@ninebix/nmt-system 1.0.0

package/dist/core/hierarchical-chunker.d.ts

@@ -0,0 +1,108 @@
+/**
+ * Hierarchical Chunker - Multi-level text segmentation
+ * 다중 레벨 텍스트 청킹 (문장 → 구 → 어절 → 음절)
+ *
+ * @module core/hierarchical-chunker
+ */
+import { Language } from './language-analyzers.js';
+/**
+ * Chunk level types
+ */
+export type ChunkLevel = 'document' | 'paragraph' | 'sentence' | 'phrase' | 'word' | 'morpheme' | 'syllable';
+/**
+ * Hierarchical chunk node
+ */
+export interface HierarchicalChunk {
+    id: string;
+    level: ChunkLevel;
+    text: string;
+    hash: string;
+    parentId: string | null;
+    childIds: string[];
+    position: {
+        start: number;
+        end: number;
+    };
+    metadata: {
+        createdAt: string;
+        source?: string;
+    };
+}
+/**
+ * Chunk tree
+ */
+export interface ChunkTree {
+    root: HierarchicalChunk;
+    nodes: Map<string, HierarchicalChunk>;
+    levels: Map<ChunkLevel, HierarchicalChunk[]>;
+}
+/**
+ * Hierarchical Chunker
+ * 텍스트를 다중 레벨로 분해 (다국어 지원)
+ */
+export declare class HierarchicalChunker {
+    private analyzer;
+    private detectedLanguage;
+    /**
+     * 텍스트를 계층적으로 청킹 (언어 자동 감지)
+     */
+    chunk(text: string, source?: string, language?: Language): ChunkTree;
+    /**
+     * 대화 입력을 계층적으로 처리
+     */
+    chunkConversation(input: string, speaker?: string): ChunkTree;
+    /**
+     * 질문에서 핵심 청크 추출
+     */
+    extractQueryChunks(query: string): {
+        intent: string;
+        keywords: string[];
+        chunks: HierarchicalChunk[];
+    };
+    private createChunk;
+    private splitParagraphs;
+    private splitSentences;
+    private splitPhrases;
+    private splitWords;
+    private splitMorphemes;
+    private splitSyllables;
+    /**
+     * 감지된 언어 반환
+     */
+    getDetectedLanguage(): Language;
+}
+/**
+ * 대화 학습 매니저
+ */
+export declare class ConversationLearner {
+    private chunker;
+    private memory;
+    constructor();
+    /**
+     * 대화 입력 학습
+     */
+    learn(input: string, speaker?: string): ChunkTree;
+    /**
+     * 질문에 대한 답변 청크 검색
+     */
+    query(question: string): {
+        intent: string;
+        matches: Array<{
+            text: string;
+            level: ChunkLevel;
+            relevance: number;
+            context: string;
+        }>;
+    };
+    /**
+     * 통계
+     */
+    getStats(): {
+        totalConversations: number;
+        totalChunks: number;
+        chunksByLevel: Record<ChunkLevel, number>;
+    };
+}
+export declare function createHierarchicalChunker(): HierarchicalChunker;
+export declare function createConversationLearner(): ConversationLearner;
+//# sourceMappingURL=hierarchical-chunker.d.ts.map

package/dist/core/hierarchical-chunker.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"hierarchical-chunker.d.ts","sourceRoot":"","sources":["../../src/core/hierarchical-chunker.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAIH,OAAO,EACL,QAAQ,EAMT,MAAM,yBAAyB,CAAC;AAEjC;;GAEG;AACH,MAAM,MAAM,UAAU,GAClB,UAAU,GACV,WAAW,GACX,UAAU,GACV,QAAQ,GACR,MAAM,GACN,UAAU,GACV,UAAU,CAAC;AAEf;;GAEG;AACH,MAAM,WAAW,iBAAiB;IAChC,EAAE,EAAE,MAAM,CAAC;IACX,KAAK,EAAE,UAAU,CAAC;IAClB,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC;IACxB,QAAQ,EAAE,MAAM,EAAE,CAAC;IACnB,QAAQ,EAAE;QACR,KAAK,EAAE,MAAM,CAAC;QACd,GAAG,EAAE,MAAM,CAAC;KACb,CAAC;IACF,QAAQ,EAAE;QACR,SAAS,EAAE,MAAM,CAAC;QAClB,MAAM,CAAC,EAAE,MAAM,CAAC;KACjB,CAAC;CACH;AAED;;GAEG;AACH,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,iBAAiB,CAAC;IACxB,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,iBAAiB,CAAC,CAAC;IACtC,MAAM,EAAE,GAAG,CAAC,UAAU,EAAE,iBAAiB,EAAE,CAAC,CAAC;CAC9C;AAED;;;GAGG;AACH,qBAAa,mBAAmB;IAC9B,OAAO,CAAC,QAAQ,CAAiC;IACjD,OAAO,CAAC,gBAAgB,CAAoB;IAE5C;;OAEG;IACH,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,MAAM,EAAE,QAAQ,GAAE,QAAiB,GAAG,SAAS;IAoF5E;;OAEG;IACH,iBAAiB,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,MAAM,GAAG,SAAS;IAK7D;;OAEG;IACH,kBAAkB,CAAC,KAAK,EAAE,MAAM,GAAG;QACjC,MAAM,EAAE,MAAM,CAAC;QACf,QAAQ,EAAE,MAAM,EAAE,CAAC;QACnB,MAAM,EAAE,iBAAiB,EAAE,CAAC;KAC7B;IAqCD,OAAO,CAAC,WAAW;IAuBnB,OAAO,CAAC,eAAe;IAIvB,OAAO,CAAC,cAAc;IAQtB,OAAO,CAAC,YAAY;IAQpB,OAAO,CAAC,UAAU;IAQlB,OAAO,CAAC,cAAc;IAQtB,OAAO,CAAC,cAAc;IAKtB;;OAEG;IACH,mBAAmB,IAAI,QAAQ;CAGhC;AAED;;GAEG;AACH,qBAAa,mBAAmB;IAC9B,OAAO,CAAC,OAAO,CAAsB;IACrC,OAAO,CAAC,MAAM,CAIX;;IAOH;;OAEG;IACH,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,GAAE,MAAe,GAAG,SAAS;IAazD;;OAEG;IACH,KAAK,CAAC,QAAQ,EAAE,MAAM,GAAG;QACvB,MAAM,EAAE,MAAM,CAAC;QACf,OAAO,EAAE,KAAK,CAAC;YACb,IAAI,EAAE,MAAM,CAAC;YACb,KAAK,EAAE,UAAU,CAAC;YAClB,SAAS,EAAE,MAAM,CAAC;YAClB,OAAO,EAAE,MAAM,CAAC;SACjB,CAAC,CAAC;KACJ;IAuDD;;OAEG;IACH,QAAQ,IAAI;QACV,kBAAkB,EAAE,MAAM,CAAC;QAC3B,WAAW,EAAE,MAAM,CAAC;QACpB,aAAa,EAAE,MAAM,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;KAC3C;CA0BF;AAGD,wBAAgB,yBAAyB,IAAI,mBAAmB,CAE/D;AAED,wBAAgB,yBAAyB,IAAI,mBAAmB,CAE/D"}

package/dist/core/hierarchical-chunker.js

@@ -0,0 +1,296 @@
+/**
+ * Hierarchical Chunker - Multi-level text segmentation
+ * 다중 레벨 텍스트 청킹 (문장 → 구 → 어절 → 음절)
+ *
+ * @module core/hierarchical-chunker
+ */
+import { generateUUID } from '../utils/uuid.js';
+import { hash } from '../utils/hash.js';
+import { detectLanguage, getAnalyzer } from './language-analyzers.js';
+/**
+ * Hierarchical Chunker
+ * 텍스트를 다중 레벨로 분해 (다국어 지원)
+ */
+export class HierarchicalChunker {
+    analyzer = null;
+    detectedLanguage = 'auto';
+    /**
+     * 텍스트를 계층적으로 청킹 (언어 자동 감지)
+     */
+    chunk(text, source, language = 'auto') {
+        // 언어 감지 및 분석기 설정
+        this.detectedLanguage = language === 'auto' ? detectLanguage(text) : language;
+        this.analyzer = getAnalyzer(this.detectedLanguage);
+        const nodes = new Map();
+        const levels = new Map();
+        // Initialize level arrays
+        const allLevels = ['document', 'paragraph', 'sentence', 'phrase', 'word', 'syllable'];
+        for (const level of allLevels) {
+            levels.set(level, []);
+        }
+        // Level 1: Document (전체 문서)
+        const docChunk = this.createChunk('document', text, 0, text.length, null, source);
+        nodes.set(docChunk.id, docChunk);
+        levels.get('document').push(docChunk);
+        // Level 2: Paragraphs (문단)
+        const paragraphs = this.splitParagraphs(text);
+        let offset = 0;
+        for (const para of paragraphs) {
+            const paraChunk = this.createChunk('paragraph', para, offset, offset + para.length, docChunk.id, source);
+            nodes.set(paraChunk.id, paraChunk);
+            levels.get('paragraph').push(paraChunk);
+            docChunk.childIds.push(paraChunk.id);
+            // Level 3: Sentences (문장)
+            const sentences = this.splitSentences(para);
+            let sentOffset = offset;
+            for (const sent of sentences) {
+                const sentChunk = this.createChunk('sentence', sent, sentOffset, sentOffset + sent.length, paraChunk.id, source);
+                nodes.set(sentChunk.id, sentChunk);
+                levels.get('sentence').push(sentChunk);
+                paraChunk.childIds.push(sentChunk.id);
+                // Level 4: Phrases (구/절)
+                const phrases = this.splitPhrases(sent);
+                let phraseOffset = sentOffset;
+                for (const phrase of phrases) {
+                    const phraseChunk = this.createChunk('phrase', phrase, phraseOffset, phraseOffset + phrase.length, sentChunk.id, source);
+                    nodes.set(phraseChunk.id, phraseChunk);
+                    levels.get('phrase').push(phraseChunk);
+                    sentChunk.childIds.push(phraseChunk.id);
+                    // Level 5: Words (어절/단어)
+                    const words = this.splitWords(phrase);
+                    let wordOffset = phraseOffset;
+                    for (const word of words) {
+                        const wordChunk = this.createChunk('word', word, wordOffset, wordOffset + word.length, phraseChunk.id, source);
+                        nodes.set(wordChunk.id, wordChunk);
+                        levels.get('word').push(wordChunk);
+                        phraseChunk.childIds.push(wordChunk.id);
+                        // Level 6: Syllables (음절)
+                        const syllables = this.splitSyllables(word);
+                        let syllOffset = wordOffset;
+                        for (const syll of syllables) {
+                            const syllChunk = this.createChunk('syllable', syll, syllOffset, syllOffset + syll.length, wordChunk.id, source);
+                            nodes.set(syllChunk.id, syllChunk);
+                            levels.get('syllable').push(syllChunk);
+                            wordChunk.childIds.push(syllChunk.id);
+                            syllOffset += syll.length;
+                        }
+                        wordOffset += word.length + 1; // +1 for space
+                    }
+                    phraseOffset += phrase.length;
+                }
+                sentOffset += sent.length;
+            }
+            offset += para.length + 1; // +1 for newline
+        }
+        return {
+            root: docChunk,
+            nodes,
+            levels
+        };
+    }
+    /**
+     * 대화 입력을 계층적으로 처리
+     */
+    chunkConversation(input, speaker) {
+        const tree = this.chunk(input, `conversation:${speaker || 'user'}`);
+        return tree;
+    }
+    /**
+     * 질문에서 핵심 청크 추출
+     */
+    extractQueryChunks(query) {
+        const tree = this.chunk(query);
+        const words = tree.levels.get('word') || [];
+        // 의도 파악 (간단한 규칙 기반)
+        let intent = 'unknown';
+        const queryLower = query.toLowerCase();
+        if (queryLower.includes('누구') || queryLower.includes('who')) {
+            intent = 'identity';
+        }
+        else if (queryLower.includes('뭐') || queryLower.includes('무엇') || queryLower.includes('what')) {
+            intent = 'definition';
+        }
+        else if (queryLower.includes('어디') || queryLower.includes('where')) {
+            intent = 'location';
+        }
+        else if (queryLower.includes('언제') || queryLower.includes('when')) {
+            intent = 'time';
+        }
+        else if (queryLower.includes('왜') || queryLower.includes('why')) {
+            intent = 'reason';
+        }
+        else if (queryLower.includes('어떻게') || queryLower.includes('how')) {
+            intent = 'method';
+        }
+        // 핵심 키워드 추출 (조사, 어미 제외)
+        const stopWords = new Set(['은', '는', '이', '가', '을', '를', '에', '의', '로', '와', '과', '도', '만', '?', '.', '!']);
+        const keywords = words
+            .map(w => w.text.trim())
+            .filter(w => w.length > 0 && !stopWords.has(w));
+        return {
+            intent,
+            keywords,
+            chunks: [...(tree.levels.get('word') || []), ...(tree.levels.get('phrase') || [])]
+        };
+    }
+    // ==================== Private Methods ====================
+    createChunk(level, text, start, end, parentId, source) {
+        return {
+            id: generateUUID(),
+            level,
+            text: text.trim(),
+            hash: hash(text),
+            parentId,
+            childIds: [],
+            position: { start, end },
+            metadata: {
+                createdAt: new Date().toISOString(),
+                source
+            }
+        };
+    }
+    splitParagraphs(text) {
+        return text.split(/\n\s*\n/).filter(p => p.trim().length > 0);
+    }
+    splitSentences(text) {
+        // 언어별 문장 분리
+        if (this.analyzer) {
+            return this.analyzer.splitSentences(text);
+        }
+        return text.split(/(?<=[.!?。！？])\s+/).filter(s => s.trim().length > 0);
+    }
+    splitPhrases(text) {
+        // 언어별 구/절 분리
+        if (this.analyzer) {
+            return this.analyzer.splitPhrases(text);
+        }
+        return text.split(/[,，、;；]/g).filter(p => p.trim().length > 0);
+    }
+    splitWords(text) {
+        // 언어별 어절/단어 분리
+        if (this.analyzer) {
+            return this.analyzer.splitWords(text);
+        }
+        return text.split(/\s+/).filter(w => w.trim().length > 0);
+    }
+    splitMorphemes(text) {
+        // 언어별 형태소 분리
+        if (this.analyzer) {
+            return this.analyzer.splitMorphemes(text);
+        }
+        return text.split(/\s+/).filter(w => w.trim().length > 0);
+    }
+    splitSyllables(text) {
+        // 글자 단위 분리
+        return [...text].filter(c => c.trim().length > 0);
+    }
+    /**
+     * 감지된 언어 반환
+     */
+    getDetectedLanguage() {
+        return this.detectedLanguage;
+    }
+}
+/**
+ * 대화 학습 매니저
+ */
+export class ConversationLearner {
+    chunker;
+    memory;
+    constructor() {
+        this.chunker = new HierarchicalChunker();
+        this.memory = new Map();
+    }
+    /**
+     * 대화 입력 학습
+     */
+    learn(input, speaker = 'user') {
+        const tree = this.chunker.chunkConversation(input, speaker);
+        // 메모리에 저장
+        this.memory.set(tree.root.id, {
+            tree,
+            timestamp: Date.now(),
+            speaker
+        });
+        return tree;
+    }
+    /**
+     * 질문에 대한 답변 청크 검색
+     */
+    query(question) {
+        const { intent, keywords } = this.chunker.extractQueryChunks(question);
+        const matches = [];
+        // 모든 저장된 대화에서 검색
+        for (const [, entry] of this.memory) {
+            const { tree, speaker } = entry;
+            // 각 레벨에서 키워드 매칭
+            for (const [level, chunks] of tree.levels) {
+                for (const chunk of chunks) {
+                    let relevance = 0;
+                    // 키워드 매칭 점수
+                    for (const keyword of keywords) {
+                        if (chunk.text.includes(keyword)) {
+                            relevance += 1;
+                        }
+                    }
+                    // 의도 기반 매칭
+                    if (intent === 'identity' && (chunk.text.includes('저는') || chunk.text.includes('나는'))) {
+                        relevance += 2;
+                    }
+                    if (relevance > 0) {
+                        // 컨텍스트 찾기 (부모 청크)
+                        let context = chunk.text;
+                        if (chunk.parentId) {
+                            const parent = tree.nodes.get(chunk.parentId);
+                            if (parent)
+                                context = parent.text;
+                        }
+                        matches.push({
+                            text: chunk.text,
+                            level: level,
+                            relevance,
+                            context
+                        });
+                    }
+                }
+            }
+        }
+        // 관련도 순으로 정렬
+        matches.sort((a, b) => b.relevance - a.relevance);
+        return { intent, matches };
+    }
+    /**
+     * 통계
+     */
+    getStats() {
+        const chunksByLevel = {
+            document: 0,
+            paragraph: 0,
+            sentence: 0,
+            phrase: 0,
+            word: 0,
+            morpheme: 0,
+            syllable: 0
+        };
+        let totalChunks = 0;
+        for (const [, entry] of this.memory) {
+            for (const [level, chunks] of entry.tree.levels) {
+                chunksByLevel[level] += chunks.length;
+                totalChunks += chunks.length;
+            }
+        }
+        return {
+            totalConversations: this.memory.size,
+            totalChunks,
+            chunksByLevel
+        };
+    }
+}
+// Export
+export function createHierarchicalChunker() {
+    return new HierarchicalChunker();
+}
+export function createConversationLearner() {
+    return new ConversationLearner();
+}
+//# sourceMappingURL=hierarchical-chunker.js.map

package/dist/core/hierarchical-chunker.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"hierarchical-chunker.js","sourceRoot":"","sources":["../../src/core/hierarchical-chunker.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAChD,OAAO,EAAE,IAAI,EAAE,MAAM,kBAAkB,CAAC;AACxC,OAAO,EAIL,cAAc,EACd,WAAW,EAEZ,MAAM,yBAAyB,CAAC;AA2CjC;;;GAGG;AACH,MAAM,OAAO,mBAAmB;IACtB,QAAQ,GAA4B,IAAI,CAAC;IACzC,gBAAgB,GAAa,MAAM,CAAC;IAE5C;;OAEG;IACH,KAAK,CAAC,IAAY,EAAE,MAAe,EAAE,WAAqB,MAAM;QAC9D,iBAAiB;QACjB,IAAI,CAAC,gBAAgB,GAAG,QAAQ,KAAK,MAAM,CAAC,CAAC,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;QAC9E,IAAI,CAAC,QAAQ,GAAG,WAAW,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;QACnD,MAAM,KAAK,GAAG,IAAI,GAAG,EAA6B,CAAC;QACnD,MAAM,MAAM,GAAG,IAAI,GAAG,EAAmC,CAAC;QAE1D,0BAA0B;QAC1B,MAAM,SAAS,GAAiB,CAAC,UAAU,EAAE,WAAW,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,EAAE,UAAU,CAAC,CAAC;QACpG,KAAK,MAAM,KAAK,IAAI,SAAS,EAAE,CAAC;YAC9B,MAAM,CAAC,GAAG,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QACxB,CAAC;QAED,4BAA4B;QAC5B,MAAM,QAAQ,GAAG,IAAI,CAAC,WAAW,CAAC,UAAU,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,CAAC,MAAM,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC;QAClF,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAC;QACjC,MAAM,CAAC,GAAG,CAAC,UAAU,CAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAEvC,2BAA2B;QAC3B,MAAM,UAAU,GAAG,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,CAAC;QAC9C,IAAI,MAAM,GAAG,CAAC,CAAC;QACf,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,CAAC,WAAW,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC,MAAM,EAAE,QAAQ,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC;YACzG,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,EAAE,SAAS,CAAC,CAAC;YACnC,MAAM,CAAC,GAAG,CAAC,WAAW,CAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YACzC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC;YAErC,0BAA0B;YAC1B,MAAM,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;YAC5C,IAAI,UAAU,GAAG,MAAM,CAAC;YACxB,KAAK,MAAM,IAAI,IAAI,SAAS,EAAE,CAAC;gBAC7B,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,CAAC,UAAU,EAAE,IAAI,EAAE,UAAU,EAAE,UAAU,GAAG,IAAI,CAAC,MAAM,EAAE,SAAS,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC;gBACjH,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,EAAE,SAAS,CAAC,CAAC;gBACnC,MAAM,CAAC,GAAG,CAAC,UAAU,CAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;gBACxC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC;gBAEtC,yBAAyB;gBACzB,MAAM,OAAO,GAAG,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;gBACxC,IAAI,YAAY,GAAG,UAAU,CAAC;gBAC9B,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;oBAC7B,MAAM,WAAW,GAAG,IAAI,CAAC,WAAW,CAAC,QAAQ,EAAE,MAAM,EAAE,YAAY,EAAE,YAAY,GAAG,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC;oBACzH,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC,EAAE,EAAE,WAAW,CAAC,CAAC;oBACvC,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAE,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;oBACxC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC;oBAExC,yBAAyB;oBACzB,MAAM,KAAK,GAAG,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC;oBACtC,IAAI,UAAU,GAAG,YAAY,CAAC;oBAC9B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;wBACzB,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,CAAC,MAAM,EAAE,IAAI,EAAE,UAAU,EAAE,UAAU,GAAG,IAAI,CAAC,MAAM,EAAE,WAAW,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC;wBAC/G,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,EAAE,SAAS,CAAC,CAAC;wBACnC,MAAM,CAAC,GAAG,CAAC,MAAM,CAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;wBACpC,WAAW,CAAC,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC;wBAExC,0BAA0B;wBAC1B,MAAM,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;wBAC5C,IAAI,UAAU,GAAG,UAAU,CAAC;wBAC5B,KAAK,MAAM,IAAI,IAAI,SAAS,EAAE,CAAC;4BAC7B,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,CAAC,UAAU,EAAE,IAAI,EAAE,UAAU,EAAE,UAAU,GAAG,IAAI,CAAC,MAAM,EAAE,SAAS,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC;4BACjH,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,EAAE,SAAS,CAAC,CAAC;4BACnC,MAAM,CAAC,GAAG,CAAC,UAAU,CAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;4BACxC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC;4BACtC,UAAU,IAAI,IAAI,CAAC,MAAM,CAAC;wBAC5B,CAAC;wBAED,UAAU,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,eAAe;oBAChD,CAAC;oBAED,YAAY,IAAI,MAAM,CAAC,MAAM,CAAC;gBAChC,CAAC;gBAED,UAAU,IAAI,IAAI,CAAC,MAAM,CAAC;YAC5B,CAAC;YAED,MAAM,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,iBAAiB;QAC9C,CAAC;QAED,OAAO;YACL,IAAI,EAAE,QAAQ;YACd,KAAK;YACL,MAAM;SACP,CAAC;IACJ,CAAC;IAED;;OAEG;IACH,iBAAiB,CAAC,KAAa,EAAE,OAAgB;QAC/C,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,gBAAgB,OAAO,IAAI,MAAM,EAAE,CAAC,CAAC;QACpE,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;OAEG;IACH,kBAAkB,CAAC,KAAa;QAK9B,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QAC/B,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;QAE5C,oBAAoB;QACpB,IAAI,MAAM,GAAG,SAAS,CAAC;QACvB,MAAM,UAAU,GAAG,KAAK,CAAC,WAAW,EAAE,CAAC;QAEvC,IAAI,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;YAC5D,MAAM,GAAG,UAAU,CAAC;QACtB,CAAC;aAAM,IAAI,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;YAChG,MAAM,GAAG,YAAY,CAAC;QACxB,CAAC;aAAM,IAAI,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,UAAU,CAAC,QAAQ,CAAC,OAAO,CAAC,EAAE,CAAC;YACrE,MAAM,GAAG,UAAU,CAAC;QACtB,CAAC;aAAM,IAAI,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;YACpE,MAAM,GAAG,MAAM,CAAC;QAClB,CAAC;aAAM,IAAI,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;YAClE,MAAM,GAAG,QAAQ,CAAC;QACpB,CAAC;aAAM,IAAI,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;YACpE,MAAM,GAAG,QAAQ,CAAC;QACpB,CAAC;QAED,wBAAwB;QACxB,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;QAC5G,MAAM,QAAQ,GAAG,KAAK;aACnB,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;aACvB,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QAElD,OAAO;YACL,MAAM;YACN,QAAQ;YACR,MAAM,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,EAAE,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC;SACnF,CAAC;IACJ,CAAC;IAED,4DAA4D;IAEpD,WAAW,CACjB,KAAiB,EACjB,IAAY,EACZ,KAAa,EACb,GAAW,EACX,QAAuB,EACvB,MAAe;QAEf,OAAO;YACL,EAAE,EAAE,YAAY,EAAE;YAClB,KAAK;YACL,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE;YACjB,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC;YAChB,QAAQ;YACR,QAAQ,EAAE,EAAE;YACZ,QAAQ,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE;YACxB,QAAQ,EAAE;gBACR,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;gBACnC,MAAM;aACP;SACF,CAAC;IACJ,CAAC;IAEO,eAAe,CAAC,IAAY;QAClC,OAAO,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAChE,CAAC;IAEO,cAAc,CAAC,IAAY;QACjC,YAAY;QACZ,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClB,OAAO,IAAI,CAAC,QAAQ,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;QAC5C,CAAC;QACD,OAAO,IAAI,CAAC,KAAK,CAAC,kBAAkB,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACzE,CAAC;IAEO,YAAY,CAAC,IAAY;QAC/B,aAAa;QACb,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClB,OAAO,IAAI,CAAC,QAAQ,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;QAC1C,CAAC;QACD,OAAO,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACjE,CAAC;IAEO,UAAU,CAAC,IAAY;QAC7B,eAAe;QACf,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClB,OAAO,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;QACxC,CAAC;QACD,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC5D,CAAC;IAEO,cAAc,CAAC,IAAY;QACjC,aAAa;QACb,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClB,OAAO,IAAI,CAAC,QAAQ,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;QAC5C,CAAC;QACD,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC5D,CAAC;IAEO,cAAc,CAAC,IAAY;QACjC,WAAW;QACX,OAAO,CAAC,GAAG,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACpD,CAAC;IAED;;OAEG;IACH,mBAAmB;QACjB,OAAO,IAAI,CAAC,gBAAgB,CAAC;IAC/B,CAAC;CACF;AAED;;GAEG;AACH,MAAM,OAAO,mBAAmB;IACtB,OAAO,CAAsB;IAC7B,MAAM,CAIX;IAEH;QACE,IAAI,CAAC,OAAO,GAAG,IAAI,mBAAmB,EAAE,CAAC;QACzC,IAAI,CAAC,MAAM,GAAG,IAAI,GAAG,EAAE,CAAC;IAC1B,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,KAAa,EAAE,UAAkB,MAAM;QAC3C,MAAM,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,iBAAiB,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;QAE5D,UAAU;QACV,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE;YAC5B,IAAI;YACJ,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE;YACrB,OAAO;SACR,CAAC,CAAC;QAEH,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,QAAgB;QASpB,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,kBAAkB,CAAC,QAAQ,CAAC,CAAC;QACvE,MAAM,OAAO,GAKR,EAAE,CAAC;QAER,iBAAiB;QACjB,KAAK,MAAM,CAAC,EAAE,KAAK,CAAC,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YACpC,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,GAAG,KAAK,CAAC;YAEhC,gBAAgB;YAChB,KAAK,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;gBAC1C,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;oBAC3B,IAAI,SAAS,GAAG,CAAC,CAAC;oBAElB,YAAY;oBACZ,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;wBAC/B,IAAI,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,EAAE,CAAC;4BACjC,SAAS,IAAI,CAAC,CAAC;wBACjB,CAAC;oBACH,CAAC;oBAED,WAAW;oBACX,IAAI,MAAM,KAAK,UAAU,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;wBACtF,SAAS,IAAI,CAAC,CAAC;oBACjB,CAAC;oBAED,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;wBAClB,kBAAkB;wBAClB,IAAI,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC;wBACzB,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;4BACnB,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;4BAC9C,IAAI,MAAM;gCAAE,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC;wBACpC,CAAC;wBAED,OAAO,CAAC,IAAI,CAAC;4BACX,IAAI,EAAE,KAAK,CAAC,IAAI;4BAChB,KAAK,EAAE,KAAmB;4BAC1B,SAAS;4BACT,OAAO;yBACR,CAAC,CAAC;oBACL,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;QAED,aAAa;QACb,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,GAAG,CAAC,CAAC,SAAS,CAAC,CAAC;QAElD,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,CAAC;IAC7B,CAAC;IAED;;OAEG;IACH,QAAQ;QAKN,MAAM,aAAa,GAA+B;YAChD,QAAQ,EAAE,CAAC;YACX,SAAS,EAAE,CAAC;YACZ,QAAQ,EAAE,CAAC;YACX,MAAM,EAAE,CAAC;YACT,IAAI,EAAE,CAAC;YACP,QAAQ,EAAE,CAAC;YACX,QAAQ,EAAE,CAAC;SACZ,CAAC;QAEF,IAAI,WAAW,GAAG,CAAC,CAAC;QAEpB,KAAK,MAAM,CAAC,EAAE,KAAK,CAAC,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YACpC,KAAK,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;gBAChD,aAAa,CAAC,KAAmB,CAAC,IAAI,MAAM,CAAC,MAAM,CAAC;gBACpD,WAAW,IAAI,MAAM,CAAC,MAAM,CAAC;YAC/B,CAAC;QACH,CAAC;QAED,OAAO;YACL,kBAAkB,EAAE,IAAI,CAAC,MAAM,CAAC,IAAI;YACpC,WAAW;YACX,aAAa;SACd,CAAC;IACJ,CAAC;CACF;AAED,SAAS;AACT,MAAM,UAAU,yBAAyB;IACvC,OAAO,IAAI,mBAAmB,EAAE,CAAC;AACnC,CAAC;AAED,MAAM,UAAU,yBAAyB;IACvC,OAAO,IAAI,mBAAmB,EAAE,CAAC;AACnC,CAAC"}

package/dist/core/hnsw-index.d.ts

@@ -0,0 +1,111 @@
+/**
+ * HNSW Index - Hierarchical Navigable Small World Graph
+ * @module core/hnsw-index
+ */
+import type { UUID, Embedding384, HNSWParams, HNSWNode, HNSWIndexData, SearchResult } from '../types/index.js';
+/**
+ * HNSW Index implementation for approximate nearest neighbor search
+ */
+export declare class HNSWIndex {
+    private params;
+    private layers;
+    private nodes;
+    private entryPoint;
+    private maxLayer;
+    constructor(params?: Partial<HNSWParams>);
+    /**
+     * Get the number of nodes in the index
+     */
+    get size(): number;
+    /**
+     * Insert a new node into the index
+     * @param id - Node identifier
+     * @param embedding - 384-dimensional embedding vector
+     */
+    insert(id: UUID, embedding: Embedding384): void;
+    /**
+     * Search for k nearest neighbors
+     * @param query - Query embedding
+     * @param k - Number of neighbors to return
+     * @param ef - Search expansion factor (default from params)
+     * @returns Array of search results sorted by similarity
+     */
+    search(query: Embedding384, k: number, ef?: number): SearchResult[];
+    /**
+     * Delete a node from the index
+     * @param id - Node identifier to delete
+     */
+    delete(id: UUID): boolean;
+    /**
+     * Check if a node exists in the index
+     * @param id - Node identifier
+     */
+    has(id: UUID): boolean;
+    /**
+     * Get node by ID
+     * @param id - Node identifier
+     */
+    getNode(id: UUID): HNSWNode | undefined;
+    /**
+     * Get all node IDs
+     */
+    getAllIds(): UUID[];
+    /**
+     * Serialize the index to a JSON-friendly format
+     */
+    serialize(): HNSWIndexData;
+    /**
+     * Deserialize index from JSON format
+     */
+    static deserialize(data: HNSWIndexData): HNSWIndex;
+    /**
+     * Assign layer for new node using exponential distribution
+     */
+    private assignLayer;
+    /**
+     * Greedy search to find closest node at a given layer
+     */
+    private greedySearch;
+    /**
+     * Search a layer for candidates using beam search
+     */
+    private searchLayer;
+    /**
+     * Select best neighbors from candidates
+     */
+    private selectNeighbors;
+    /**
+     * Prune connections to maintain max connections limit
+     */
+    private pruneConnections;
+    /**
+     * Update entry point after deletion
+     */
+    private updateEntryPoint;
+    /**
+     * Ensure layer exists
+     */
+    private ensureLayer;
+    /**
+     * Get layer data
+     */
+    private getLayer;
+    /**
+     * Compute similarity between two embeddings
+     */
+    private similarity;
+    /**
+     * Get index statistics
+     */
+    getStats(): {
+        totalNodes: number;
+        layerDistribution: Map<number, number>;
+        avgConnectionsPerLayer: Map<number, number>;
+        maxLayer: number;
+    };
+}
+/**
+ * Create an HNSWIndex instance with default parameters
+ */
+export declare function createHNSWIndex(params?: Partial<HNSWParams>): HNSWIndex;
+//# sourceMappingURL=hnsw-index.d.ts.map

package/dist/core/hnsw-index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"hnsw-index.d.ts","sourceRoot":"","sources":["../../src/core/hnsw-index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,KAAK,EACV,IAAI,EACJ,YAAY,EACZ,UAAU,EACV,QAAQ,EAER,aAAa,EACb,YAAY,EAEb,MAAM,mBAAmB,CAAC;AAc3B;;GAEG;AACH,qBAAa,SAAS;IACpB,OAAO,CAAC,MAAM,CAAa;IAC3B,OAAO,CAAC,MAAM,CAAyB;IACvC,OAAO,CAAC,KAAK,CAAsB;IACnC,OAAO,CAAC,UAAU,CAAc;IAChC,OAAO,CAAC,QAAQ,CAAS;gBAEb,MAAM,GAAE,OAAO,CAAC,UAAU,CAAM;IAc5C;;OAEG;IACH,IAAI,IAAI,IAAI,MAAM,CAEjB;IAED;;;;OAIG;IACH,MAAM,CAAC,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE,YAAY,GAAG,IAAI;IAsF/C;;;;;;OAMG;IACH,MAAM,CAAC,KAAK,EAAE,YAAY,EAAE,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC,EAAE,MAAM,GAAG,YAAY,EAAE;IAoBnE;;;OAGG;IACH,MAAM,CAAC,EAAE,EAAE,IAAI,GAAG,OAAO;IAiCzB;;;OAGG;IACH,GAAG,CAAC,EAAE,EAAE,IAAI,GAAG,OAAO;IAItB;;;OAGG;IACH,OAAO,CAAC,EAAE,EAAE,IAAI,GAAG,QAAQ,GAAG,SAAS;IAIvC;;OAEG;IACH,SAAS,IAAI,IAAI,EAAE;IAInB;;OAEG;IACH,SAAS,IAAI,aAAa;IAkC1B;;OAEG;IACH,MAAM,CAAC,WAAW,CAAC,IAAI,EAAE,aAAa,GAAG,SAAS;IA8ClD;;OAEG;IACH,OAAO,CAAC,WAAW;IAMnB;;OAEG;IACH,OAAO,CAAC,YAAY;IA4BpB;;OAEG;IACH,OAAO,CAAC,WAAW;IAiEnB;;OAEG;IACH,OAAO,CAAC,eAAe;IAKvB;;OAEG;IACH,OAAO,CAAC,gBAAgB;IA2BxB;;OAEG;IACH,OAAO,CAAC,gBAAgB;IAgBxB;;OAEG;IACH,OAAO,CAAC,WAAW;IAMnB;;OAEG;IACH,OAAO,CAAC,QAAQ;IAIhB;;OAEG;IACH,OAAO,CAAC,UAAU;IAIlB;;OAEG;IACH,QAAQ,IAAI;QACV,UAAU,EAAE,MAAM,CAAC;QACnB,iBAAiB,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QACvC,sBAAsB,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QAC5C,QAAQ,EAAE,MAAM,CAAC;KAClB;CAgCF;AAED;;GAEG;AACH,wBAAgB,eAAe,CAAC,MAAM,CAAC,EAAE,OAAO,CAAC,UAAU,CAAC,GAAG,SAAS,CAEvE"}