allprofanity 2.1.1 → 2.2.0

package/dist/algos/bloom-filter.js

@@ -0,0 +1,208 @@
+/**
+ * Bloom Filter implementation for efficient set membership testing
+ */
+export class BloomFilter {
+    constructor(expectedItems, falsePositiveRate = 0.01) {
+        this.itemCount = 0;
+        // Calculate optimal size and hash count
+        this.size = this.calculateOptimalSize(expectedItems, falsePositiveRate);
+        this.hashCount = this.calculateOptimalHashCount(this.size, expectedItems);
+        this.bitArray = new Uint8Array(Math.ceil(this.size / 8));
+    }
+    /**
+     * Calculate optimal bit array size
+     */
+    calculateOptimalSize(n, p) {
+        return Math.ceil((-n * Math.log(p)) / Math.log(2) ** 2);
+    }
+    /**
+     * Calculate optimal number of hash functions
+     */
+    calculateOptimalHashCount(m, n) {
+        return Math.ceil((m / n) * Math.log(2));
+    }
+    /**
+     * Hash function 1 (FNV-1a variant)
+     */
+    hash1(item) {
+        let hash = 2166136261;
+        for (let i = 0; i < item.length; i++) {
+            hash ^= item.charCodeAt(i);
+            hash *= 16777619;
+        }
+        return Math.abs(hash) % this.size;
+    }
+    /**
+     * Hash function 2 (djb2 variant)
+     */
+    hash2(item) {
+        let hash = 5381;
+        for (let i = 0; i < item.length; i++) {
+            hash = (hash << 5) + hash + item.charCodeAt(i);
+        }
+        return Math.abs(hash) % this.size;
+    }
+    /**
+     * Generate k hash values for an item using double hashing
+     */
+    getHashes(item) {
+        const hash1 = this.hash1(item);
+        const hash2 = this.hash2(item);
+        const hashes = [];
+        for (let i = 0; i < this.hashCount; i++) {
+            const hash = (hash1 + i * hash2) % this.size;
+            hashes.push(Math.abs(hash));
+        }
+        return hashes;
+    }
+    /**
+     * Set a bit in the bit array
+     */
+    setBit(index) {
+        const byteIndex = Math.floor(index / 8);
+        const bitIndex = index % 8;
+        this.bitArray[byteIndex] |= 1 << bitIndex;
+    }
+    /**
+     * Get a bit from the bit array
+     */
+    getBit(index) {
+        const byteIndex = Math.floor(index / 8);
+        const bitIndex = index % 8;
+        return (this.bitArray[byteIndex] & (1 << bitIndex)) !== 0;
+    }
+    /**
+     * Add an item to the bloom filter
+     */
+    add(item) {
+        const hashes = this.getHashes(item);
+        for (const hash of hashes) {
+            this.setBit(hash);
+        }
+        this.itemCount++;
+    }
+    /**
+     * Add multiple items to the bloom filter
+     */
+    addAll(items) {
+        for (const item of items) {
+            this.add(item);
+        }
+    }
+    /**
+     * Test if an item might be in the set
+     */
+    mightContain(item) {
+        const hashes = this.getHashes(item);
+        for (const hash of hashes) {
+            if (!this.getBit(hash)) {
+                return false;
+            }
+        }
+        return true;
+    }
+    /**
+     * Test multiple items at once
+     */
+    mightContainAny(items) {
+        return items.some((item) => this.mightContain(item));
+    }
+    /**
+     * Filter items that might be in the set
+     */
+    filter(items) {
+        return items.filter((item) => this.mightContain(item));
+    }
+    /**
+     * Clear the bloom filter
+     */
+    clear() {
+        this.bitArray.fill(0);
+        this.itemCount = 0;
+    }
+    /**
+     * Get current false positive probability
+     */
+    getCurrentFalsePositiveRate() {
+        const ratio = this.itemCount / this.size;
+        return Math.pow(1 - Math.exp(-this.hashCount * ratio), this.hashCount);
+    }
+    /**
+     * Get bloom filter statistics
+     */
+    getStats() {
+        let bitsSet = 0;
+        for (let i = 0; i < this.size; i++) {
+            if (this.getBit(i)) {
+                bitsSet++;
+            }
+        }
+        const loadFactor = bitsSet / this.size;
+        const estimatedFalsePositiveRate = Math.pow(loadFactor, this.hashCount);
+        return {
+            size: this.size,
+            hashCount: this.hashCount,
+            itemCount: this.itemCount,
+            bitsSet,
+            loadFactor,
+            estimatedFalsePositiveRate,
+        };
+    }
+    /**
+     * Serialize bloom filter to JSON
+     */
+    toJSON() {
+        return {
+            size: this.size,
+            hashCount: this.hashCount,
+            itemCount: this.itemCount,
+            bitArray: Array.from(this.bitArray),
+        };
+    }
+    /**
+     * Deserialize bloom filter from JSON
+     */
+    static fromJSON(data) {
+        const filter = Object.create(BloomFilter.prototype);
+        filter.size = data.size;
+        filter.hashCount = data.hashCount;
+        filter.itemCount = data.itemCount;
+        filter.bitArray = new Uint8Array(data.bitArray);
+        return filter;
+    }
+    /**
+     * Union operation with another bloom filter
+     */
+    union(other) {
+        if (this.size !== other.size || this.hashCount !== other.hashCount) {
+            throw new Error("Bloom filters must have same size and hash count for union operation");
+        }
+        const result = new BloomFilter(1, 0.01);
+        result.size = this.size;
+        result.hashCount = this.hashCount;
+        result.bitArray = new Uint8Array(this.bitArray.length);
+        result.itemCount = this.itemCount + other.itemCount;
+        for (let i = 0; i < this.bitArray.length; i++) {
+            result.bitArray[i] = this.bitArray[i] | other.bitArray[i];
+        }
+        return result;
+    }
+    /**
+     * Intersection operation with another bloom filter
+     */
+    intersect(other) {
+        if (this.size !== other.size || this.hashCount !== other.hashCount) {
+            throw new Error("Bloom filters must have same size and hash count for intersection operation");
+        }
+        const result = new BloomFilter(1, 0.01);
+        result.size = this.size;
+        result.hashCount = this.hashCount;
+        result.bitArray = new Uint8Array(this.bitArray.length);
+        result.itemCount = Math.min(this.itemCount, other.itemCount);
+        for (let i = 0; i < this.bitArray.length; i++) {
+            result.bitArray[i] = this.bitArray[i] & other.bitArray[i];
+        }
+        return result;
+    }
+}
+//# sourceMappingURL=bloom-filter.js.map

package/dist/algos/bloom-filter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"bloom-filter.js","sourceRoot":"","sources":["../../src/algos/bloom-filter.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,MAAM,OAAO,WAAW;IAMtB,YAAY,aAAqB,EAAE,oBAA4B,IAAI;QAF3D,cAAS,GAAW,CAAC,CAAC;QAG5B,wCAAwC;QACxC,IAAI,CAAC,IAAI,GAAG,IAAI,CAAC,oBAAoB,CAAC,aAAa,EAAE,iBAAiB,CAAC,CAAC;QACxE,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC,yBAAyB,CAAC,IAAI,CAAC,IAAI,EAAE,aAAa,CAAC,CAAC;QAC1E,IAAI,CAAC,QAAQ,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC;IAC3D,CAAC;IAED;;OAEG;IACK,oBAAoB,CAAC,CAAS,EAAE,CAAS;QAC/C,OAAO,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IAC1D,CAAC;IAED;;OAEG;IACK,yBAAyB,CAAC,CAAS,EAAE,CAAS;QACpD,OAAO,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1C,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,IAAY;QACxB,IAAI,IAAI,GAAG,UAAU,CAAC;QACtB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YACpC,IAAI,IAAI,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;YAC3B,IAAI,IAAI,QAAQ,CAAC;SAClB;QACD,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC;IACpC,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,IAAY;QACxB,IAAI,IAAI,GAAG,IAAI,CAAC;QAChB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YACpC,IAAI,GAAG,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,IAAI,GAAG,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;SAChD;QACD,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC;IACpC,CAAC;IAED;;OAEG;IACK,SAAS,CAAC,IAAY;QAC5B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC/B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC/B,MAAM,MAAM,GAAa,EAAE,CAAC;QAE5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,SAAS,EAAE,CAAC,EAAE,EAAE;YACvC,MAAM,IAAI,GAAG,CAAC,KAAK,GAAG,CAAC,GAAG,KAAK,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC;YAC7C,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC;SAC7B;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;OAEG;IACK,MAAM,CAAC,KAAa;QAC1B,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;QACxC,MAAM,QAAQ,GAAG,KAAK,GAAG,CAAC,CAAC;QAC3B,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,QAAQ,CAAC;IAC5C,CAAC;IAED;;OAEG;IACK,MAAM,CAAC,KAAa;QAC1B,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;QACxC,MAAM,QAAQ,GAAG,KAAK,GAAG,CAAC,CAAC;QAC3B,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,IAAI,QAAQ,CAAC,CAAC,KAAK,CAAC,CAAC;IAC5D,CAAC;IAED;;OAEG;IACH,GAAG,CAAC,IAAY;QACd,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;QACpC,KAAK,MAAM,IAAI,IAAI,MAAM,EAAE;YACzB,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;SACnB;QACD,IAAI,CAAC,SAAS,EAAE,CAAC;IACnB,CAAC;IAED;;OAEG;IACH,MAAM,CAAC,KAAe;QACpB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE;YACxB,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;SAChB;IACH,CAAC;IAED;;OAEG;IACH,YAAY,CAAC,IAAY;QACvB,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;QACpC,KAAK,MAAM,IAAI,IAAI,MAAM,EAAE;YACzB,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE;gBACtB,OAAO,KAAK,CAAC;aACd;SACF;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;OAEG;IACH,eAAe,CAAC,KAAe;QAC7B,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC;IACvD,CAAC;IAED;;OAEG;IACH,MAAM,CAAC,KAAe;QACpB,OAAO,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC;IACzD,CAAC;IAED;;OAEG;IACH,KAAK;QACH,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACtB,IAAI,CAAC,SAAS,GAAG,CAAC,CAAC;IACrB,CAAC;IAED;;OAEG;IACH,2BAA2B;QACzB,MAAM,KAAK,GAAG,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC,IAAI,CAAC;QACzC,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,SAAS,GAAG,KAAK,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC;IACzE,CAAC;IAED;;OAEG;IACH,QAAQ;QAQN,IAAI,OAAO,GAAG,CAAC,CAAC;QAChB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,EAAE,EAAE;YAClC,IAAI,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE;gBAClB,OAAO,EAAE,CAAC;aACX;SACF;QAED,MAAM,UAAU,GAAG,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC;QACvC,MAAM,0BAA0B,GAAG,IAAI,CAAC,GAAG,CAAC,UAAU,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC;QAExE,OAAO;YACL,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,OAAO;YACP,UAAU;YACV,0BAA0B;SAC3B,CAAC;IACJ,CAAC;IAED;;OAEG;IACH,MAAM;QAMJ,OAAO;YACL,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,QAAQ,EAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC;SACpC,CAAC;IACJ,CAAC;IAED;;OAEG;IACH,MAAM,CAAC,QAAQ,CAAC,IAKf;QACC,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,WAAW,CAAC,SAAS,CAAC,CAAC;QACpD,MAAM,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;QACxB,MAAM,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC;QAClC,MAAM,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC;QAClC,MAAM,CAAC,QAAQ,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAChD,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,KAAkB;QACtB,IAAI,IAAI,CAAC,IAAI,KAAK,KAAK,CAAC,IAAI,IAAI,IAAI,CAAC,SAAS,KAAK,KAAK,CAAC,SAAS,EAAE;YAClE,MAAM,IAAI,KAAK,CACb,sEAAsE,CACvE,CAAC;SACH;QAED,MAAM,MAAM,GAAG,IAAI,WAAW,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;QACxC,MAAM,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;QACxB,MAAM,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC;QAClC,MAAM,CAAC,QAAQ,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;QACvD,MAAM,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,GAAG,KAAK,CAAC,SAAS,CAAC;QAEpD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YAC7C,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC;SAC3D;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;OAEG;IACH,SAAS,CAAC,KAAkB;QAC1B,IAAI,IAAI,CAAC,IAAI,KAAK,KAAK,CAAC,IAAI,IAAI,IAAI,CAAC,SAAS,KAAK,KAAK,CAAC,SAAS,EAAE;YAClE,MAAM,IAAI,KAAK,CACb,6EAA6E,CAC9E,CAAC;SACH;QAED,MAAM,MAAM,GAAG,IAAI,WAAW,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;QACxC,MAAM,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;QACxB,MAAM,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC;QAClC,MAAM,CAAC,QAAQ,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;QACvD,MAAM,CAAC,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,EAAE,KAAK,CAAC,SAAS,CAAC,CAAC;QAE7D,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YAC7C,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC;SAC3D;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;CACF"}

package/dist/algos/context-patterns.d.ts

@@ -0,0 +1,88 @@
+/**
+ * Universal context patterns for multi-language profanity detection
+ */
+export interface UniversalContextPattern {
+    type: "negation" | "possessive" | "compound" | "proper_noun" | "article" | "quotation" | "medical" | "anatomical";
+    pattern: RegExp;
+    weight: number;
+    languages: string[];
+    description: string;
+    examples: string[];
+}
+export interface ContextRule {
+    pattern: RegExp;
+    action: "reduce_score" | "increase_score" | "whitelist" | "flag";
+    weight: number;
+    priority: number;
+}
+/**
+ * Universal context patterns that work across multiple languages
+ */
+export declare const UNIVERSAL_CONTEXT_PATTERNS: UniversalContextPattern[];
+/**
+ * Language-specific context patterns
+ */
+export declare const LANGUAGE_SPECIFIC_PATTERNS: Record<string, UniversalContextPattern[]>;
+/**
+ * Context rule generator
+ */
+export declare class ContextPatternMatcher {
+    private patterns;
+    private languagePatterns;
+    constructor(languages?: string[]);
+    /**
+     * Generate context rules for a specific word
+     */
+    generateRules(word: string, languages?: string[]): ContextRule[];
+    /**
+     * Get priority for pattern type
+     */
+    private getPriority;
+    /**
+     * Escape regex special characters
+     */
+    private escapeRegex;
+    /**
+     * Add custom pattern
+     */
+    addPattern(pattern: UniversalContextPattern): void;
+    /**
+     * Add language-specific pattern
+     */
+    addLanguagePattern(language: string, pattern: UniversalContextPattern): void;
+    /**
+     * Get all patterns for debugging
+     */
+    getAllPatterns(): {
+        universal: UniversalContextPattern[];
+        languageSpecific: Map<string, UniversalContextPattern[]>;
+    };
+}
+/**
+ * Context analyzer for scoring matches
+ */
+export declare class ContextAnalyzer {
+    private patternMatcher;
+    private contextWindow;
+    constructor(languages?: string[]);
+    /**
+     * Analyze context around a potential profanity match
+     */
+    analyzeContext(text: string, matchStart: number, matchEnd: number, word: string): {
+        score: number;
+        confidence: "high" | "medium" | "low";
+        appliedRules: Array<{
+            rule: ContextRule;
+            matched: boolean;
+        }>;
+        context: string;
+    };
+    /**
+     * Set context window size
+     */
+    setContextWindow(size: number): void;
+    /**
+     * Add custom pattern to the analyzer
+     */
+    addCustomPattern(pattern: UniversalContextPattern): void;
+}

package/dist/algos/context-patterns.js

@@ -0,0 +1,298 @@
+/**
+ * Universal context patterns for multi-language profanity detection
+ */
+/**
+ * Universal context patterns that work across multiple languages
+ */
+export const UNIVERSAL_CONTEXT_PATTERNS = [
+    // Negation patterns
+    {
+        type: "negation",
+        pattern: /\b(not|don't|won't|can't|never|ne|pas|nicht|no|नहीं|不|non|niente|нет|nie)\b.{0,30}PROFANE_WORD/i,
+        weight: 0.2,
+        languages: ["*"],
+        description: "Negation words that reduce profanity likelihood",
+        examples: ["not bad", "don't call me that", "never say that"],
+    },
+    // Possessive patterns
+    {
+        type: "possessive",
+        pattern: /\b\w+(['s]|du|de|का|की|के|の|del|della|от|od)\s+PROFANE_WORD\b/i,
+        weight: 0.4,
+        languages: ["*"],
+        description: "Possessive constructions that may be innocent",
+        examples: ["dog's mouth", "cat's ass", "bird's ass"],
+    },
+    // Article patterns
+    {
+        type: "article",
+        pattern: /\b(the|a|an|le|la|les|un|une|der|die|das|ein|eine|el|la|los|las|il|lo|gli|le)\s+PROFANE_WORD\b/i,
+        weight: 0.6,
+        languages: ["*"],
+        description: "Articles that may indicate neutral reference",
+        examples: ["the ass of the donkey", "a hell of a time"],
+    },
+    // Compound word patterns
+    {
+        type: "compound",
+        pattern: /\b(smart|silly|cute|funny|little|big|old|new|good|bad|nice|sweet)\s*[-]?\s*PROFANE_WORD\b/i,
+        weight: 0.5,
+        languages: ["*"],
+        description: "Adjective-noun compounds that may be innocent",
+        examples: ["smart-ass", "silly ass", "cute little ass"],
+    },
+    // Proper noun patterns
+    {
+        type: "proper_noun",
+        pattern: /\b[A-Z][a-z]+\s+PROFANE_WORD\b/,
+        weight: 0.3,
+        languages: ["en", "fr", "de", "es", "it"],
+        description: "Proper nouns followed by potential profanity",
+        examples: ["Hell Michigan", "Ass River"],
+    },
+    // Quotation patterns
+    {
+        type: "quotation",
+        pattern: /["'«»„"‚'].*PROFANE_WORD.*["'«»„"‚']/i,
+        weight: 0.7,
+        languages: ["*"],
+        description: "Quoted text which may be reporting speech",
+        examples: ['"Don\'t be an ass"', "'What the hell'"],
+    },
+    // Medical/anatomical context
+    {
+        type: "medical",
+        pattern: /\b(medical|anatomy|doctor|hospital|clinic|patient|diagnosis|treatment|surgical|clinical)\b.{0,50}PROFANE_WORD/i,
+        weight: 0.1,
+        languages: ["*"],
+        description: "Medical contexts where anatomical terms are appropriate",
+        examples: [
+            "medical examination of the ass",
+            "doctor checked the damn thing",
+        ],
+    },
+    // Anatomical context
+    {
+        type: "anatomical",
+        pattern: /\b(body|part|muscle|bone|skin|tissue|organ|limb|extremity)\b.{0,30}PROFANE_WORD/i,
+        weight: 0.3,
+        languages: ["*"],
+        description: "Anatomical contexts for body parts",
+        examples: ["body part called ass", "muscle in the ass"],
+    },
+];
+/**
+ * Language-specific context patterns
+ */
+export const LANGUAGE_SPECIFIC_PATTERNS = {
+    en: [
+        {
+            type: "compound",
+            pattern: /\b(jack|dumb|smart|bad|kick)\s*[-]?\s*PROFANE_WORD\b/i,
+            weight: 0.4,
+            languages: ["en"],
+            description: "English-specific compound patterns",
+            examples: ["jackass", "dumbass", "badass"],
+        },
+    ],
+    fr: [
+        {
+            type: "negation",
+            pattern: /\b(ne|n'|pas|point|jamais|rien|personne)\b.{0,30}PROFANE_WORD/i,
+            weight: 0.2,
+            languages: ["fr"],
+            description: "French negation patterns",
+            examples: ["ne pas dire", "jamais ça"],
+        },
+    ],
+    de: [
+        {
+            type: "compound",
+            pattern: /\bPROFANE_WORD(kopf|zeug|ding|sache)\b/i,
+            weight: 0.5,
+            languages: ["de"],
+            description: "German compound word patterns",
+            examples: ["Scheißzeug", "Arschloch"],
+        },
+    ],
+    es: [
+        {
+            type: "possessive",
+            pattern: /\b(el|la|los|las)\s+PROFANE_WORD\s+(de|del|de la)\b/i,
+            weight: 0.4,
+            languages: ["es"],
+            description: "Spanish possessive patterns",
+            examples: ["el culo de la mesa"],
+        },
+    ],
+};
+/**
+ * Context rule generator
+ */
+export class ContextPatternMatcher {
+    constructor(languages = ["en"]) {
+        this.patterns = [...UNIVERSAL_CONTEXT_PATTERNS];
+        this.languagePatterns = new Map();
+        // Load language-specific patterns
+        for (const lang of languages) {
+            if (LANGUAGE_SPECIFIC_PATTERNS[lang]) {
+                this.languagePatterns.set(lang, LANGUAGE_SPECIFIC_PATTERNS[lang]);
+            }
+        }
+    }
+    /**
+     * Generate context rules for a specific word
+     */
+    generateRules(word, languages = ["en"]) {
+        const rules = [];
+        const allPatterns = [...this.patterns];
+        // Add language-specific patterns
+        for (const lang of languages) {
+            const langPatterns = this.languagePatterns.get(lang) || [];
+            allPatterns.push(...langPatterns);
+        }
+        for (const pattern of allPatterns) {
+            // Skip if pattern doesn't apply to any of the specified languages
+            if (!pattern.languages.includes("*") &&
+                !pattern.languages.some((lang) => languages.includes(lang))) {
+                continue;
+            }
+            // Replace PROFANE_WORD placeholder with actual word
+            const regexSource = pattern.pattern.source.replace("PROFANE_WORD", this.escapeRegex(word));
+            const regex = new RegExp(regexSource, pattern.pattern.flags);
+            let action;
+            if (pattern.weight < 0.3) {
+                action = "reduce_score";
+            }
+            else if (pattern.weight > 0.8) {
+                action = "increase_score";
+            }
+            else {
+                action = "reduce_score";
+            }
+            rules.push({
+                pattern: regex,
+                action,
+                weight: pattern.weight,
+                priority: this.getPriority(pattern.type),
+            });
+        }
+        return rules.sort((a, b) => a.priority - b.priority);
+    }
+    /**
+     * Get priority for pattern type
+     */
+    getPriority(type) {
+        const priorities = {
+            medical: 1,
+            anatomical: 2,
+            negation: 3,
+            quotation: 4,
+            proper_noun: 5,
+            possessive: 6,
+            article: 7,
+            compound: 8,
+        };
+        return priorities[type] || 9;
+    }
+    /**
+     * Escape regex special characters
+     */
+    escapeRegex(str) {
+        return str.replace(/[\\^$.*+?()[\]{}|]/g, "\\$&");
+    }
+    /**
+     * Add custom pattern
+     */
+    addPattern(pattern) {
+        this.patterns.push(pattern);
+    }
+    /**
+     * Add language-specific pattern
+     */
+    addLanguagePattern(language, pattern) {
+        if (!this.languagePatterns.has(language)) {
+            this.languagePatterns.set(language, []);
+        }
+        this.languagePatterns.get(language).push(pattern);
+    }
+    /**
+     * Get all patterns for debugging
+     */
+    getAllPatterns() {
+        return {
+            universal: [...this.patterns],
+            languageSpecific: new Map(this.languagePatterns),
+        };
+    }
+}
+/**
+ * Context analyzer for scoring matches
+ */
+export class ContextAnalyzer {
+    constructor(languages = ["en"]) {
+        this.contextWindow = 50; // Characters before and after the match
+        this.patternMatcher = new ContextPatternMatcher(languages);
+    }
+    /**
+     * Analyze context around a potential profanity match
+     */
+    analyzeContext(text, matchStart, matchEnd, word) {
+        // Extract context window
+        const contextStart = Math.max(0, matchStart - this.contextWindow);
+        const contextEnd = Math.min(text.length, matchEnd + this.contextWindow);
+        const context = text.substring(contextStart, contextEnd);
+        // Get rules for this word
+        const rules = this.patternMatcher.generateRules(word);
+        let score = 1.0; // Start with full profanity score
+        const appliedRules = [];
+        // Apply context rules
+        for (const rule of rules) {
+            const matched = rule.pattern.test(context);
+            appliedRules.push({ rule, matched });
+            if (matched) {
+                if (rule.action === "reduce_score") {
+                    score *= rule.weight;
+                }
+                else if (rule.action === "increase_score") {
+                    score *= 2 - rule.weight; // Increase score
+                }
+                else if (rule.action === "whitelist") {
+                    score = 0; // Complete whitelist
+                    break;
+                }
+            }
+        }
+        // Determine confidence based on number of matching rules
+        const matchingRules = appliedRules.filter((ar) => ar.matched).length;
+        let confidence;
+        if (matchingRules === 0) {
+            confidence = "high"; // No context rules matched, likely profanity
+        }
+        else if (matchingRules <= 2) {
+            confidence = "medium";
+        }
+        else {
+            confidence = "low"; // Many context rules matched, likely innocent
+        }
+        return {
+            score: Math.max(0, Math.min(1, score)),
+            confidence,
+            appliedRules,
+            context,
+        };
+    }
+    /**
+     * Set context window size
+     */
+    setContextWindow(size) {
+        this.contextWindow = Math.max(10, Math.min(200, size));
+    }
+    /**
+     * Add custom pattern to the analyzer
+     */
+    addCustomPattern(pattern) {
+        this.patternMatcher.addPattern(pattern);
+    }
+}
+//# sourceMappingURL=context-patterns.js.map

package/dist/algos/context-patterns.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"context-patterns.js","sourceRoot":"","sources":["../../src/algos/context-patterns.ts"],"names":[],"mappings":"AAAA;;GAEG;AA0BH;;GAEG;AACH,MAAM,CAAC,MAAM,0BAA0B,GAA8B;IACnE,oBAAoB;IACpB;QACE,IAAI,EAAE,UAAU;QAChB,OAAO,EACL,iGAAiG;QACnG,MAAM,EAAE,GAAG;QACX,SAAS,EAAE,CAAC,GAAG,CAAC;QAChB,WAAW,EAAE,iDAAiD;QAC9D,QAAQ,EAAE,CAAC,SAAS,EAAE,oBAAoB,EAAE,gBAAgB,CAAC;KAC9D;IAED,sBAAsB;IACtB;QACE,IAAI,EAAE,YAAY;QAClB,OAAO,EAAE,gEAAgE;QACzE,MAAM,EAAE,GAAG;QACX,SAAS,EAAE,CAAC,GAAG,CAAC;QAChB,WAAW,EAAE,+CAA+C;QAC5D,QAAQ,EAAE,CAAC,aAAa,EAAE,WAAW,EAAE,YAAY,CAAC;KACrD;IAED,mBAAmB;IACnB;QACE,IAAI,EAAE,SAAS;QACf,OAAO,EACL,iGAAiG;QACnG,MAAM,EAAE,GAAG;QACX,SAAS,EAAE,CAAC,GAAG,CAAC;QAChB,WAAW,EAAE,8CAA8C;QAC3D,QAAQ,EAAE,CAAC,uBAAuB,EAAE,kBAAkB,CAAC;KACxD;IAED,yBAAyB;IACzB;QACE,IAAI,EAAE,UAAU;QAChB,OAAO,EACL,4FAA4F;QAC9F,MAAM,EAAE,GAAG;QACX,SAAS,EAAE,CAAC,GAAG,CAAC;QAChB,WAAW,EAAE,+CAA+C;QAC5D,QAAQ,EAAE,CAAC,WAAW,EAAE,WAAW,EAAE,iBAAiB,CAAC;KACxD;IAED,uBAAuB;IACvB;QACE,IAAI,EAAE,aAAa;QACnB,OAAO,EAAE,gCAAgC;QACzC,MAAM,EAAE,GAAG;QACX,SAAS,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC;QACzC,WAAW,EAAE,8CAA8C;QAC3D,QAAQ,EAAE,CAAC,eAAe,EAAE,WAAW,CAAC;KACzC;IAED,qBAAqB;IACrB;QACE,IAAI,EAAE,WAAW;QACjB,OAAO,EAAE,uCAAuC;QAChD,MAAM,EAAE,GAAG;QACX,SAAS,EAAE,CAAC,GAAG,CAAC;QAChB,WAAW,EAAE,2CAA2C;QACxD,QAAQ,EAAE,CAAC,oBAAoB,EAAE,iBAAiB,CAAC;KACpD;IAED,6BAA6B;IAC7B;QACE,IAAI,EAAE,SAAS;QACf,OAAO,EACL,gHAAgH;QAClH,MAAM,EAAE,GAAG;QACX,SAAS,EAAE,CAAC,GAAG,CAAC;QAChB,WAAW,EAAE,yDAAyD;QACtE,QAAQ,EAAE;YACR,gCAAgC;YAChC,+BAA+B;SAChC;KACF;IAED,qBAAqB;IACrB;QACE,IAAI,EAAE,YAAY;QAClB,OAAO,EACL,kFAAkF;QACpF,MAAM,EAAE,GAAG;QACX,SAAS,EAAE,CAAC,GAAG,CAAC;QAChB,WAAW,EAAE,oCAAoC;QACjD,QAAQ,EAAE,CAAC,sBAAsB,EAAE,mBAAmB,CAAC;KACxD;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,CAAC,MAAM,0BAA0B,GAGnC;IACF,EAAE,EAAE;QACF;YACE,IAAI,EAAE,UAAU;YAChB,OAAO,EAAE,uDAAuD;YAChE,MAAM,EAAE,GAAG;YACX,SAAS,EAAE,CAAC,IAAI,CAAC;YACjB,WAAW,EAAE,oCAAoC;YACjD,QAAQ,EAAE,CAAC,SAAS,EAAE,SAAS,EAAE,QAAQ,CAAC;SAC3C;KACF;IAED,EAAE,EAAE;QACF;YACE,IAAI,EAAE,UAAU;YAChB,OAAO,EAAE,gEAAgE;YACzE,MAAM,EAAE,GAAG;YACX,SAAS,EAAE,CAAC,IAAI,CAAC;YACjB,WAAW,EAAE,0BAA0B;YACvC,QAAQ,EAAE,CAAC,aAAa,EAAE,WAAW,CAAC;SACvC;KACF;IAED,EAAE,EAAE;QACF;YACE,IAAI,EAAE,UAAU;YAChB,OAAO,EAAE,yCAAyC;YAClD,MAAM,EAAE,GAAG;YACX,SAAS,EAAE,CAAC,IAAI,CAAC;YACjB,WAAW,EAAE,+BAA+B;YAC5C,QAAQ,EAAE,CAAC,YAAY,EAAE,WAAW,CAAC;SACtC;KACF;IAED,EAAE,EAAE;QACF;YACE,IAAI,EAAE,YAAY;YAClB,OAAO,EAAE,sDAAsD;YAC/D,MAAM,EAAE,GAAG;YACX,SAAS,EAAE,CAAC,IAAI,CAAC;YACjB,WAAW,EAAE,6BAA6B;YAC1C,QAAQ,EAAE,CAAC,oBAAoB,CAAC;SACjC;KACF;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,qBAAqB;IAIhC,YAAY,YAAsB,CAAC,IAAI,CAAC;QACtC,IAAI,CAAC,QAAQ,GAAG,CAAC,GAAG,0BAA0B,CAAC,CAAC;QAChD,IAAI,CAAC,gBAAgB,GAAG,IAAI,GAAG,EAAE,CAAC;QAElC,kCAAkC;QAClC,KAAK,MAAM,IAAI,IAAI,SAAS,EAAE;YAC5B,IAAI,0BAA0B,CAAC,IAAI,CAAC,EAAE;gBACpC,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,IAAI,EAAE,0BAA0B,CAAC,IAAI,CAAC,CAAC,CAAC;aACnE;SACF;IACH,CAAC;IAED;;OAEG;IACH,aAAa,CAAC,IAAY,EAAE,YAAsB,CAAC,IAAI,CAAC;QACtD,MAAM,KAAK,GAAkB,EAAE,CAAC;QAChC,MAAM,WAAW,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC;QAEvC,iCAAiC;QACjC,KAAK,MAAM,IAAI,IAAI,SAAS,EAAE;YAC5B,MAAM,YAAY,GAAG,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;YAC3D,WAAW,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC;SACnC;QAED,KAAK,MAAM,OAAO,IAAI,WAAW,EAAE;YACjC,kEAAkE;YAClE,IACE,CAAC,OAAO,CAAC,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC;gBAChC,CAAC,OAAO,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,EAC3D;gBACA,SAAS;aACV;YAED,oDAAoD;YACpD,MAAM,WAAW,GAAG,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,OAAO,CAChD,cAAc,EACd,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,CACvB,CAAC;YACF,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,WAAW,EAAE,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;YAE7D,IAAI,MAAgE,CAAC;YACrE,IAAI,OAAO,CAAC,MAAM,GAAG,GAAG,EAAE;gBACxB,MAAM,GAAG,cAAc,CAAC;aACzB;iBAAM,IAAI,OAAO,CAAC,MAAM,GAAG,GAAG,EAAE;gBAC/B,MAAM,GAAG,gBAAgB,CAAC;aAC3B;iBAAM;gBACL,MAAM,GAAG,cAAc,CAAC;aACzB;YAED,KAAK,CAAC,IAAI,CAAC;gBACT,OAAO,EAAE,KAAK;gBACd,MAAM;gBACN,MAAM,EAAE,OAAO,CAAC,MAAM;gBACtB,QAAQ,EAAE,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,IAAI,CAAC;aACzC,CAAC,CAAC;SACJ;QAED,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;IACvD,CAAC;IAED;;OAEG;IACK,WAAW,CAAC,IAAY;QAC9B,MAAM,UAAU,GAA2B;YACzC,OAAO,EAAE,CAAC;YACV,UAAU,EAAE,CAAC;YACb,QAAQ,EAAE,CAAC;YACX,SAAS,EAAE,CAAC;YACZ,WAAW,EAAE,CAAC;YACd,UAAU,EAAE,CAAC;YACb,OAAO,EAAE,CAAC;YACV,QAAQ,EAAE,CAAC;SACZ,CAAC;QACF,OAAO,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC/B,CAAC;IAED;;OAEG;IACK,WAAW,CAAC,GAAW;QAC7B,OAAO,GAAG,CAAC,OAAO,CAAC,qBAAqB,EAAE,MAAM,CAAC,CAAC;IACpD,CAAC;IAED;;OAEG;IACH,UAAU,CAAC,OAAgC;QACzC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IAC9B,CAAC;IAED;;OAEG;IACH,kBAAkB,CAAC,QAAgB,EAAE,OAAgC;QACnE,IAAI,CAAC,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE;YACxC,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;SACzC;QACD,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,QAAQ,CAAE,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACrD,CAAC;IAED;;OAEG;IACH,cAAc;QAIZ,OAAO;YACL,SAAS,EAAE,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC;YAC7B,gBAAgB,EAAE,IAAI,GAAG,CAAC,IAAI,CAAC,gBAAgB,CAAC;SACjD,CAAC;IACJ,CAAC;CACF;AAED;;GAEG;AACH,MAAM,OAAO,eAAe;IAI1B,YAAY,YAAsB,CAAC,IAAI,CAAC;QAFhC,kBAAa,GAAW,EAAE,CAAC,CAAC,wCAAwC;QAG1E,IAAI,CAAC,cAAc,GAAG,IAAI,qBAAqB,CAAC,SAAS,CAAC,CAAC;IAC7D,CAAC;IAED;;OAEG;IACH,cAAc,CACZ,IAAY,EACZ,UAAkB,EAClB,QAAgB,EAChB,IAAY;QAOZ,yBAAyB;QACzB,MAAM,YAAY,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,UAAU,GAAG,IAAI,CAAC,aAAa,CAAC,CAAC;QAClE,MAAM,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,EAAE,QAAQ,GAAG,IAAI,CAAC,aAAa,CAAC,CAAC;QACxE,MAAM,OAAO,GAAG,IAAI,CAAC,SAAS,CAAC,YAAY,EAAE,UAAU,CAAC,CAAC;QAEzD,0BAA0B;QAC1B,MAAM,KAAK,GAAG,IAAI,CAAC,cAAc,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC;QAEtD,IAAI,KAAK,GAAG,GAAG,CAAC,CAAC,kCAAkC;QACnD,MAAM,YAAY,GAAmD,EAAE,CAAC;QAExE,sBAAsB;QACtB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE;YACxB,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YAC3C,YAAY,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,CAAC;YAErC,IAAI,OAAO,EAAE;gBACX,IAAI,IAAI,CAAC,MAAM,KAAK,cAAc,EAAE;oBAClC,KAAK,IAAI,IAAI,CAAC,MAAM,CAAC;iBACtB;qBAAM,IAAI,IAAI,CAAC,MAAM,KAAK,gBAAgB,EAAE;oBAC3C,KAAK,IAAI,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,iBAAiB;iBAC5C;qBAAM,IAAI,IAAI,CAAC,MAAM,KAAK,WAAW,EAAE;oBACtC,KAAK,GAAG,CAAC,CAAC,CAAC,qBAAqB;oBAChC,MAAM;iBACP;aACF;SACF;QAED,yDAAyD;QACzD,MAAM,aAAa,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC;QACrE,IAAI,UAAqC,CAAC;QAE1C,IAAI,aAAa,KAAK,CAAC,EAAE;YACvB,UAAU,GAAG,MAAM,CAAC,CAAC,6CAA6C;SACnE;aAAM,IAAI,aAAa,IAAI,CAAC,EAAE;YAC7B,UAAU,GAAG,QAAQ,CAAC;SACvB;aAAM;YACL,UAAU,GAAG,KAAK,CAAC,CAAC,8CAA8C;SACnE;QAED,OAAO;YACL,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;YACtC,UAAU;YACV,YAAY;YACZ,OAAO;SACR,CAAC;IACJ,CAAC;IAED;;OAEG;IACH,gBAAgB,CAAC,IAAY;QAC3B,IAAI,CAAC,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC,CAAC;IACzD,CAAC;IAED;;OAEG;IACH,gBAAgB,CAAC,OAAgC;QAC/C,IAAI,CAAC,cAAc,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;IAC1C,CAAC;CACF"}