@arizeai/phoenix-evals/dist/src/code/createFBetaEvaluator.js

A library for running evaluations for AI use cases

"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
exports.createFBetaEvaluator = createFBetaEvaluator;
const classificationMetrics_1 = require("./classificationMetrics");
const createClassificationMetricEvaluator_1 = require("./createClassificationMetricEvaluator");
/**
 * Creates a code evaluator that computes the F-beta score: the weighted
 * harmonic mean of precision and recall, where `beta` controls how much more
 * weight recall gets relative to precision (`beta = 1` is the standard F1
 * score).
 *
 * Supports binary classification (via `positiveLabel`, or auto-detected when
 * `average` is at its default `"macro"` and labels are the numeric set
 * `{0, 1}`) and multi-class classification (via the `average` strategy).
 *
 * @example
 * ```typescript
 * const f2 = createFBetaEvaluator({ beta: 2 });
 * const result = await f2.evaluate({
 *   expected: ["cat", "dog", "cat", "bird"],
 *   output: ["cat", "cat", "cat", "bird"],
 * });
 * ```
 */
function createFBetaEvaluator(options = {}) {
    const { beta = 1 } = options;
    const suffix = (0, classificationMetrics_1.getAverageMetricNameSuffix)(options);
    return (0, createClassificationMetricEvaluator_1.createClassificationMetricEvaluator)(`${(0, classificationMetrics_1.formatBetaForMetricName)(beta)}${suffix}`, "fScore", options);
}
//# sourceMappingURL=createFBetaEvaluator.js.map