@astermind/astermind-synthetic-data/dist/encoders/StringEncoder.d.ts

OmegaSynth - Label-Conditioned Synthetic Data Generator for AsterMind ELM/KELM Pipelines

/**
 * StringEncoder - Encodes strings to vectors and decodes back
 * Compatible with ELM/KELM pipelines
 */
import { CharVocab } from './CharVocab';
export interface StringEncoderConfig {
    maxLength: number;
    charSet?: string;
    useOneHot?: boolean;
}
export declare class StringEncoder {
    private vocab;
    private config;
    constructor(config: StringEncoderConfig);
    /**
     * Build vocabulary from training samples
     */
    buildVocab(samples: string[]): void;
    /**
     * Encode a string to a vector
     * @param str String to encode
     * @returns Encoded vector (either indices or one-hot)
     */
    encode(str: string): number[];
    /**
     * Decode a vector back to a string
     * @param vector Encoded vector
     * @returns Decoded string
     */
    decode(vector: number[]): string;
    /**
     * Encode multiple strings
     */
    encodeBatch(strings: string[]): number[][];
    /**
     * Decode multiple vectors
     */
    decodeBatch(vectors: number[][]): string[];
    /**
     * Get the output vector size
     */
    getVectorSize(): number;
    /**
     * Get vocabulary size
     */
    getVocabSize(): number;
    /**
     * Get vocabulary
     */
    getVocab(): CharVocab;
}