,e.byteOffset+e.byteLength)}catch{return null}}async readRange(t,e,s){await this.#a();try{const i=await this.#s.open(this.#r(t),"r"),n=s-e,a=Buffer.alloc(n);return await i.read(a,0,n,e),await i.close(),a.buffer.slice(a.byteOffset,a.byteOffset+a.byteLength)}catch{return null}}async remove(t){await this.#a();try{await this.#s.unlink(this.#r(t))}catch{}}async listFiles(){await this.#a();try{return await this.#s.readdir(this.#n)}catch{return[]}}async clearAll(){await this.#a();try{const t=await this.#s.readdir(this.#n);for(const e of t)await this.#s.unlink(this.#i.join(this.#n,e))}catch{}}async getFileSize(t){await this.#a();try{return(await this.#s.stat(this.#r(t))).size}catch{return 0}}}const s="search_meta.json",i="deleted_ids.bin",n="added_ids.bin";class a{#o;#h={wordSegments:[],charSegments:[]};#c=new Set;#d=new Set;constructor(t){this.#o=t}async load(){const t=await this.#o.read(s);if(t){const e=(new TextDecoder).decode(t);this.#h=JSON.parse(e)}else this.#h={wordSegments:[],charSegments:[]};const e=await this.#o.read(i);if(e){const t=new DataView(e);let s=0;const i=e.byteLength;for(;s<i&&!(s+4>i);){const e=t.getUint32(s,!0);this.#c.add(e),s+=4,s<i&&30===t.getUint8(s)&&(s+=1)}}const a=await this.#o.read(n);if(a){const t=new DataView(a);let e=0;const s=a.byteLength;for(;e<s&&!(e+4>s);){const i=t.getUint32(e,!0);this.#d.add(i),e+=4,e<s&&30===t.getUint8(e)&&(e+=1)}}}async save(){const t=JSON.stringify(this.#h);if(await this.#o.write(s,(new TextEncoder).encode(t).buffer),0===this.#c.size)await this.#o.remove(i);else{const t=4*this.#c.size+this.#c.size,e=new ArrayBuffer(t),s=new DataView(e);let n=0;for(const t of this.#c)s.setUint32(n,t,!0),n+=4,s.setUint8(n,30),n+=1;await this.#o.write(i,e)}if(0===this.#d.size)await this.#o.remove(n);else{const t=4*this.#d.size+this.#d.size,e=new ArrayBuffer(t),s=new DataView(e);let i=0;for(const t of this.#d)s.setUint32(i,t,!0),i+=4,s.setUint8(i,30),i+=1;await this.#o.write(n,e)}}getSegments(t){return"word"===t?this.#h.wordSegments:this.#h.charSegments}getDeletedIds(){return this.#c}addDeletedId(t){this.#c.add(t)}isDeleted(t){return this.#c.has(t)}addAddedId(t){this.#d.add(t)}removeAddedId(t){this.#d.delete(t)}isAdded(t){return this.#d.has(t)}getAddedIds(){return this.#d}hasDocument(t){return this.#d.has(t)||this.#c.has(t)}getLastSegmentInfo(t){const e=this.getSegments(t);return 0===e.length?null:e[e.length-1]}updateSegment(t,e,s,i,n,a){const r="word"===t?this.#h.wordSegments:this.#h.charSegments;if(a)r.push({filename:e,start:s,end:i,tokenCount:n});else{const t=r[r.length-1];t&&t.filename===e&&(t.end=i,t.tokenCount=n)}}reset(){this.#h={wordSegments:[],charSegments:[]},this.#c.clear(),this.#d.clear()}}class r{static SEPARATOR=30;#o;constructor(t){this.#o=t}async appendBatch(t,e){if(0===e.length)return await this.#o.getFileSize(t);const s=new TextEncoder;let i=0;for(const t of e){i+=8;for(const e of t.tokens){i+=2+Math.min(s.encode(e).byteLength,65535)}i+=1}const n=new Uint8Array(i);let a=0;for(const t of e){const e=[];for(const i of t.tokens){const t=s.encode(i),n=t.byteLength>65535?t.slice(0,65535):t;e.push(n)}const i=new DataView(n.buffer,a);i.setUint32(0,t.id,!0),i.setUint32(4,e.length,!0),a+=8;for(const t of e)new DataView(n.buffer,a).setUint16(0,t.byteLength,!0),a+=2,n.set(t,a),a+=t.byteLength;n[a++]=r.SEPARATOR}return await this.#o.append(t,n.buffer),await this.#o.getFileSize(t)}async readRange(t,e,s){const i=await this.#o.readRange(t,e,s);if(!i||0===i.byteLength)return[];const n=new DataView(i),a=new Uint8Array(i),o=new TextDecoder,h=[];let c=0;const d=i.byteLength;for(;c<d&&!(c+8>d);){const t=n.getUint32(c,!0);c+=4;const e=n.getUint32(c,!0);c+=4;const s=[];for(let t=0;t<e&&!(c+2>d);t++){const t=n.getUint16(c,!0);if(c+=2,c+t>d)break;const e=new Uint8Array(i,c,t);s.push(o.decode(e)),c+=t}c<d&&a[c]===r.SEPARATOR&&(c+=1),h.push({id:t,tokens:s})}return h}async getCurrentSize(t){return await this.#o.getFileSize(t)}}class o{#g;#o;#l=null;#f=null;static hash(t){let e=5381;for(let s=0;s<t.length;s++)e=(e<<5)+e^t.charCodeAt(s);return e>>>0}constructor(t,e){this.#g=t,this.#o=e}async loadIndex(){return!!this.#l||(this.#l=await this.#o.read(this.#g),!!this.#l&&(this.#f=new DataView(this.#l),!0))}async buildAndSave(t){const e=new Map;for(const s of t){const t=new Map;for(const i of s.tokens)if(!t.has(i)){t.set(i,!0);const n=o.hash(i);e.has(n)||e.set(n,[]),e.get(n).push(s.id)}}const s=Array.from(e.keys()).sort((t,e)=>t-e);let i=0;const n=new Array(s.length);for(let t=0;t<s.length;t++){const a=s[t],r=e.get(a);n[t]=r,i+=r.length}const a=12*s.length,r=new ArrayBuffer(8+a+4*i),h=new DataView(r);h.setUint32(0,1229866072),h.setUint32(4,s.length);let c=8,d=8+a;for(let t=0;t<s.length;t++){const e=s[t],i=n[t];h.setUint32(c,e),h.setUint32(c+4,d),h.setUint32(c+8,i.length),c+=12;for(let t=0;t<i.length;t++)h.setUint32(d,i[t],!0),d+=4}await this.#o.write(this.#g,r),this.#l=r,this.#f=h}search(t){if(!this.#f||!this.#l)return[];const e=o.hash(t);let s=0,i=this.#f.getUint32(4)-1;for(;s<=i;){const t=s+i>>>1,n=8+12*t,a=this.#f.getUint32(n);if(a<e)s=t+1;else{if(!(a>e)){const t=this.#f.getUint32(n+4),e=this.#f.getUint32(n+8),s=[];for(let i=0;i<e;i++)s.push(this.#f.getUint32(t+4*i,!0));return s}i=t-1}}return[]}}const h="word_cache.bin",c="char_cache.bin";class d{#o;#h;#w;#u;#m=!1;#y;#p=!1;#D={word:0,char:0};constructor(s){if(!s.baseDir)throw new Error("SearchEngine requires 'baseDir' in config.");if(this.#y={wordSegmentTokenThreshold:1e5,charSegmentTokenThreshold:5e5,minWordTokenSave:0,minCharTokenSave:0,...s},(this.#y.minWordTokenSave||0)>=(this.#y.wordSegmentTokenThreshold||1e5))throw new Error("minWordTokenSave must be less than wordSegmentTokenThreshold");if((this.#y.minCharTokenSave||0)>=(this.#y.charSegmentTokenThreshold||5e5))throw new Error("minCharTokenSave must be less than charSegmentTokenThreshold");let i=null;if(this.#y.storage&&("object"==typeof this.#y.storage?i=this.#y.storage:"browser"===this.#y.storage?i=new t(this.#y.baseDir):"node"===this.#y.storage&&(i=new e(this.#y.baseDir))),!i){const s=typeof navigator<"u"&&navigator?.storage?.getDirectory instanceof Function,n=typeof process<"u"&&null!=process.versions&&null!=process.versions.node;s?i=new t(this.#y.baseDir):n&&(i=new e(this.#y.baseDir))}if(!i)throw new Error('Storage initialization failed. Please configure "storage" explicitly or ensure you are in a supported environment (Browser/Node).');this.#o=i,this.#h=new a(this.#o),this.#w=new r(this.#o),this.#u=new Map}async init(){if(this.#m)return;await this.#h.load();const t=[...this.#h.getSegments("word"),...this.#h.getSegments("char")];for(const e of t)this.#u.has(e.filename)||this.#u.set(e.filename,new o(e.filename,this.#o)),await this.#u.get(e.filename).loadIndex();this.#m=!0}startBatch(){this.#p=!0,this.#D={word:0,char:0}}async endBatch(){this.#p=!1,this.#D.word>0&&await this.#S("word",this.#D.word),this.#D.char>0&&await this.#S("char",this.#D.char),this.#D={word:0,char:0},await this.#h.save()}#k(t){try{if(typeof Intl<"u"&&"function"==typeof Intl.Segmenter&&"function"==typeof Array.from){const e=new Intl.Segmenter([],{granularity:"word"}).segment(t);if("object"==typeof e&&null!==e)return Array.from(e).filter(t=>t?.isWordLike).map(t=>t?.segment?.toLowerCase()||"")}}catch{}return t.toLowerCase().split(/[^a-z0-9\u4e00-\u9fa5]+/g).filter(t=>t.length>0)}#I(t){return this.#y.indexingTokenizer?this.#y.indexingTokenizer(t):this.#k(t.text)}#b(t){return this.#y.searchTokenizer?this.#y.searchTokenizer(t):this.#I(t)}async addDocument(t){return this.addDocuments([t])}async addDocumentIfMissing(t){return this.addDocumentsIfMissing([t])}async addDocumentsIfMissing(t){if(this.#m||await this.init(),0===t.length)return;const e=this.#h.getDeletedIds(),s=[],i=[],n=[];for(const a of t){if(e.has(a.id)||this.#h.isAdded(a.id))continue;const t=this.#I(a),r=[],o=[];for(const e of t)e.length>1?r.push(e):1===e.length&&o.push(e);r.length>0&&s.push({id:a.id,tokens:r}),o.length>0&&i.push({id:a.id,tokens:o}),n.push(a)}if(0===n.length)return;let a=0,r=0;if(s.length>0){await this.#w.appendBatch(h,s);for(const t of s)a+=t.tokens.length}if(i.length>0){await this.#w.appendBatch(c,i);for(const t of i)r+=t.tokens.length}for(const t of n)this.#h.addAddedId(t.id);this.#p?(this.#D.word+=a,this.#D.char+=r):(a>0&&await this.#S("word",a),r>0&&await this.#S("char",r),await this.#h.save())}async addDocuments(t){if(this.#m||await this.init(),0===t.length)return;const e=this.#h.getDeletedIds(),s=[],i=[];for(const n of t){if(e.has(n.id))throw new Error(`Document ID ${n.id} has been deleted and cannot be re-added.`);if(this.#h.isAdded(n.id))throw new Error(`Document ID ${n.id} already exists.`);const t=this.#I(n),a=[],r=[];for(const e of t)e.length>1?a.push(e):1===e.length&&r.push(e);a.length>0&&s.push({id:n.id,tokens:a}),r.length>0&&i.push({id:n.id,tokens:r})}let n=0,a=0;if(s.length>0){await this.#w.appendBatch(h,s);for(const t of s)n+=t.tokens.length}if(i.length>0){await this.#w.appendBatch(c,i);for(const t of i)a+=t.tokens.length}for(const e of t)this.#h.addAddedId(e.id);this.#p?(this.#D.word+=n,this.#D.char+=a):(n>0&&await this.#S("word",n),a>0&&await this.#S("char",a),await this.#h.save())}async#S(t,e){const s="word"===t?h:c,i=await this.#w.getCurrentSize(s),n="word"===t?this.#y.wordSegmentTokenThreshold||1e5:this.#y.charSegmentTokenThreshold||5e5,a="word"===t?this.#y.minWordTokenSave||0:this.#y.minCharTokenSave||0,r=this.#h.getLastSegmentInfo(t);let d,g,l,f;const w=()=>{const e=this.#h.getSegments(t).length+1;return`${t}_seg_${e}.bin`};if(r){const t=r.tokenCount;t>=n||t+e>=n?(d=w(),l=!0,g=r.end,f=e):(d=r.filename,l=!1,g=r.start,f=t+e)}else d=w(),l=!0,g=0,f=e;if(f<a)return void this.#h.updateSegment(t,d,g,i,f,l);const u=await this.#w.readRange(s,g,i);let m=this.#u.get(d);m||(m=new o(d,this.#o),this.#u.set(d,m)),await m.buildAndSave(u),this.#h.updateSegment(t,d,g,i,f,l)}async search(t,e){this.#m||await this.init();const s="string"==typeof t?{text:t}:t,i=this.#b(s),n=i.filter(t=>t.length>1),a=i.filter(t=>1===t.length),r=this.#h.getDeletedIds(),h=new Map,c=new Map,d=t=>{const e=this.#h.getSegments(t);for(const t of e){const e=t.filename;!this.#u.has(e)&&!c.has(e)&&c.set(e,new o(e,this.#o))}};d("word"),d("char"),await Promise.all(Array.from(c.entries()).map(([t,e])=>e.loadIndex().then(s=>{s&&this.#u.set(t,e)})));const g=async(t,e)=>{if(0===e.length)return;const s=this.#h.getSegments(t);for(const t of s){const s=t.filename,i=this.#u.get(s);if(i)for(const t of e){const e=i.search(t),s=1+.1*t.length;for(const i of e)if(!r.has(i))if(h.has(i)){const e=h.get(i);e.score+=s,e.tokens.add(t)}else h.set(i,{score:0,tokens:new Set([t])})}}};await g("word",n),await g("char",a);const l=[];return h.forEach((t,e)=>{l.push({id:e,score:t.score,tokens:Array.from(t.tokens)})}),l.sort((t,e)=>e.score-t.score),"number"==typeof e&&e>0?l.slice(0,e):l}async removeDocument(t){this.#m||await this.init(),this.#h.addDeletedId(t),this.#h.removeAddedId(t),await this.#h.save()}async clearAll(){await this.#o.clearAll(),this.#u.clear(),this.#h.reset(),this.#m=!1,this.#p=!1,this.#D={word:0,char:0}}async getStatus(){return this.#m||await this.init(),{wordSegments:this.#h.getSegments("word").length,charSegments:this.#h.getSegments("char").length,deleted:this.#h.getDeletedIds().size,wordCacheSize:await this.#w.getCurrentSize(h),charCacheSize:await this.#w.getCurrentSize(c),inBatch:this.#p}}async hasDocument(t){return this.#m||await this.init(),this.#h.hasDocument(t)}}exports.BrowserStorage=t,exports.NodeStorage=e,exports.SearchEngine=d,exports.SimpleSearch=class{static#T=null;static#v={baseDir:"simple_search_data",wordSegmentTokenThreshold:1e5,minWordTokenSave:0};static configure(t){const e={...this.#v,...t};this.#T=new d(e)}static#z(){return this.#T||(this.#T=new d(this.#v)),this.#T}static async startBatch(){this.#z().startBatch()}static async endBatch(){return this.#z().endBatch()}static async addDocument(t){return this.#z().addDocument(t)}static async addDocumentIfMissing(t){return this.#z().addDocumentIfMissing(t)}static async addDocuments(t){return this.#z().addDocuments(t)}static async addDocumentsIfMissing(t){return this.#z().addDocumentsIfMissing(t)}static async search(t,e){return this.#z().search(t,e)}static async removeDocument(t){return this.#z().removeDocument(t)}static async clearAll(){return this.#z().clearAll()}static async getStatus(){return this.#z().getStatus()}static async hasDocument(t){return this.#z().hasDocument(t)}};
+"use strict";var e=require('./type'),t=require('./core'),r=require('./simple'),o=require('./browser'),n=require('./node');Object.keys(e).forEach(function(t){"default"!==t&&!Object.prototype.hasOwnProperty.call(exports,t)&&Object.defineProperty(exports,t,{enumerable:!0,get:function(){return e[t]}})}),Object.keys(t).forEach(function(e){"default"!==e&&!Object.prototype.hasOwnProperty.call(exports,e)&&Object.defineProperty(exports,e,{enumerable:!0,get:function(){return t[e]}})}),Object.keys(r).forEach(function(e){"default"!==e&&!Object.prototype.hasOwnProperty.call(exports,e)&&Object.defineProperty(exports,e,{enumerable:!0,get:function(){return r[e]}})}),Object.keys(o).forEach(function(e){"default"!==e&&!Object.prototype.hasOwnProperty.call(exports,e)&&Object.defineProperty(exports,e,{enumerable:!0,get:function(){return o[e]}})}),Object.keys(n).forEach(function(e){"default"!==e&&!Object.prototype.hasOwnProperty.call(exports,e)&&Object.defineProperty(exports,e,{enumerable:!0,get:function(){return n[e]}})});

package/lib/index.d.ts

@@ -1,225 +1,5 @@
-/**
- * 核心类型定义
- */
-interface IDocumentBase {
-    text: string;
-}
-interface IDocument extends IDocumentBase {
-    id: number;
-}
-interface IResult {
-    id: number;
-    score: number;
-    tokens: string[];
-}
-interface ISegmentMeta {
-    filename: string;
-    start: number;
-    end: number;
-    tokenCount: number;
-}
-interface IIndexMeta {
-    wordSegments: ISegmentMeta[];
-    charSegments: ISegmentMeta[];
-}
-interface ITokenizedDoc {
-    id: number;
-    tokens: string[];
-}
-type IndexType = 'word' | 'char';
-/**
- * 存储层接口 (外部化)
- */
-interface IStorage {
-    write(filename: string, data: ArrayBuffer): Promise<void>;
-    append(filename: string, data: ArrayBuffer): Promise<void>;
-    read(filename: string): Promise<ArrayBuffer | null>;
-    readRange(filename: string, start: number, end: number): Promise<ArrayBuffer | null>;
-    remove(filename: string): Promise<void>;
-    listFiles(): Promise<string[]>;
-    clearAll(): Promise<void>;
-    getFileSize(filename: string): Promise<number>;
-}
-/**
- * 核心搜索引擎配置
- */
-interface ISearchEngineConfig {
-    /** * 数据存储的基础目录 (必填)
-     * 用于区分不同的搜索引擎实例
-     */
-    baseDir: string;
-    /**
-     * 存储实现配置 (可选)
-     * - 'browser': 强制使用 OPFS (BrowserStorage)
-     * - 'node': 强制使用 Node.js fs (NodeStorage)
-     * - IStorage: 传入自定义的存储实例
-     * - undefined: 自动检测环境
-     */
-    storage?: 'browser' | 'node' | IStorage;
-    /**
-     * 索引时使用的分词器 (算法核心配置)
-     * - 作用: 将文档文本转换为索引用的token序列
-     * - 算法: 自定义实现的分词逻辑，需满足返回字符串数组的格式要求
-     * - 建议: 针对不同语言(中文/英文/日文等)使用专门的分词实现
-     * - 影响: 直接决定索引的粒度和搜索的准确性
-     */
-    indexingTokenizer?: <T extends IDocument = IDocument>(doc: T) => string[];
-    /**
-     * 搜索时使用的分词器 (算法核心配置)
-     * - 作用: 将查询文本转换为搜索用的token序列
-     * - 算法: 自定义实现的分词逻辑，需满足返回字符串数组的格式要求
-     * - 建议: 与indexingTokenizer保持一致的分词策略以确保搜索准确性
-     * - 影响: 直接决定搜索匹配的范围和结果的相关性
-     */
-    searchTokenizer?: <T extends IDocumentBase = IDocumentBase>(doc: T) => string[];
-    /**
-     * 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
-     * - 算法: 基于Token数量的分段策略，当新增Token数加上已有Token数超过阈值时触发分段
-     * - 默认值: 100000
-     * - 影响: 过小会导致索引文件过多，过大可能影响搜索性能
-     */
-    wordSegmentTokenThreshold?: number;
-    /**
-     * 字索引分段阈值 (Token数) - 分段算法配置
-     * - 作用: 控制字索引文件的大小，超过阈值时创建新的索引段
-     * - 算法: 基于Token数量的分段策略，当新增Token数加上已有Token数超过阈值时触发分段
-     * - 默认值: 500000
-     * - 影响: 过小会导致索引文件过多，过大可能影响搜索性能
-     */
-    charSegmentTokenThreshold?: number;
-    /**
-     * 词索引最小保存阈值 (Token数) - 缓存算法配置
-     * - 作用: 控制词索引是否立即写入磁盘，低于阈值时只保存在内存缓存中
-     * - 算法: 基于Token数量的缓存策略，当累计Token数达到阈值时才进行持久化
-     * - 默认值: 0
-     * - 影响: 适当设置可减少磁盘IO次数，提高索引性能
-     */
-    minWordTokenSave?: number;
-    /**
-     * 字索引最小保存阈值 (Token数) - 缓存算法配置
-     * - 作用: 控制字索引是否立即写入磁盘，低于阈值时只保存在内存缓存中
-     * - 算法: 基于Token数量的缓存策略，当累计Token数达到阈值时才进行持久化
-     * - 默认值: 0
-     * - 影响: 适当设置可减少磁盘IO次数，提高索引性能
-     */
-    minCharTokenSave?: number;
-}
-
-/**
- * 核心搜索引擎类 (多实例支持)
- */
-declare class SearchEngine {
-    #private;
-    constructor(config: ISearchEngineConfig);
-    init(): Promise<void>;
-    /**
-     * 开启批处理
-     * 批处理期间 addDocuments 只写入缓存，不触发索引段构建
-     */
-    startBatch(): void;
-    /**
-     * 结束批处理
-     * 触发索引构建检查并保存元数据
-     */
-    endBatch(): Promise<void>;
-    addDocument<T extends IDocument = IDocument>(doc: T): Promise<void>;
-    /**
-     * 添加单个文档，如果文档ID已存在则跳过
-     * 用于在批量添加中途出错后的恢复添加行为，也可直接用于单个文档添加
-     */
-    addDocumentIfMissing<T extends IDocument = IDocument>(doc: T): Promise<void>;
-    /**
-     * 添加多个文档，跳过已存在的文档ID
-     * 用于在批量添加中途出错后的恢复添加行为，也可直接用于批量添加
-     */
-    addDocumentsIfMissing<T extends IDocument = IDocument>(docs: T[]): Promise<void>;
-    addDocuments<T extends IDocument = IDocument>(docs: T[]): Promise<void>;
-    search<T extends IDocumentBase = IDocumentBase>(query: T | string, limit?: number): Promise<IResult[]>;
-    removeDocument(id: number): Promise<void>;
-    clearAll(): Promise<void>;
-    getStatus(): Promise<{
-        wordSegments: number;
-        charSegments: number;
-        deleted: number;
-        wordCacheSize: number;
-        charCacheSize: number;
-        inBatch: boolean;
-    }>;
-    /**
-     * 检查文档ID是否曾经添加过（包括已删除的）
-     * @param id 文档ID
-     * @returns 文档是否曾经添加过的布尔值
-     */
-    hasDocument(id: number): Promise<boolean>;
-}
-
-/**
- * 快速使用封装
- * 提供单例模式和默认配置
- */
-declare class SimpleSearch {
-    #private;
-    /**
-     * 配置并初始化单例
-     */
-    static configure(config: Partial<ISearchEngineConfig>): void;
-    static startBatch(): Promise<void>;
-    static endBatch(): Promise<void>;
-    static addDocument<T extends IDocument = IDocument>(doc: T): Promise<void>;
-    static addDocumentIfMissing<T extends IDocument = IDocument>(doc: T): Promise<void>;
-    static addDocuments<T extends IDocument = IDocument>(docs: T[]): Promise<void>;
-    static addDocumentsIfMissing<T extends IDocument = IDocument>(docs: T[]): Promise<void>;
-    static search<T extends IDocumentBase = IDocumentBase>(query: T | string, limit?: number): Promise<IResult[]>;
-    static removeDocument(id: number): Promise<void>;
-    static clearAll(): Promise<void>;
-    static getStatus(): Promise<{
-        wordSegments: number;
-        charSegments: number;
-        deleted: number;
-        wordCacheSize: number;
-        charCacheSize: number;
-        inBatch: boolean;
-    }>;
-    /**
-     * 检查文档ID是否曾经添加过（包括已删除的）
-     * @param id 文档ID
-     * @returns 文档是否曾经添加过的布尔值
-     */
-    static hasDocument(id: number): Promise<boolean>;
-}
-
-/**
- * 浏览器实现 (OPFS - 子目录隔离)
- * 支持 Main Thread 和 Web Worker
- */
-declare class BrowserStorage implements IStorage {
-    #private;
-    constructor(baseDir: string);
-    write(filename: string, data: ArrayBuffer): Promise<void>;
-    append(filename: string, data: ArrayBuffer): Promise<void>;
-    read(filename: string): Promise<ArrayBuffer | null>;
-    readRange(filename: string, start: number, end: number): Promise<ArrayBuffer | null>;
-    remove(filename: string): Promise<void>;
-    listFiles(): Promise<string[]>;
-    clearAll(): Promise<void>;
-    getFileSize(filename: string): Promise<number>;
-}
-/**
- * Node.js 实现
- */
-declare class NodeStorage implements IStorage {
-    #private;
-    constructor(baseDir: string);
-    write(filename: string, data: ArrayBuffer): Promise<void>;
-    append(filename: string, data: ArrayBuffer): Promise<void>;
-    read(filename: string): Promise<ArrayBuffer | null>;
-    readRange(filename: string, start: number, end: number): Promise<ArrayBuffer | null>;
-    remove(filename: string): Promise<void>;
-    listFiles(): Promise<string[]>;
-    clearAll(): Promise<void>;
-    getFileSize(filename: string): Promise<number>;
-}
-
-export { BrowserStorage, NodeStorage, SearchEngine, SimpleSearch };
-export type { IDocument, IDocumentBase, IIndexMeta, IResult, ISearchEngineConfig, ISegmentMeta, IStorage, ITokenizedDoc, IndexType };
+export * from './type';
+export * from './core';
+export * from './simple';
+export * from './browser';
+export * from './node';

package/lib/index.js

@@ -1 +1 @@
-class t{#t;constructor(t){this.#t=t}async#e(){return await(await navigator.storage.getDirectory()).getDirectoryHandle(this.#t,{create:!0})}async write(t,e){const s=await(await(await this.#e()).getFileHandle(t,{create:!0})).createWritable();await s.write(e),await s.close()}async append(t,e){const s=await this.#e();let i;try{i=await s.getFileHandle(t,{create:!0})}catch{i=await s.getFileHandle(t,{create:!0})}const n=await i.getFile(),a=await i.createWritable({keepExistingData:!0});await a.seek(n.size),await a.write(e),await a.close()}async read(t){const e=await this.#e();try{return await(await(await e.getFileHandle(t)).getFile()).arrayBuffer()}catch{return null}}async readRange(t,e,s){const i=await this.#e();try{return await(await(await i.getFileHandle(t)).getFile()).slice(e,s).arrayBuffer()}catch{return null}}async remove(t){const e=await this.#e();try{await e.removeEntry(t)}catch{}}async listFiles(){const t=await this.#e(),e=[];for await(const s of t.keys())e.push(s);return e}async clearAll(){const t=await this.#e();for await(const e of t.keys())await t.removeEntry(e,{recursive:!0})}async getFileSize(t){const e=await this.#e();try{return(await(await e.getFileHandle(t)).getFile()).size}catch{return 0}}}class e{#s=null;#i=null;#t;#n="";constructor(t){this.#t=t}async#a(){if(this.#s)return;const t=await import("node:fs"),e=await import("node:path");this.#s=t.promises,this.#i=e.default||e,this.#n=this.#i.join(process.cwd(),this.#t);try{await this.#s.access(this.#n)}catch{await this.#s.mkdir(this.#n,{recursive:!0})}}#r(t){return this.#i.join(this.#n,t)}async write(t,e){await this.#a(),await this.#s.writeFile(this.#r(t),Buffer.from(e))}async append(t,e){await this.#a(),await this.#s.appendFile(this.#r(t),Buffer.from(e))}async read(t){await this.#a();try{const e=await this.#s.readFile(this.#r(t));return e.buffer.slice(e.byteOffset,e.byteOffset+e.byteLength)}catch{return null}}async readRange(t,e,s){await this.#a();try{const i=await this.#s.open(this.#r(t),"r"),n=s-e,a=Buffer.alloc(n);return await i.read(a,0,n,e),await i.close(),a.buffer.slice(a.byteOffset,a.byteOffset+a.byteLength)}catch{return null}}async remove(t){await this.#a();try{await this.#s.unlink(this.#r(t))}catch{}}async listFiles(){await this.#a();try{return await this.#s.readdir(this.#n)}catch{return[]}}async clearAll(){await this.#a();try{const t=await this.#s.readdir(this.#n);for(const e of t)await this.#s.unlink(this.#i.join(this.#n,e))}catch{}}async getFileSize(t){await this.#a();try{return(await this.#s.stat(this.#r(t))).size}catch{return 0}}}const s="search_meta.json",i="deleted_ids.bin",n="added_ids.bin";class a{#o;#h={wordSegments:[],charSegments:[]};#c=new Set;#d=new Set;constructor(t){this.#o=t}async load(){const t=await this.#o.read(s);if(t){const e=(new TextDecoder).decode(t);this.#h=JSON.parse(e)}else this.#h={wordSegments:[],charSegments:[]};const e=await this.#o.read(i);if(e){const t=new DataView(e);let s=0;const i=e.byteLength;for(;s<i&&!(s+4>i);){const e=t.getUint32(s,!0);this.#c.add(e),s+=4,s<i&&30===t.getUint8(s)&&(s+=1)}}const a=await this.#o.read(n);if(a){const t=new DataView(a);let e=0;const s=a.byteLength;for(;e<s&&!(e+4>s);){const i=t.getUint32(e,!0);this.#d.add(i),e+=4,e<s&&30===t.getUint8(e)&&(e+=1)}}}async save(){const t=JSON.stringify(this.#h);if(await this.#o.write(s,(new TextEncoder).encode(t).buffer),0===this.#c.size)await this.#o.remove(i);else{const t=4*this.#c.size+this.#c.size,e=new ArrayBuffer(t),s=new DataView(e);let n=0;for(const t of this.#c)s.setUint32(n,t,!0),n+=4,s.setUint8(n,30),n+=1;await this.#o.write(i,e)}if(0===this.#d.size)await this.#o.remove(n);else{const t=4*this.#d.size+this.#d.size,e=new ArrayBuffer(t),s=new DataView(e);let i=0;for(const t of this.#d)s.setUint32(i,t,!0),i+=4,s.setUint8(i,30),i+=1;await this.#o.write(n,e)}}getSegments(t){return"word"===t?this.#h.wordSegments:this.#h.charSegments}getDeletedIds(){return this.#c}addDeletedId(t){this.#c.add(t)}isDeleted(t){return this.#c.has(t)}addAddedId(t){this.#d.add(t)}removeAddedId(t){this.#d.delete(t)}isAdded(t){return this.#d.has(t)}getAddedIds(){return this.#d}hasDocument(t){return this.#d.has(t)||this.#c.has(t)}getLastSegmentInfo(t){const e=this.getSegments(t);return 0===e.length?null:e[e.length-1]}updateSegment(t,e,s,i,n,a){const r="word"===t?this.#h.wordSegments:this.#h.charSegments;if(a)r.push({filename:e,start:s,end:i,tokenCount:n});else{const t=r[r.length-1];t&&t.filename===e&&(t.end=i,t.tokenCount=n)}}reset(){this.#h={wordSegments:[],charSegments:[]},this.#c.clear(),this.#d.clear()}}class r{static SEPARATOR=30;#o;constructor(t){this.#o=t}async appendBatch(t,e){if(0===e.length)return await this.#o.getFileSize(t);const s=new TextEncoder;let i=0;for(const t of e){i+=8;for(const e of t.tokens){i+=2+Math.min(s.encode(e).byteLength,65535)}i+=1}const n=new Uint8Array(i);let a=0;for(const t of e){const e=[];for(const i of t.tokens){const t=s.encode(i),n=t.byteLength>65535?t.slice(0,65535):t;e.push(n)}const i=new DataView(n.buffer,a);i.setUint32(0,t.id,!0),i.setUint32(4,e.length,!0),a+=8;for(const t of e)new DataView(n.buffer,a).setUint16(0,t.byteLength,!0),a+=2,n.set(t,a),a+=t.byteLength;n[a++]=r.SEPARATOR}return await this.#o.append(t,n.buffer),await this.#o.getFileSize(t)}async readRange(t,e,s){const i=await this.#o.readRange(t,e,s);if(!i||0===i.byteLength)return[];const n=new DataView(i),a=new Uint8Array(i),o=new TextDecoder,h=[];let c=0;const d=i.byteLength;for(;c<d&&!(c+8>d);){const t=n.getUint32(c,!0);c+=4;const e=n.getUint32(c,!0);c+=4;const s=[];for(let t=0;t<e&&!(c+2>d);t++){const t=n.getUint16(c,!0);if(c+=2,c+t>d)break;const e=new Uint8Array(i,c,t);s.push(o.decode(e)),c+=t}c<d&&a[c]===r.SEPARATOR&&(c+=1),h.push({id:t,tokens:s})}return h}async getCurrentSize(t){return await this.#o.getFileSize(t)}}class o{#g;#o;#l=null;#f=null;static hash(t){let e=5381;for(let s=0;s<t.length;s++)e=(e<<5)+e^t.charCodeAt(s);return e>>>0}constructor(t,e){this.#g=t,this.#o=e}async loadIndex(){return!!this.#l||(this.#l=await this.#o.read(this.#g),!!this.#l&&(this.#f=new DataView(this.#l),!0))}async buildAndSave(t){const e=new Map;for(const s of t){const t=new Map;for(const i of s.tokens)if(!t.has(i)){t.set(i,!0);const n=o.hash(i);e.has(n)||e.set(n,[]),e.get(n).push(s.id)}}const s=Array.from(e.keys()).sort((t,e)=>t-e);let i=0;const n=new Array(s.length);for(let t=0;t<s.length;t++){const a=s[t],r=e.get(a);n[t]=r,i+=r.length}const a=12*s.length,r=new ArrayBuffer(8+a+4*i),h=new DataView(r);h.setUint32(0,1229866072),h.setUint32(4,s.length);let c=8,d=8+a;for(let t=0;t<s.length;t++){const e=s[t],i=n[t];h.setUint32(c,e),h.setUint32(c+4,d),h.setUint32(c+8,i.length),c+=12;for(let t=0;t<i.length;t++)h.setUint32(d,i[t],!0),d+=4}await this.#o.write(this.#g,r),this.#l=r,this.#f=h}search(t){if(!this.#f||!this.#l)return[];const e=o.hash(t);let s=0,i=this.#f.getUint32(4)-1;for(;s<=i;){const t=s+i>>>1,n=8+12*t,a=this.#f.getUint32(n);if(a<e)s=t+1;else{if(!(a>e)){const t=this.#f.getUint32(n+4),e=this.#f.getUint32(n+8),s=[];for(let i=0;i<e;i++)s.push(this.#f.getUint32(t+4*i,!0));return s}i=t-1}}return[]}}const h="word_cache.bin",c="char_cache.bin";class d{#o;#h;#w;#u;#m=!1;#y;#p=!1;#D={word:0,char:0};constructor(s){if(!s.baseDir)throw new Error("SearchEngine requires 'baseDir' in config.");if(this.#y={wordSegmentTokenThreshold:1e5,charSegmentTokenThreshold:5e5,minWordTokenSave:0,minCharTokenSave:0,...s},(this.#y.minWordTokenSave||0)>=(this.#y.wordSegmentTokenThreshold||1e5))throw new Error("minWordTokenSave must be less than wordSegmentTokenThreshold");if((this.#y.minCharTokenSave||0)>=(this.#y.charSegmentTokenThreshold||5e5))throw new Error("minCharTokenSave must be less than charSegmentTokenThreshold");let i=null;if(this.#y.storage&&("object"==typeof this.#y.storage?i=this.#y.storage:"browser"===this.#y.storage?i=new t(this.#y.baseDir):"node"===this.#y.storage&&(i=new e(this.#y.baseDir))),!i){const s=typeof navigator<"u"&&navigator?.storage?.getDirectory instanceof Function,n=typeof process<"u"&&null!=process.versions&&null!=process.versions.node;s?i=new t(this.#y.baseDir):n&&(i=new e(this.#y.baseDir))}if(!i)throw new Error('Storage initialization failed. Please configure "storage" explicitly or ensure you are in a supported environment (Browser/Node).');this.#o=i,this.#h=new a(this.#o),this.#w=new r(this.#o),this.#u=new Map}async init(){if(this.#m)return;await this.#h.load();const t=[...this.#h.getSegments("word"),...this.#h.getSegments("char")];for(const e of t)this.#u.has(e.filename)||this.#u.set(e.filename,new o(e.filename,this.#o)),await this.#u.get(e.filename).loadIndex();this.#m=!0}startBatch(){this.#p=!0,this.#D={word:0,char:0}}async endBatch(){this.#p=!1,this.#D.word>0&&await this.#S("word",this.#D.word),this.#D.char>0&&await this.#S("char",this.#D.char),this.#D={word:0,char:0},await this.#h.save()}#k(t){try{if(typeof Intl<"u"&&"function"==typeof Intl.Segmenter&&"function"==typeof Array.from){const e=new Intl.Segmenter([],{granularity:"word"}).segment(t);if("object"==typeof e&&null!==e)return Array.from(e).filter(t=>t?.isWordLike).map(t=>t?.segment?.toLowerCase()||"")}}catch{}return t.toLowerCase().split(/[^a-z0-9\u4e00-\u9fa5]+/g).filter(t=>t.length>0)}#I(t){return this.#y.indexingTokenizer?this.#y.indexingTokenizer(t):this.#k(t.text)}#b(t){return this.#y.searchTokenizer?this.#y.searchTokenizer(t):this.#I(t)}async addDocument(t){return this.addDocuments([t])}async addDocumentIfMissing(t){return this.addDocumentsIfMissing([t])}async addDocumentsIfMissing(t){if(this.#m||await this.init(),0===t.length)return;const e=this.#h.getDeletedIds(),s=[],i=[],n=[];for(const a of t){if(e.has(a.id)||this.#h.isAdded(a.id))continue;const t=this.#I(a),r=[],o=[];for(const e of t)e.length>1?r.push(e):1===e.length&&o.push(e);r.length>0&&s.push({id:a.id,tokens:r}),o.length>0&&i.push({id:a.id,tokens:o}),n.push(a)}if(0===n.length)return;let a=0,r=0;if(s.length>0){await this.#w.appendBatch(h,s);for(const t of s)a+=t.tokens.length}if(i.length>0){await this.#w.appendBatch(c,i);for(const t of i)r+=t.tokens.length}for(const t of n)this.#h.addAddedId(t.id);this.#p?(this.#D.word+=a,this.#D.char+=r):(a>0&&await this.#S("word",a),r>0&&await this.#S("char",r),await this.#h.save())}async addDocuments(t){if(this.#m||await this.init(),0===t.length)return;const e=this.#h.getDeletedIds(),s=[],i=[];for(const n of t){if(e.has(n.id))throw new Error(`Document ID ${n.id} has been deleted and cannot be re-added.`);if(this.#h.isAdded(n.id))throw new Error(`Document ID ${n.id} already exists.`);const t=this.#I(n),a=[],r=[];for(const e of t)e.length>1?a.push(e):1===e.length&&r.push(e);a.length>0&&s.push({id:n.id,tokens:a}),r.length>0&&i.push({id:n.id,tokens:r})}let n=0,a=0;if(s.length>0){await this.#w.appendBatch(h,s);for(const t of s)n+=t.tokens.length}if(i.length>0){await this.#w.appendBatch(c,i);for(const t of i)a+=t.tokens.length}for(const e of t)this.#h.addAddedId(e.id);this.#p?(this.#D.word+=n,this.#D.char+=a):(n>0&&await this.#S("word",n),a>0&&await this.#S("char",a),await this.#h.save())}async#S(t,e){const s="word"===t?h:c,i=await this.#w.getCurrentSize(s),n="word"===t?this.#y.wordSegmentTokenThreshold||1e5:this.#y.charSegmentTokenThreshold||5e5,a="word"===t?this.#y.minWordTokenSave||0:this.#y.minCharTokenSave||0,r=this.#h.getLastSegmentInfo(t);let d,g,l,f;const w=()=>{const e=this.#h.getSegments(t).length+1;return`${t}_seg_${e}.bin`};if(r){const t=r.tokenCount;t>=n||t+e>=n?(d=w(),l=!0,g=r.end,f=e):(d=r.filename,l=!1,g=r.start,f=t+e)}else d=w(),l=!0,g=0,f=e;if(f<a)return void this.#h.updateSegment(t,d,g,i,f,l);const u=await this.#w.readRange(s,g,i);let m=this.#u.get(d);m||(m=new o(d,this.#o),this.#u.set(d,m)),await m.buildAndSave(u),this.#h.updateSegment(t,d,g,i,f,l)}async search(t,e){this.#m||await this.init();const s="string"==typeof t?{text:t}:t,i=this.#b(s),n=i.filter(t=>t.length>1),a=i.filter(t=>1===t.length),r=this.#h.getDeletedIds(),h=new Map,c=new Map,d=t=>{const e=this.#h.getSegments(t);for(const t of e){const e=t.filename;!this.#u.has(e)&&!c.has(e)&&c.set(e,new o(e,this.#o))}};d("word"),d("char"),await Promise.all(Array.from(c.entries()).map(([t,e])=>e.loadIndex().then(s=>{s&&this.#u.set(t,e)})));const g=async(t,e)=>{if(0===e.length)return;const s=this.#h.getSegments(t);for(const t of s){const s=t.filename,i=this.#u.get(s);if(i)for(const t of e){const e=i.search(t),s=1+.1*t.length;for(const i of e)if(!r.has(i))if(h.has(i)){const e=h.get(i);e.score+=s,e.tokens.add(t)}else h.set(i,{score:0,tokens:new Set([t])})}}};await g("word",n),await g("char",a);const l=[];return h.forEach((t,e)=>{l.push({id:e,score:t.score,tokens:Array.from(t.tokens)})}),l.sort((t,e)=>e.score-t.score),"number"==typeof e&&e>0?l.slice(0,e):l}async removeDocument(t){this.#m||await this.init(),this.#h.addDeletedId(t),this.#h.removeAddedId(t),await this.#h.save()}async clearAll(){await this.#o.clearAll(),this.#u.clear(),this.#h.reset(),this.#m=!1,this.#p=!1,this.#D={word:0,char:0}}async getStatus(){return this.#m||await this.init(),{wordSegments:this.#h.getSegments("word").length,charSegments:this.#h.getSegments("char").length,deleted:this.#h.getDeletedIds().size,wordCacheSize:await this.#w.getCurrentSize(h),charCacheSize:await this.#w.getCurrentSize(c),inBatch:this.#p}}async hasDocument(t){return this.#m||await this.init(),this.#h.hasDocument(t)}}class g{static#T=null;static#v={baseDir:"simple_search_data",wordSegmentTokenThreshold:1e5,minWordTokenSave:0};static configure(t){const e={...this.#v,...t};this.#T=new d(e)}static#z(){return this.#T||(this.#T=new d(this.#v)),this.#T}static async startBatch(){this.#z().startBatch()}static async endBatch(){return this.#z().endBatch()}static async addDocument(t){return this.#z().addDocument(t)}static async addDocumentIfMissing(t){return this.#z().addDocumentIfMissing(t)}static async addDocuments(t){return this.#z().addDocuments(t)}static async addDocumentsIfMissing(t){return this.#z().addDocumentsIfMissing(t)}static async search(t,e){return this.#z().search(t,e)}static async removeDocument(t){return this.#z().removeDocument(t)}static async clearAll(){return this.#z().clearAll()}static async getStatus(){return this.#z().getStatus()}static async hasDocument(t){return this.#z().hasDocument(t)}}export{t as BrowserStorage,e as NodeStorage,d as SearchEngine,g as SimpleSearch};
+export*from'./type';export*from'./core';export*from'./simple';export*from'./browser';export*from'./node';

package/lib/node.cjs

@@ -0,0 +1 @@
+"use strict";Object.create,Object.defineProperty,Object.getOwnPropertyDescriptor,Object.getOwnPropertyNames,Object.getPrototypeOf,Object.prototype.hasOwnProperty;exports.NodeStorage=class{#t=null;#i=null;#a;#s="";constructor(t){this.#a=t}async#e(){if(this.#t)return;const t=await import("fs/promises"),i=await import("path");this.#t=t,this.#i=i,this.#s=this.#i.join(process.cwd(),this.#a);try{await this.#t.access(this.#s)}catch{await this.#t.mkdir(this.#s,{recursive:!0})}}#r(t){return this.#i.join(this.#s,t)}async write(t,i){await this.#e(),await this.#t.writeFile(this.#r(t),Buffer.from(i))}async append(t,i){await this.#e(),await this.#t.appendFile(this.#r(t),Buffer.from(i))}async read(t){await this.#e();try{const i=await this.#t.readFile(this.#r(t));return i.buffer.slice(i.byteOffset,i.byteOffset+i.byteLength)}catch{return null}}async readRange(t,i,a){await this.#e();try{const s=await this.#t.open(this.#r(t),"r"),e=a-i,r=Buffer.alloc(e);return await s.read(r,0,e,i),await s.close(),r.buffer.slice(r.byteOffset,r.byteOffset+r.byteLength)}catch{return null}}async remove(t){await this.#e();try{await this.#t.unlink(this.#r(t))}catch{}}async listFiles(){await this.#e();try{return await this.#t.readdir(this.#s)}catch{return[]}}async clearAll(){await this.#e();try{const t=await this.#t.readdir(this.#s);for(const i of t)await this.#t.unlink(this.#i.join(this.#s,i))}catch{}}async getFileSize(t){await this.#e();try{return(await this.#t.stat(this.#r(t))).size}catch{return 0}}};

package/lib/node.d.ts

@@ -0,0 +1,20 @@
+import { IStorage } from './type';
+
+/**
+ * Node.js 实现
+ */
+
+declare class NodeStorage implements IStorage {
+    #private;
+    constructor(baseDir: string);
+    write(filename: string, data: ArrayBuffer): Promise<void>;
+    append(filename: string, data: ArrayBuffer): Promise<void>;
+    read(filename: string): Promise<ArrayBuffer | null>;
+    readRange(filename: string, start: number, end: number): Promise<ArrayBuffer | null>;
+    remove(filename: string): Promise<void>;
+    listFiles(): Promise<string[]>;
+    clearAll(): Promise<void>;
+    getFileSize(filename: string): Promise<number>;
+}
+
+export { NodeStorage };

package/lib/node.js

@@ -0,0 +1 @@
+class t{#t=null;#i=null;#a;#s="";constructor(t){this.#a=t}async#r(){if(this.#t)return;const t=await import("fs/promises"),i=await import("path");this.#t=t,this.#i=i,this.#s=this.#i.join(process.cwd(),this.#a);try{await this.#t.access(this.#s)}catch{await this.#t.mkdir(this.#s,{recursive:!0})}}#e(t){return this.#i.join(this.#s,t)}async write(t,i){await this.#r(),await this.#t.writeFile(this.#e(t),Buffer.from(i))}async append(t,i){await this.#r(),await this.#t.appendFile(this.#e(t),Buffer.from(i))}async read(t){await this.#r();try{const i=await this.#t.readFile(this.#e(t));return i.buffer.slice(i.byteOffset,i.byteOffset+i.byteLength)}catch{return null}}async readRange(t,i,a){await this.#r();try{const s=await this.#t.open(this.#e(t),"r"),r=a-i,e=Buffer.alloc(r);return await s.read(e,0,r,i),await s.close(),e.buffer.slice(e.byteOffset,e.byteOffset+e.byteLength)}catch{return null}}async remove(t){await this.#r();try{await this.#t.unlink(this.#e(t))}catch{}}async listFiles(){await this.#r();try{return await this.#t.readdir(this.#s)}catch{return[]}}async clearAll(){await this.#r();try{const t=await this.#t.readdir(this.#s);for(const i of t)await this.#t.unlink(this.#i.join(this.#s,i))}catch{}}async getFileSize(t){await this.#r();try{return(await this.#t.stat(this.#e(t))).size}catch{return 0}}}export{t as NodeStorage};

package/lib/simple.cjs

@@ -0,0 +1 @@
+"use strict";var t=require('./core');exports.SimpleSearch=class{static#t=null;static#e={baseDir:"simple_search_data",wordSegmentTokenThreshold:1e5,minWordTokenSave:0};static configure(e){const n={...this.#e,...e};this.#t=new t.SearchEngine(n)}static#n(){return this.#t||(this.#t=new t.SearchEngine(this.#e)),this.#t}static async startBatch(){this.#n().startBatch()}static async endBatch(){return this.#n().endBatch()}static async addDocument(t){return this.#n().addDocument(t)}static async addDocumentIfMissing(t){return this.#n().addDocumentIfMissing(t)}static async addDocuments(t){return this.#n().addDocuments(t)}static async addDocumentsIfMissing(t){return this.#n().addDocumentsIfMissing(t)}static async search(t,e){return this.#n().search(t,e)}static async removeDocument(t){return this.#n().removeDocument(t)}static async clearAll(){return this.#n().clearAll()}static async getStatus(){return this.#n().getStatus()}static async hasDocument(t){return this.#n().hasDocument(t)}};

package/lib/simple.d.ts

@@ -0,0 +1,39 @@
+import * as ___type from './type';
+import { ISearchEngineConfig, IDocument, IDocumentBase } from './type';
+
+/**
+ * 快速使用封装
+ * 提供单例模式和默认配置
+ */
+declare class SimpleSearch {
+    #private;
+    /**
+     * 配置并初始化单例
+     */
+    static configure(config: Partial<ISearchEngineConfig>): void;
+    static startBatch(): Promise<void>;
+    static endBatch(): Promise<void>;
+    static addDocument<T extends IDocument = IDocument>(doc: T): Promise<void>;
+    static addDocumentIfMissing<T extends IDocument = IDocument>(doc: T): Promise<void>;
+    static addDocuments<T extends IDocument = IDocument>(docs: T[]): Promise<void>;
+    static addDocumentsIfMissing<T extends IDocument = IDocument>(docs: T[]): Promise<void>;
+    static search<T extends IDocumentBase = IDocumentBase>(query: T | string, limit?: number): Promise<___type.IResult[]>;
+    static removeDocument(id: number): Promise<void>;
+    static clearAll(): Promise<void>;
+    static getStatus(): Promise<{
+        wordSegments: number;
+        charSegments: number;
+        deleted: number;
+        wordCacheSize: number;
+        charCacheSize: number;
+        inBatch: boolean;
+    }>;
+    /**
+     * 检查文档ID是否曾经添加过（包括已删除的）
+     * @param id 文档ID
+     * @returns 文档是否曾经添加过的布尔值
+     */
+    static hasDocument(id: number): Promise<boolean>;
+}
+
+export { SimpleSearch };

package/lib/simple.js

@@ -0,0 +1 @@
+import{SearchEngine as t}from'./core';class s{static#t=null;static#s={baseDir:"simple_search_data",wordSegmentTokenThreshold:1e5,minWordTokenSave:0};static configure(s){const n={...this.#s,...s};this.#t=new t(n)}static#n(){return this.#t||(this.#t=new t(this.#s)),this.#t}static async startBatch(){this.#n().startBatch()}static async endBatch(){return this.#n().endBatch()}static async addDocument(t){return this.#n().addDocument(t)}static async addDocumentIfMissing(t){return this.#n().addDocumentIfMissing(t)}static async addDocuments(t){return this.#n().addDocuments(t)}static async addDocumentsIfMissing(t){return this.#n().addDocumentsIfMissing(t)}static async search(t,s){return this.#n().search(t,s)}static async removeDocument(t){return this.#n().removeDocument(t)}static async clearAll(){return this.#n().clearAll()}static async getStatus(){return this.#n().getStatus()}static async hasDocument(t){return this.#n().hasDocument(t)}}export{s as SimpleSearch};

package/lib/type.cjs

@@ -0,0 +1 @@
+"use strict";

package/lib/type.d.ts

@@ -0,0 +1,135 @@
+/**
+ * 文档基础接口
+ */
+interface IDocumentBase {
+    text: string;
+}
+/**
+ * 文档接口（包含ID）
+ */
+interface IDocument extends IDocumentBase {
+    id: number;
+}
+
+/**
+ * 搜索结果接口
+ */
+interface IResult {
+    id: number;
+    score: number;
+    tokens: string[];
+}
+
+/**
+ * 索引段元数据接口
+ */
+interface ISegmentMeta {
+    filename: string;
+    start: number;
+    end: number;
+    tokenCount: number;
+}
+
+/**
+ * 索引元数据接口
+ */
+
+interface IIndexMeta {
+    wordSegments: ISegmentMeta[];
+    charSegments: ISegmentMeta[];
+}
+
+/**
+ * 分词后的文档接口
+ */
+interface ITokenizedDoc {
+    id: number;
+    tokens: string[];
+}
+
+/**
+ * 存储层接口 (外部化)
+ */
+interface IStorage {
+    write(filename: string, data: ArrayBuffer): Promise<void>;
+    append(filename: string, data: ArrayBuffer): Promise<void>;
+    read(filename: string): Promise<ArrayBuffer | null>;
+    readRange(filename: string, start: number, end: number): Promise<ArrayBuffer | null>;
+    remove(filename: string): Promise<void>;
+    listFiles(): Promise<string[]>;
+    clearAll(): Promise<void>;
+    getFileSize(filename: string): Promise<number>;
+}
+
+/**
+ * 索引类型
+ */
+type IndexType = 'word' | 'char';
+
+interface ISearchEngineConfig {
+    /**
+     * 数据存储的基础目录 (必填)
+     * 用于区分不同的搜索引擎实例
+     */
+    baseDir: string;
+    /**
+     * 存储实现配置 (可选)
+     * - 'browser': 强制使用 OPFS (BrowserStorage)
+     * - 'node': 强制使用 Node.js fs (NodeStorage)
+     * - IStorage: 传入自定义的存储实例
+     * - undefined: 自动检测环境
+     */
+    storage?: 'browser' | 'node' | IStorage;
+    /**
+     * 索引时使用的分词器 (算法核心配置)
+     * - 作用: 将文档文本转换为索引用的token序列
+     * - 算法: 自定义实现的分词逻辑，需满足返回字符串数组的格式要求
+     * - 建议: 针对不同语言(中文/英文/日文等)使用专门的分词实现
+     * - 影响: 直接决定索引的粒度和搜索的准确性
+     */
+    indexingTokenizer?: <T extends IDocument = IDocument>(doc: T) => string[];
+    /**
+     * 搜索时使用的分词器 (算法核心配置)
+     * - 作用: 将查询文本转换为搜索用的token序列
+     * - 算法: 自定义实现的分词逻辑，需满足返回字符串数组的格式要求
+     * - 建议: 与indexingTokenizer保持一致的分词策略以确保搜索准确性
+     * - 影响: 直接决定搜索匹配的范围和结果的相关性
+     */
+    searchTokenizer?: <T extends IDocumentBase = IDocumentBase>(doc: T) => string[];
+    /**
+     * 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
+     * - 算法: 基于Token数量的分段策略，当新增Token数加上已有Token数超过阈值时触发分段
+     * - 默认值: 100000
+     * - 影响: 过小会导致索引文件过多，过大可能影响搜索性能
+     */
+    wordSegmentTokenThreshold?: number;
+    /**
+     * 字索引分段阈值 (Token数) - 分段算法配置
+     * - 作用: 控制字索引文件的大小，超过阈值时创建新的索引段
+     * - 算法: 基于Token数量的分段策略，当新增Token数加上已有Token数超过阈值时触发分段
+     * - 默认值: 500000
+     * - 影响: 过小会导致索引文件过多，过大可能影响搜索性能
+     */
+    charSegmentTokenThreshold?: number;
+    /**
+     * 词索引最小保存阈值 (Token数) - 缓存算法配置
+     * - 作用: 控制词索引是否立即写入磁盘，低于阈值时只保存在内存缓存中
+     * - 算法: 基于Token数量的缓存策略，当累计Token数达到阈值时才进行持久化
+     * - 默认值: 0
+     * - 影响: 适当设置可减少磁盘IO次数，提高索引性能
+     */
+    minWordTokenSave?: number;
+    /**
+     * 字索引最小保存阈值 (Token数) - 缓存算法配置
+     * - 作用: 控制字索引是否立即写入磁盘，低于阈值时只保存在内存缓存中
+     * - 算法: 基于Token数量的缓存策略，当累计Token数达到阈值时才进行持久化
+     * - 默认值: 0
+     * - 影响: 适当设置可减少磁盘IO次数，提高索引性能
+     */
+    minCharTokenSave?: number;
+}
+
+
+
+export type { IDocument, IDocumentBase, IIndexMeta, IResult, ISearchEngineConfig, ISegmentMeta, IStorage, ITokenizedDoc, IndexType };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "gs-search",
-  "version": "0.1.2",
+  "version": "0.1.4",
   "type": "module",
   "main": "lib/index.cjs",
   "module": "lib/index.js",
@@ -9,6 +9,31 @@
       "import": "./lib/index.js",
       "require": "./lib/index.cjs",
       "types": "./lib/index.d.ts"
+    },
+    "./browser": {
+      "import": "./lib/browser.js",
+      "require": "./lib/browser.cjs",
+      "types": "./lib/browser.d.ts"
+    },
+    "./core": {
+      "import": "./lib/core.js",
+      "require": "./lib/core.cjs",
+      "types": "./lib/core.d.ts"
+    },
+    "./node": {
+      "import": "./lib/node.js",
+      "require": "./lib/node.cjs",
+      "types": "./lib/node.d.ts"
+    },
+    "./simple": {
+      "import": "./lib/simple.js",
+      "require": "./lib/simple.cjs",
+      "types": "./lib/simple.d.ts"
+    },
+    "./type": {
+      "import": "./lib/type.js",
+      "require": "./lib/type.cjs",
+      "types": "./lib/type.d.ts"
     }
   },
   "types": "lib/index.d.ts",