zon-format/dist/evals/datasets.js

ZON: The most token-efficient serialization format for LLMs - beats CSV, TOON, JSON, and all competitors

"use strict";
/**
 * Dataset Management for ZON Evaluations
 */
Object.defineProperty(exports, "__esModule", { value: true });
exports.globalRegistry = exports.DatasetRegistry = void 0;
exports.createDataset = createDataset;
/**
 * Registry for managing evaluation datasets
 */
class DatasetRegistry {
    constructor() {
        this.datasets = new Map();
    }
    /**
     * Register a dataset
     */
    register(dataset) {
        const key = `${dataset.id}@${dataset.version}`;
        this.datasets.set(key, dataset);
    }
    /**
     * Get a dataset by ID and optional version
     */
    get(id, version) {
        if (version) {
            return this.datasets.get(`${id}@${version}`) || null;
        }
        const matching = Array.from(this.datasets.values())
            .filter(d => d.id === id)
            .sort((a, b) => this.compareVersions(b.version, a.version));
        return matching[0] || null;
    }
    /**
     * Get all datasets with a specific tag
     */
    getByTag(tag) {
        return Array.from(this.datasets.values())
            .filter(d => d.tags.includes(tag));
    }
    /**
     * Get all golden (baseline) datasets
     */
    getGolden() {
        return this.getByTag('golden');
    }
    /**
     * List all registered datasets
     */
    list() {
        return Array.from(this.datasets.values());
    }
    /**
     * Remove a dataset
     */
    remove(id, version) {
        return this.datasets.delete(`${id}@${version}`);
    }
    /**
     * Clear all datasets
     */
    clear() {
        this.datasets.clear();
    }
    /**
     * Compare semantic versions
     */
    compareVersions(v1, v2) {
        const parts1 = v1.split('.').map(Number);
        const parts2 = v2.split('.').map(Number);
        for (let i = 0; i < Math.max(parts1.length, parts2.length); i++) {
            const p1 = parts1[i] || 0;
            const p2 = parts2[i] || 0;
            if (p1 > p2)
                return 1;
            if (p1 < p2)
                return -1;
        }
        return 0;
    }
}
exports.DatasetRegistry = DatasetRegistry;
/**
 * Global dataset registry
 */
exports.globalRegistry = new DatasetRegistry();
/**
 * Helper to create a dataset
 */
function createDataset(id, version, name, data, questions, options) {
    return {
        id,
        version,
        name,
        data,
        questions,
        schema: options === null || options === void 0 ? void 0 : options.schema,
        tags: (options === null || options === void 0 ? void 0 : options.tags) || []
    };
}