@nahisaho/katashiro-analyzer 0.1.0

package/dist/text/text-analyzer.js

@@ -0,0 +1,281 @@
+/**
+ * TextAnalyzer - テキスト分析
+ *
+ * @requirement REQ-ANALYZE-001, REQ-ANALYZE-006
+ * @design DES-KATASHIRO-001 §2.2 Analyzer Container
+ * @task TSK-020
+ */
+import { ok, err, generateId, formatTimestamp, } from '@nahisaho/katashiro-core';
+/**
+ * 停止語リスト（英語）
+ */
+const STOP_WORDS = new Set([
+    'a', 'an', 'the', 'and', 'or', 'but', 'in', 'on', 'at', 'to', 'for',
+    'of', 'with', 'by', 'from', 'as', 'is', 'was', 'are', 'were', 'been',
+    'be', 'have', 'has', 'had', 'do', 'does', 'did', 'will', 'would', 'could',
+    'should', 'may', 'might', 'must', 'shall', 'can', 'need', 'dare', 'ought',
+    'used', 'it', 'its', 'this', 'that', 'these', 'those', 'i', 'you', 'he',
+    'she', 'we', 'they', 'what', 'which', 'who', 'whom', 'when', 'where',
+    'why', 'how', 'all', 'each', 'every', 'both', 'few', 'more', 'most',
+    'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so',
+    'than', 'too', 'very', 'just', 'also', 'now', 'over', 'into', 'about',
+]);
+/**
+ * ポジティブワードリスト
+ */
+const POSITIVE_WORDS = new Set([
+    'good', 'great', 'excellent', 'amazing', 'wonderful', 'fantastic', 'love',
+    'best', 'happy', 'joy', 'beautiful', 'perfect', 'awesome', 'brilliant',
+    'outstanding', 'superb', 'magnificent', 'delightful', 'pleasant', 'nice',
+    'positive', 'success', 'successful', 'win', 'winning', 'benefit', 'improve',
+    'improvement', 'helpful', 'useful', 'valuable', 'impressive', 'incredible',
+]);
+/**
+ * ネガティブワードリスト
+ */
+const NEGATIVE_WORDS = new Set([
+    'bad', 'terrible', 'awful', 'horrible', 'worst', 'hate', 'disappointed',
+    'disappointing', 'frustrating', 'frustrated', 'angry', 'sad', 'poor',
+    'failure', 'fail', 'failed', 'problem', 'issue', 'wrong', 'error',
+    'mistake', 'difficult', 'hard', 'impossible', 'negative', 'loss', 'lose',
+    'losing', 'damage', 'harmful', 'dangerous', 'risk', 'threat', 'concern',
+    'worried', 'worry', 'fear', 'scary', 'terrible', 'dreadful', 'miserable',
+]);
+/**
+ * テキスト分析実装
+ */
+export class TextAnalyzer {
+    /**
+     * テキストを要約
+     */
+    async summarize(content, maxLength) {
+        const text = content.body?.trim() ?? '';
+        if (!text) {
+            return err(new Error('Content text is empty'));
+        }
+        const sentences = this.splitSentences(text);
+        const keyPoints = this.extractKeyPoints(sentences);
+        const summaryText = this.generateSummary(sentences, maxLength ?? 200);
+        const summary = {
+            id: generateId('SUM'),
+            text: summaryText,
+            keyPoints,
+            wordCount: this.countWords(text),
+            createdAt: formatTimestamp(),
+        };
+        return ok(summary);
+    }
+    /**
+     * 感情分析（ISentimentAnalyzerインターフェース実装）
+     */
+    async analyze(content) {
+        return this.analyzeSentiment(content);
+    }
+    /**
+     * 感情分析
+     */
+    async analyzeSentiment(content) {
+        const text = content.body?.trim() ?? '';
+        if (!text) {
+            return err(new Error('Content text is empty'));
+        }
+        const words = this.tokenize(text.toLowerCase());
+        let positiveCount = 0;
+        let negativeCount = 0;
+        for (const word of words) {
+            if (POSITIVE_WORDS.has(word))
+                positiveCount++;
+            if (NEGATIVE_WORDS.has(word))
+                negativeCount++;
+        }
+        const total = positiveCount + negativeCount;
+        let score = 0;
+        let sentiment = 'neutral';
+        if (total > 0) {
+            score = (positiveCount - negativeCount) / total;
+            if (score > 0.2) {
+                sentiment = 'positive';
+            }
+            else if (score < -0.2) {
+                sentiment = 'negative';
+            }
+            else if (positiveCount > 0 && negativeCount > 0) {
+                sentiment = 'mixed';
+            }
+        }
+        // Confidence based on total sentiment words found
+        const confidence = Math.min(total / 10, 1);
+        const result = {
+            sentiment,
+            score,
+            confidence,
+            aspects: this.extractAspectSentiments(text),
+        };
+        return ok(result);
+    }
+    /**
+     * キーワードを抽出
+     */
+    extractKeywords(text, maxKeywords = 10) {
+        const words = this.tokenize(text.toLowerCase());
+        const wordFreq = new Map();
+        for (const word of words) {
+            if (word.length < 3 || STOP_WORDS.has(word))
+                continue;
+            wordFreq.set(word, (wordFreq.get(word) ?? 0) + 1);
+        }
+        return Array.from(wordFreq.entries())
+            .sort((a, b) => b[1] - a[1])
+            .slice(0, maxKeywords)
+            .map(([word]) => word);
+    }
+    /**
+     * 単語数をカウント
+     */
+    countWords(text) {
+        const words = text.trim().split(/\s+/).filter(w => w.length > 0);
+        return words.length;
+    }
+    /**
+     * 文に分割
+     */
+    splitSentences(text) {
+        // Simple sentence splitting - handles common abbreviations
+        const normalized = text
+            .replace(/([.!?])\s+/g, '$1\n')
+            .replace(/\n+/g, '\n');
+        return normalized
+            .split('\n')
+            .map(s => s.trim())
+            .filter(s => s.length > 0);
+    }
+    /**
+     * トークン化
+     */
+    tokenize(text) {
+        return text
+            .replace(/[^\w\s]/g, ' ')
+            .split(/\s+/)
+            .filter(word => word.length > 0);
+    }
+    /**
+     * キーポイントを抽出
+     */
+    extractKeyPoints(sentences) {
+        // Score sentences by importance indicators
+        const scoredSentences = sentences.map(sentence => {
+            let score = 0;
+            const lower = sentence.toLowerCase();
+            // Indicators of important sentences
+            if (lower.includes('important'))
+                score += 2;
+            if (lower.includes('key'))
+                score += 2;
+            if (lower.includes('main'))
+                score += 2;
+            if (lower.includes('significant'))
+                score += 2;
+            if (lower.includes('conclusion'))
+                score += 3;
+            if (lower.includes('result'))
+                score += 2;
+            if (lower.includes('finding'))
+                score += 2;
+            if (/^\d+\.\s/.test(sentence))
+                score += 1; // Numbered items
+            if (sentence.includes(':'))
+                score += 1;
+            // Penalize very short or very long sentences
+            const wordCount = this.countWords(sentence);
+            if (wordCount < 5)
+                score -= 2;
+            if (wordCount > 40)
+                score -= 1;
+            return { sentence, score };
+        });
+        // Get top sentences as key points
+        return scoredSentences
+            .filter(s => s.score > 0 || scoredSentences.length <= 5)
+            .sort((a, b) => b.score - a.score)
+            .slice(0, 5)
+            .map(s => s.sentence);
+    }
+    /**
+     * 要約を生成
+     */
+    generateSummary(sentences, maxLength) {
+        if (sentences.length === 0)
+            return '';
+        // Use extractive summarization - pick most important sentences
+        const scoredSentences = sentences.map((sentence, index) => {
+            let score = 0;
+            // Position score - first sentences often contain key info
+            if (index === 0)
+                score += 3;
+            if (index < 3)
+                score += 1;
+            // Length score - medium length sentences are often better summaries
+            const wordCount = this.countWords(sentence);
+            if (wordCount >= 10 && wordCount <= 25)
+                score += 2;
+            // Keyword density
+            const keywords = this.extractKeywords(sentence, 3);
+            score += keywords.length * 0.5;
+            return { sentence, score, index };
+        });
+        // Sort by score and build summary
+        const sortedSentences = scoredSentences
+            .sort((a, b) => b.score - a.score);
+        let summary = '';
+        const usedIndices = new Set();
+        for (const { sentence, index } of sortedSentences) {
+            if (summary.length + sentence.length > maxLength) {
+                if (summary.length > 0)
+                    break;
+            }
+            if (!usedIndices.has(index)) {
+                usedIndices.add(index);
+                summary += (summary ? ' ' : '') + sentence;
+            }
+        }
+        return summary || sentences[0] || '';
+    }
+    /**
+     * アスペクトごとの感情を抽出
+     */
+    extractAspectSentiments(text) {
+        // Simple aspect extraction based on common patterns
+        const aspects = [];
+        const sentences = this.splitSentences(text);
+        const aspectPatterns = [
+            { pattern: /quality/i, aspect: 'quality' },
+            { pattern: /price|cost/i, aspect: 'price' },
+            { pattern: /service/i, aspect: 'service' },
+            { pattern: /design|look/i, aspect: 'design' },
+            { pattern: /performance|speed/i, aspect: 'performance' },
+        ];
+        for (const { pattern, aspect } of aspectPatterns) {
+            for (const sentence of sentences) {
+                if (pattern.test(sentence)) {
+                    const words = this.tokenize(sentence.toLowerCase());
+                    let positiveCount = 0;
+                    let negativeCount = 0;
+                    for (const word of words) {
+                        if (POSITIVE_WORDS.has(word))
+                            positiveCount++;
+                        if (NEGATIVE_WORDS.has(word))
+                            negativeCount++;
+                    }
+                    const total = positiveCount + negativeCount;
+                    if (total > 0) {
+                        const score = (positiveCount - negativeCount) / total;
+                        const sentiment = score > 0.1 ? 'positive' : score < -0.1 ? 'negative' : 'neutral';
+                        aspects.push({ aspect, sentiment, score });
+                    }
+                }
+            }
+        }
+        return aspects;
+    }
+}
+//# sourceMappingURL=text-analyzer.js.map

package/dist/text/text-analyzer.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"text-analyzer.js","sourceRoot":"","sources":["../../src/text/text-analyzer.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAEL,EAAE,EACF,GAAG,EACH,UAAU,EACV,eAAe,GAEhB,MAAM,0BAA0B,CAAC;AAIlC;;GAEG;AACH,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC;IACzB,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK;IACnE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM;IACpE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO;IACzE,QAAQ,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO;IACzE,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,EAAE,GAAG,EAAE,KAAK,EAAE,IAAI;IACvE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO;IACpE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM;IACnE,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI;IACxE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO;CACtE,CAAC,CAAC;AAEH;;GAEG;AACH,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC;IAC7B,MAAM,EAAE,OAAO,EAAE,WAAW,EAAE,SAAS,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM;IACzE,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,WAAW,EAAE,SAAS,EAAE,SAAS,EAAE,WAAW;IACtE,aAAa,EAAE,QAAQ,EAAE,aAAa,EAAE,YAAY,EAAE,UAAU,EAAE,MAAM;IACxE,UAAU,EAAE,SAAS,EAAE,YAAY,EAAE,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,SAAS;IAC3E,aAAa,EAAE,SAAS,EAAE,QAAQ,EAAE,UAAU,EAAE,YAAY,EAAE,YAAY;CAC3E,CAAC,CAAC;AAEH;;GAEG;AACH,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC;IAC7B,KAAK,EAAE,UAAU,EAAE,OAAO,EAAE,UAAU,EAAE,OAAO,EAAE,MAAM,EAAE,cAAc;IACvE,eAAe,EAAE,aAAa,EAAE,YAAY,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM;IACpE,SAAS,EAAE,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO;IACjE,SAAS,EAAE,WAAW,EAAE,MAAM,EAAE,YAAY,EAAE,UAAU,EAAE,MAAM,EAAE,MAAM;IACxE,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,EAAE,QAAQ,EAAE,SAAS;IACvE,SAAS,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,UAAU,EAAE,UAAU,EAAE,WAAW;CACzE,CAAC,CAAC;AAEH;;GAEG;AACH,MAAM,OAAO,YAAY;IACvB;;OAEG;IACH,KAAK,CAAC,SAAS,CACb,OAAgB,EAChB,SAAkB;QAElB,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;QAExC,IAAI,CAAC,IAAI,EAAE,CAAC;YACV,OAAO,GAAG,CAAC,IAAI,KAAK,CAAC,uBAAuB,CAAC,CAAC,CAAC;QACjD,CAAC;QAED,MAAM,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;QAC5C,MAAM,SAAS,GAAG,IAAI,CAAC,gBAAgB,CAAC,SAAS,CAAC,CAAC;QACnD,MAAM,WAAW,GAAG,IAAI,CAAC,eAAe,CAAC,SAAS,EAAE,SAAS,IAAI,GAAG,CAAC,CAAC;QAEtE,MAAM,OAAO,GAAY;YACvB,EAAE,EAAE,UAAU,CAAC,KAAK,CAAC;YACrB,IAAI,EAAE,WAAW;YACjB,SAAS;YACT,SAAS,EAAE,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC;YAChC,SAAS,EAAE,eAAe,EAAE;SAC7B,CAAC;QAEF,OAAO,EAAE,CAAC,OAAO,CAAC,CAAC;IACrB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,OAAO,CAAC,OAAgB;QAC5B,OAAO,IAAI,CAAC,gBAAgB,CAAC,OAAO,CAAC,CAAC;IACxC,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,gBAAgB,CAAC,OAAgB;QACrC,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;QAExC,IAAI,CAAC,IAAI,EAAE,CAAC;YACV,OAAO,GAAG,CAAC,IAAI,KAAK,CAAC,uBAAuB,CAAC,CAAC,CAAC;QACjD,CAAC;QAED,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;QAChD,IAAI,aAAa,GAAG,CAAC,CAAC;QACtB,IAAI,aAAa,GAAG,CAAC,CAAC;QAEtB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,IAAI,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC;gBAAE,aAAa,EAAE,CAAC;YAC9C,IAAI,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC;gBAAE,aAAa,EAAE,CAAC;QAChD,CAAC;QAED,MAAM,KAAK,GAAG,aAAa,GAAG,aAAa,CAAC;QAC5C,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,IAAI,SAAS,GAAkD,SAAS,CAAC;QAEzE,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YACd,KAAK,GAAG,CAAC,aAAa,GAAG,aAAa,CAAC,GAAG,KAAK,CAAC;YAEhD,IAAI,KAAK,GAAG,GAAG,EAAE,CAAC;gBAChB,SAAS,GAAG,UAAU,CAAC;YACzB,CAAC;iBAAM,IAAI,KAAK,GAAG,CAAC,GAAG,EAAE,CAAC;gBACxB,SAAS,GAAG,UAAU,CAAC;YACzB,CAAC;iBAAM,IAAI,aAAa,GAAG,CAAC,IAAI,aAAa,GAAG,CAAC,EAAE,CAAC;gBAClD,SAAS,GAAG,OAAO,CAAC;YACtB,CAAC;QACH,CAAC;QAED,kDAAkD;QAClD,MAAM,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC;QAE3C,MAAM,MAAM,GAAoB;YAC9B,SAAS;YACT,KAAK;YACL,UAAU;YACV,OAAO,EAAE,IAAI,CAAC,uBAAuB,CAAC,IAAI,CAAC;SAC5C,CAAC;QAEF,OAAO,EAAE,CAAC,MAAM,CAAC,CAAC;IACpB,CAAC;IAED;;OAEG;IACH,eAAe,CAAC,IAAY,EAAE,cAAsB,EAAE;QACpD,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;QAChD,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;QAE3C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC;gBAAE,SAAS;YACtD,QAAQ,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QACpD,CAAC;QAED,OAAO,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,CAAC;aAClC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;aAC3B,KAAK,CAAC,CAAC,EAAE,WAAW,CAAC;aACrB,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC;IAC3B,CAAC;IAED;;OAEG;IACH,UAAU,CAAC,IAAY;QACrB,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QACjE,OAAO,KAAK,CAAC,MAAM,CAAC;IACtB,CAAC;IAED;;OAEG;IACH,cAAc,CAAC,IAAY;QACzB,2DAA2D;QAC3D,MAAM,UAAU,GAAG,IAAI;aACpB,OAAO,CAAC,aAAa,EAAE,MAAM,CAAC;aAC9B,OAAO,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;QAEzB,OAAO,UAAU;aACd,KAAK,CAAC,IAAI,CAAC;aACX,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;aAClB,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC/B,CAAC;IAED;;OAEG;IACK,QAAQ,CAAC,IAAY;QAC3B,OAAO,IAAI;aACR,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC;aACxB,KAAK,CAAC,KAAK,CAAC;aACZ,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACrC,CAAC;IAED;;OAEG;IACK,gBAAgB,CAAC,SAAmB;QAC1C,2CAA2C;QAC3C,MAAM,eAAe,GAAG,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE;YAC/C,IAAI,KAAK,GAAG,CAAC,CAAC;YACd,MAAM,KAAK,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC;YAErC,oCAAoC;YACpC,IAAI,KAAK,CAAC,QAAQ,CAAC,WAAW,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAC5C,IAAI,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YACtC,IAAI,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YACvC,IAAI,KAAK,CAAC,QAAQ,CAAC,aAAa,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAC9C,IAAI,KAAK,CAAC,QAAQ,CAAC,YAAY,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAC7C,IAAI,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YACzC,IAAI,KAAK,CAAC,QAAQ,CAAC,SAAS,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAC1C,IAAI,UAAU,CAAC,IAAI,CAAC,QAAQ,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC,CAAC,iBAAiB;YAC5D,IAAI,QAAQ,CAAC,QAAQ,CAAC,GAAG,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAEvC,6CAA6C;YAC7C,MAAM,SAAS,GAAG,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC;YAC5C,IAAI,SAAS,GAAG,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAC9B,IAAI,SAAS,GAAG,EAAE;gBAAE,KAAK,IAAI,CAAC,CAAC;YAE/B,OAAO,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC;QAC7B,CAAC,CAAC,CAAC;QAEH,kCAAkC;QAClC,OAAO,eAAe;aACnB,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,IAAI,eAAe,CAAC,MAAM,IAAI,CAAC,CAAC;aACvD,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC;aACjC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC;aACX,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC;IAC1B,CAAC;IAED;;OAEG;IACK,eAAe,CAAC,SAAmB,EAAE,SAAiB;QAC5D,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEtC,+DAA+D;QAC/D,MAAM,eAAe,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE;YACxD,IAAI,KAAK,GAAG,CAAC,CAAC;YAEd,0DAA0D;YAC1D,IAAI,KAAK,KAAK,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAC5B,IAAI,KAAK,GAAG,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAE1B,oEAAoE;YACpE,MAAM,SAAS,GAAG,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC;YAC5C,IAAI,SAAS,IAAI,EAAE,IAAI,SAAS,IAAI,EAAE;gBAAE,KAAK,IAAI,CAAC,CAAC;YAEnD,kBAAkB;YAClB,MAAM,QAAQ,GAAG,IAAI,CAAC,eAAe,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC;YACnD,KAAK,IAAI,QAAQ,CAAC,MAAM,GAAG,GAAG,CAAC;YAE/B,OAAO,EAAE,QAAQ,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC;QACpC,CAAC,CAAC,CAAC;QAEH,kCAAkC;QAClC,MAAM,eAAe,GAAG,eAAe;aACpC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;QAErC,IAAI,OAAO,GAAG,EAAE,CAAC;QACjB,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;QAEtC,KAAK,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,IAAI,eAAe,EAAE,CAAC;YAClD,IAAI,OAAO,CAAC,MAAM,GAAG,QAAQ,CAAC,MAAM,GAAG,SAAS,EAAE,CAAC;gBACjD,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC;oBAAE,MAAM;YAChC,CAAC;YAED,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC5B,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;gBACvB,OAAO,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC;YAC7C,CAAC;QACH,CAAC;QAED,OAAO,OAAO,IAAI,SAAS,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IACvC,CAAC;IAED;;OAEG;IACK,uBAAuB,CAAC,IAAY;QAC1C,oDAAoD;QACpD,MAAM,OAAO,GAAsB,EAAE,CAAC;QACtC,MAAM,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;QAE5C,MAAM,cAAc,GAAG;YACrB,EAAE,OAAO,EAAE,UAAU,EAAE,MAAM,EAAE,SAAS,EAAE;YAC1C,EAAE,OAAO,EAAE,aAAa,EAAE,MAAM,EAAE,OAAO,EAAE;YAC3C,EAAE,OAAO,EAAE,UAAU,EAAE,MAAM,EAAE,SAAS,EAAE;YAC1C,EAAE,OAAO,EAAE,cAAc,EAAE,MAAM,EAAE,QAAQ,EAAE;YAC7C,EAAE,OAAO,EAAE,oBAAoB,EAAE,MAAM,EAAE,aAAa,EAAE;SACzD,CAAC;QAEF,KAAK,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,IAAI,cAAc,EAAE,CAAC;YACjD,KAAK,MAAM,QAAQ,IAAI,SAAS,EAAE,CAAC;gBACjC,IAAI,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC;oBAC3B,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC,CAAC;oBACpD,IAAI,aAAa,GAAG,CAAC,CAAC;oBACtB,IAAI,aAAa,GAAG,CAAC,CAAC;oBAEtB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;wBACzB,IAAI,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC;4BAAE,aAAa,EAAE,CAAC;wBAC9C,IAAI,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC;4BAAE,aAAa,EAAE,CAAC;oBAChD,CAAC;oBAED,MAAM,KAAK,GAAG,aAAa,GAAG,aAAa,CAAC;oBAC5C,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;wBACd,MAAM,KAAK,GAAG,CAAC,aAAa,GAAG,aAAa,CAAC,GAAG,KAAK,CAAC;wBACtD,MAAM,SAAS,GAAG,KAAK,GAAG,GAAG,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,SAAS,CAAC;wBAEnF,OAAO,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,CAAC;oBAC7C,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;CACF"}

package/dist/topic/index.d.ts

@@ -0,0 +1,6 @@
+/**
+ * Topic module exports
+ */
+export { TopicModeler, } from './topic-modeler.js';
+export type { Topic, TopicDistribution, SimilarDocument, DocumentCluster, } from './topic-modeler.js';
+//# sourceMappingURL=index.d.ts.map

package/dist/topic/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/topic/index.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,EACL,YAAY,GACb,MAAM,oBAAoB,CAAC;AAE5B,YAAY,EACV,KAAK,EACL,iBAAiB,EACjB,eAAe,EACf,eAAe,GAChB,MAAM,oBAAoB,CAAC"}

package/dist/topic/index.js

@@ -0,0 +1,5 @@
+/**
+ * Topic module exports
+ */
+export { TopicModeler, } from './topic-modeler.js';
+//# sourceMappingURL=index.js.map

package/dist/topic/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/topic/index.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,EACL,YAAY,GACb,MAAM,oBAAoB,CAAC"}

package/dist/topic/topic-modeler.d.ts

@@ -0,0 +1,79 @@
+/**
+ * TopicModeler - トピックモデリング
+ *
+ * @requirement REQ-ANALYZE-010
+ * @design DES-KATASHIRO-001 §2.2 Analyzer Container
+ * @task TSK-024
+ */
+/**
+ * トピック
+ */
+export interface Topic {
+    readonly id: string;
+    readonly name: string;
+    readonly keywords: string[];
+    readonly confidence: number;
+}
+/**
+ * トピック分布
+ */
+export interface TopicDistribution {
+    readonly topics: Array<{
+        readonly topic: Topic;
+        readonly weight: number;
+    }>;
+}
+/**
+ * 類似文書
+ */
+export interface SimilarDocument {
+    readonly index: number;
+    readonly text: string;
+    readonly similarity: number;
+}
+/**
+ * 文書クラスタ
+ */
+export interface DocumentCluster {
+    readonly id: number;
+    readonly documents: string[];
+    readonly centroid: string[];
+}
+/**
+ * トピックモデリング実装
+ */
+export declare class TopicModeler {
+    /**
+     * 文書からトピックを抽出
+     */
+    extractTopics(text: string, numTopics?: number): Topic[];
+    /**
+     * TF-IDF計算
+     */
+    calculateTfIdf(documents: string[]): Record<string, number>;
+    /**
+     * 文書クラスタリング
+     */
+    clusterDocuments(documents: string[], numClusters: number): DocumentCluster[];
+    /**
+     * 類似文書検索
+     */
+    findSimilarDocuments(query: string, corpus: string[], topK?: number): SimilarDocument[];
+    /**
+     * トピック分布取得
+     */
+    getTopicDistribution(text: string): TopicDistribution;
+    /**
+     * テキストをトークン化
+     */
+    private tokenize;
+    /**
+     * 単語頻度計算
+     */
+    private calculateTermFrequency;
+    /**
+     * Jaccard類似度計算
+     */
+    private jaccardSimilarity;
+}
+//# sourceMappingURL=topic-modeler.d.ts.map

package/dist/topic/topic-modeler.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"topic-modeler.d.ts","sourceRoot":"","sources":["../../src/topic/topic-modeler.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH;;GAEG;AACH,MAAM,WAAW,KAAK;IACpB,QAAQ,CAAC,EAAE,EAAE,MAAM,CAAC;IACpB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,QAAQ,EAAE,MAAM,EAAE,CAAC;IAC5B,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;CAC7B;AAED;;GAEG;AACH,MAAM,WAAW,iBAAiB;IAChC,QAAQ,CAAC,MAAM,EAAE,KAAK,CAAC;QACrB,QAAQ,CAAC,KAAK,EAAE,KAAK,CAAC;QACtB,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;KACzB,CAAC,CAAC;CACJ;AAED;;GAEG;AACH,MAAM,WAAW,eAAe;IAC9B,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;CAC7B;AAED;;GAEG;AACH,MAAM,WAAW,eAAe;IAC9B,QAAQ,CAAC,EAAE,EAAE,MAAM,CAAC;IACpB,QAAQ,CAAC,SAAS,EAAE,MAAM,EAAE,CAAC;IAC7B,QAAQ,CAAC,QAAQ,EAAE,MAAM,EAAE,CAAC;CAC7B;AAyBD;;GAEG;AACH,qBAAa,YAAY;IACvB;;OAEG;IACH,aAAa,CAAC,IAAI,EAAE,MAAM,EAAE,SAAS,GAAE,MAAU,GAAG,KAAK,EAAE;IAqC3D;;OAEG;IACH,cAAc,CAAC,SAAS,EAAE,MAAM,EAAE,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC;IAwC3D;;OAEG;IACH,gBAAgB,CAAC,SAAS,EAAE,MAAM,EAAE,EAAE,WAAW,EAAE,MAAM,GAAG,eAAe,EAAE;IAgF7E;;OAEG;IACH,oBAAoB,CAClB,KAAK,EAAE,MAAM,EACb,MAAM,EAAE,MAAM,EAAE,EAChB,IAAI,GAAE,MAAU,GACf,eAAe,EAAE;IA2BpB;;OAEG;IACH,oBAAoB,CAAC,IAAI,EAAE,MAAM,GAAG,iBAAiB;IAkBrD;;OAEG;IACH,OAAO,CAAC,QAAQ;IAgBhB;;OAEG;IACH,OAAO,CAAC,sBAAsB;IAQ9B;;OAEG;IACH,OAAO,CAAC,iBAAiB;CAc1B"}

package/dist/topic/topic-modeler.js

@@ -0,0 +1,267 @@
+/**
+ * TopicModeler - トピックモデリング
+ *
+ * @requirement REQ-ANALYZE-010
+ * @design DES-KATASHIRO-001 §2.2 Analyzer Container
+ * @task TSK-024
+ */
+/**
+ * 日本語ストップワード
+ */
+const JAPANESE_STOPWORDS = new Set([
+    'の', 'に', 'は', 'を', 'た', 'が', 'で', 'て', 'と', 'し', 'れ', 'さ',
+    'ある', 'いる', 'も', 'な', 'する', 'から', 'こと', 'として', 'い', 'や',
+    'など', 'なっ', 'ない', 'この', 'ため', 'その', 'あっ', 'よう', 'また',
+    'もの', 'という', 'あり', 'まで', 'られ', 'なる', 'へ', 'か', 'だ', 'これ',
+    'によって', 'により', 'おり', 'より', 'による', 'ず', 'なり', 'られる',
+    'です', 'ます', 'した', 'して', 'です', 'ました',
+]);
+/**
+ * 英語ストップワード
+ */
+const ENGLISH_STOPWORDS = new Set([
+    'the', 'a', 'an', 'and', 'or', 'but', 'in', 'on', 'at', 'to', 'for',
+    'of', 'with', 'by', 'from', 'up', 'about', 'into', 'through', 'during',
+    'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had',
+    'do', 'does', 'did', 'will', 'would', 'could', 'should', 'may', 'might',
+    'this', 'that', 'these', 'those', 'it', 'its',
+]);
+/**
+ * トピックモデリング実装
+ */
+export class TopicModeler {
+    /**
+     * 文書からトピックを抽出
+     */
+    extractTopics(text, numTopics = 3) {
+        if (!text || text.trim().length === 0) {
+            return [];
+        }
+        // Tokenize and get term frequencies
+        const tokens = this.tokenize(text);
+        const termFreq = this.calculateTermFrequency(tokens);
+        // Sort by frequency and get top terms
+        const sortedTerms = Array.from(termFreq.entries())
+            .sort((a, b) => b[1] - a[1])
+            .slice(0, numTopics * 5);
+        // Group into topics
+        const topics = [];
+        const termsPerTopic = Math.ceil(sortedTerms.length / numTopics);
+        for (let i = 0; i < numTopics && i * termsPerTopic < sortedTerms.length; i++) {
+            const topicTerms = sortedTerms.slice(i * termsPerTopic, (i + 1) * termsPerTopic);
+            if (topicTerms.length === 0)
+                continue;
+            const keywords = topicTerms.map(t => t[0]);
+            const avgFreq = topicTerms.reduce((sum, t) => sum + t[1], 0) / topicTerms.length;
+            const maxFreq = Math.max(...Array.from(termFreq.values()));
+            topics.push({
+                id: `topic-${i + 1}`,
+                name: keywords[0] ?? `Topic ${i + 1}`,
+                keywords,
+                confidence: maxFreq > 0 ? Math.min(avgFreq / maxFreq + 0.3, 1) : 0.5,
+            });
+        }
+        return topics;
+    }
+    /**
+     * TF-IDF計算
+     */
+    calculateTfIdf(documents) {
+        if (documents.length === 0) {
+            return {};
+        }
+        // Tokenize all documents
+        const tokenizedDocs = documents.map(doc => this.tokenize(doc));
+        // Calculate document frequency
+        const docFreq = new Map();
+        for (const tokens of tokenizedDocs) {
+            const uniqueTokens = new Set(tokens);
+            for (const token of uniqueTokens) {
+                docFreq.set(token, (docFreq.get(token) ?? 0) + 1);
+            }
+        }
+        // Calculate TF-IDF for each term
+        const tfidf = {};
+        const numDocs = documents.length;
+        for (const tokens of tokenizedDocs) {
+            const termFreq = this.calculateTermFrequency(tokens);
+            const maxFreq = Math.max(...Array.from(termFreq.values()), 1);
+            for (const [term, freq] of termFreq) {
+                const tf = freq / maxFreq;
+                const df = docFreq.get(term) ?? 1;
+                const idf = Math.log((numDocs + 1) / (df + 1)) + 1;
+                const score = tf * idf;
+                if (!tfidf[term] || score > tfidf[term]) {
+                    tfidf[term] = score;
+                }
+            }
+        }
+        return tfidf;
+    }
+    /**
+     * 文書クラスタリング
+     */
+    clusterDocuments(documents, numClusters) {
+        if (documents.length === 0) {
+            return [];
+        }
+        if (documents.length <= numClusters) {
+            return documents.map((doc, i) => ({
+                id: i,
+                documents: [doc],
+                centroid: this.tokenize(doc).slice(0, 5),
+            }));
+        }
+        // Simple clustering based on term overlap
+        const tokenizedDocs = documents.map(doc => new Set(this.tokenize(doc)));
+        const clusters = [];
+        const assigned = new Set();
+        for (let c = 0; c < numClusters && assigned.size < documents.length; c++) {
+            // Find unassigned document as seed
+            let seedIdx = -1;
+            for (let i = 0; i < documents.length; i++) {
+                if (!assigned.has(i)) {
+                    seedIdx = i;
+                    break;
+                }
+            }
+            if (seedIdx === -1)
+                break;
+            const clusterDocs = [];
+            const clusterIndices = [];
+            const seedTokens = tokenizedDocs[seedIdx];
+            if (!seedTokens)
+                continue;
+            // Find similar documents
+            for (let i = 0; i < documents.length; i++) {
+                if (assigned.has(i))
+                    continue;
+                const docTokens = tokenizedDocs[i];
+                if (!docTokens)
+                    continue;
+                const similarity = this.jaccardSimilarity(seedTokens, docTokens);
+                if (similarity > 0.1 || i === seedIdx) {
+                    const doc = documents[i];
+                    if (doc) {
+                        clusterDocs.push(doc);
+                        clusterIndices.push(i);
+                        assigned.add(i);
+                    }
+                }
+            }
+            if (clusterDocs.length > 0) {
+                // Calculate centroid (most common terms)
+                const termCounts = new Map();
+                for (const idx of clusterIndices) {
+                    const tokens = tokenizedDocs[idx];
+                    if (tokens) {
+                        for (const token of tokens) {
+                            termCounts.set(token, (termCounts.get(token) ?? 0) + 1);
+                        }
+                    }
+                }
+                const centroid = Array.from(termCounts.entries())
+                    .sort((a, b) => b[1] - a[1])
+                    .slice(0, 5)
+                    .map(e => e[0]);
+                clusters.push({
+                    id: c,
+                    documents: clusterDocs,
+                    centroid,
+                });
+            }
+        }
+        return clusters;
+    }
+    /**
+     * 類似文書検索
+     */
+    findSimilarDocuments(query, corpus, topK = 5) {
+        if (corpus.length === 0) {
+            return [];
+        }
+        const queryTokens = new Set(this.tokenize(query));
+        const similarities = [];
+        for (let i = 0; i < corpus.length; i++) {
+            const doc = corpus[i];
+            if (!doc)
+                continue;
+            const docTokens = new Set(this.tokenize(doc));
+            const similarity = this.jaccardSimilarity(queryTokens, docTokens);
+            similarities.push({
+                index: i,
+                text: doc,
+                similarity,
+            });
+        }
+        return similarities
+            .sort((a, b) => b.similarity - a.similarity)
+            .slice(0, topK);
+    }
+    /**
+     * トピック分布取得
+     */
+    getTopicDistribution(text) {
+        const topics = this.extractTopics(text, 3);
+        if (topics.length === 0) {
+            return { topics: [] };
+        }
+        // Normalize weights to sum to 1
+        const totalConfidence = topics.reduce((sum, t) => sum + t.confidence, 0);
+        return {
+            topics: topics.map(topic => ({
+                topic,
+                weight: totalConfidence > 0 ? topic.confidence / totalConfidence : 1 / topics.length,
+            })),
+        };
+    }
+    /**
+     * テキストをトークン化
+     */
+    tokenize(text) {
+        // Simple tokenization: split by non-word characters
+        const tokens = text
+            .toLowerCase()
+            .split(/[\s\p{P}]+/u)
+            .filter(token => {
+            if (token.length < 2)
+                return false;
+            if (JAPANESE_STOPWORDS.has(token))
+                return false;
+            if (ENGLISH_STOPWORDS.has(token))
+                return false;
+            if (/^\d+$/.test(token))
+                return false;
+            return true;
+        });
+        return tokens;
+    }
+    /**
+     * 単語頻度計算
+     */
+    calculateTermFrequency(tokens) {
+        const freq = new Map();
+        for (const token of tokens) {
+            freq.set(token, (freq.get(token) ?? 0) + 1);
+        }
+        return freq;
+    }
+    /**
+     * Jaccard類似度計算
+     */
+    jaccardSimilarity(set1, set2) {
+        if (set1.size === 0 && set2.size === 0)
+            return 1;
+        if (set1.size === 0 || set2.size === 0)
+            return 0;
+        let intersection = 0;
+        for (const item of set1) {
+            if (set2.has(item)) {
+                intersection++;
+            }
+        }
+        const union = set1.size + set2.size - intersection;
+        return union > 0 ? intersection / union : 0;
+    }
+}
+//# sourceMappingURL=topic-modeler.js.map