🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.
TypeScript analyticsautogenevaluationlangchainlarge-language-modelsllama-indexllmllm-evaluationllm-observabilityllmopsmonitoringobservabilityopen-sourceopenaiplaygroundprompt-engineeringprompt-managementself-hostedycombinator
import { describe, expect, it } from "vitest";
import {
DEFAULT_MAX_RENDERED_SERIES,
prepareVisibleSeries,
} from "@/src/features/widgets/chart-library/prepareVisibleSeries";
import { type DataPoint } from "@/src/features/widgets/chart-library/chart-props";
/** Builds one bucket per value for a dimension, so its summary == sum(values). */
const series = (dimension: string, values: number[]): DataPoint[] =>
values.map((metric, i) => ({
time_dimension: `2026-01-0${i + 1}T00:00:00Z`,
dimension,
metric,
}));
describe("prepareVisibleSeries", () => {
it("returns every series untouched (and in original order) below the cap", () => {
const dimensions = ["b", "a", "c"];
const data = [
...series("b", [1]),
...series("a", [100]),
...series("c", [10]),
];
const result = prepareVisibleSeries(data, dimensions, 25);
expect(result).toEqual({ visible: ["b", "a", "c"], total: 3, hidden: 0 });
});
it("keeps the highest-magnitude series and reports the dropped count", () => {
const dimensions = ["small", "big", "medium"];
const data = [
...series("small", [1, 1]),
...series("big", [50, 50]),
...series("medium", [10, 10]),
];
const result = prepareVisibleSeries(data, dimensions, 2);
expect(result.visible).toEqual(["big", "medium"]);
expect(result.total).toBe(3);
expect(result.hidden).toBe(1);
});
it("breaks ties by name for a deterministic selection", () => {
const dimensions = ["delta", "alpha", "charlie", "bravo"];
const data = [
...series("delta", [5]),
...series("alpha", [5]),
...series("charlie", [5]),
...series("bravo", [5]),
];
const result = prepareVisibleSeries(data, dimensions, 2);
// All equal magnitude -> alphabetical.
expect(result.visible).toEqual(["alpha", "bravo"]);
});
it("ranks series with no finite data last", () => {
const dimensions = ["empty", "tiny", "huge"];
const data = [
// 'empty' has only non-finite metrics -> null summary -> ranked last.
{
time_dimension: "2026-01-01T00:00:00Z",
dimension: "empty",
metric: NaN,
},
...series("tiny", [1]),
...series("huge", [999]),
];
const result = prepareVisibleSeries(data, dimensions, 2);
expect(result.visible).toEqual(["huge", "tiny"]);
expect(result.hidden).toBe(1);
});
it("treats a real 0-sum series as data (kept over a no-data series)", () => {
const dimensions = ["zero", "nodata"];
const data = [
...series("zero", [0, 0]),
{
time_dimension: "2026-01-01T00:00:00Z",
dimension: "nodata",
metric: NaN,
},
];
const result = prepareVisibleSeries(data, dimensions, 1);
expect(result.visible).toEqual(["zero"]);
});
it("defaults to the shared render cap", () => {
const dimensions = Array.from(
{ length: DEFAULT_MAX_RENDERED_SERIES + 30 },
(_, i) => `series-${i}`,
);
const data = dimensions.flatMap((d, i) => series(d, [i]));
const result = prepareVisibleSeries(data, dimensions);
expect(result.visible).toHaveLength(DEFAULT_MAX_RENDERED_SERIES);
expect(result.hidden).toBe(30);
expect(result.total).toBe(DEFAULT_MAX_RENDERED_SERIES + 30);
});
});