Files
paseo/packages/app/e2e/browser/agent-stream-smoothness.spec.ts
e22ad117fd Stream assistant text at a steady rate instead of in arrival lumps (#3612)
* Stream assistant text at a steady rate instead of in arrival lumps

Assistant text arrived in lumps and was painted that way. The daemon
coalesces one message per 60ms window, and that window carries however
many characters the model produced in it, which swings by an order of
magnitude within a turn. Painting each delta as it landed made the size
of those lumps visible.

Three changes:

- The agent stream coalescer is now leading + trailing. The first delta
  after an idle window flushes synchronously instead of waiting a full
  window, matching the invariant TerminalOutputCoalescer already
  documents. The window stays at 60ms; shrinking it would raise message
  rate on a loop that already contends with terminal frames and relay
  encryption, without making the stream smoother.

- The client reducer queue commits on a frame instead of an arbitrary
  48ms timer, with the timer kept as a ceiling so a hidden tab still
  advances.

- Assistant and reasoning text are revealed at a rate derived from the
  backlog rather than from arrival timing. Arrival changes the target,
  not the rate. The store keeps the full text and only the rendered
  slice is paced, so copy, selection, the chat outline, and scroll
  geometry all stay consistent with what is on screen.

The leading flush means a burst's first chunk lands as its own timeline
row. The projection already merges contiguous same-turn rows and clients
read the projected timeline, so history is unchanged; the server tests
that assert on raw rows were updated for the extra row.

Measured on Expo web against a local dev daemon. Mock bursty-stream
provider, 6s samples, paint-on-arrival vs paced: frames that advanced
the text 55 -> 244, chars-per-frame CV 2.66 -> 1.44, gap p50 100ms ->
17ms. Real Claude Haiku agent over 52s: 84% of frames advance the text,
CV 1.29, gap p50 and p95 both 17ms (one frame at 60Hz).

Adds a bursty-stream mock model with seeded burst sizes, a gated browser
perf spec, and docs/agent-stream-performance.md.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* Address review: safe reveal boundaries and a testable seam

Two findings from the PR review.

Unicode boundaries. Slicing the revealed text at an arbitrary index could
strand a lone surrogate, which paints as a replacement glyph, or split an
emoji ZWJ sequence so a family briefly rendered as its members. Both were
visible for a frame or two mid-stream — the exact flicker the paced reveal
exists to remove. The rendered cut is now pulled back off surrogate pairs,
combining marks, variation selectors, skin tone modifiers, and ZWJ joins.
The reveal counter stays monotonic and only the painted boundary moves, so
a long cluster cannot stall the reveal.

Test placement. The hook test used JSDOM, @testing-library renderHook, and
monkey-patched requestAnimationFrame, all of which docs/testing.md rules
out; it prescribes fixing the missing seam instead. The reveal policy is
now a pure state machine in text-reveal.ts (begin/retarget/advance/
complete/visible) covered by plain unit tests with no environment, and the
hook is reduced to the frame-clock binding. Rendered behavior is left to
the browser spec, which no longer carries frame sampling and statistics in
its body — those moved to e2e/support/helpers/stream-smoothness.ts.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* Measure total painted text, not the tail message

The smoothness sampler read the length of the last assistant-message
element. A real turn emits many assistant messages, so that element keeps
changing identity and its length is not monotonic — handovers read as
resets, and a stream that was plainly moving on screen reported almost no
growth at all. Sampling the total across every assistant message fixes it.

This invalidated the numbers previously recorded for the real-agent run,
so the doc table is replaced with a fresh A/B taken with the corrected
sampler against a Claude Haiku agent: frames that advanced the text 6% ->
87%, chars-per-frame CV 4.11 -> 1.86, gap p50 317ms -> 17ms, p95 383ms ->
17ms. The paced column also reported zero Unicode defects across a stream
containing emoji, which exercises the reveal boundary clamp end to end.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* perf(agent-stream): harden paced text reveal

Use standards-compliant grapheme segmentation when available and preserve full arrivals on older runtimes. Cap reveal commits at 60Hz and measure the real browser rendering path so high-refresh displays stay smooth without rendering at hardware frame rate.

* Hold back trailing fragments split by arrival boundaries

The reveal boundary is not the only place a grapheme gets cut. The
daemon's coalescing window ends a delta wherever the window closes, so a
caught-up reveal can be handed half a cluster and paint it. Measured on a
mixed-script Haiku stream: five frames in ~350 ending in a single regional
indicator, i.e. half a flag on screen for a frame.

While the turn is streaming, a trailing fragment that cannot stand alone —
a dangling high surrogate, a trailing joiner, an odd regional indicator —
now waits for the rest of its cluster. Once the turn ends everything
received is painted, because nothing more is coming. Complements the
grapheme-boundary clamp, which handles the reveal cut; this handles the
arrival cut.

The same stream now measures one such frame in ~400 rather than five. The
remainder is a flush boundary that coincides with the end of an assistant
message, putting the two halves in different messages, which the client
cannot rejoin — that one belongs in the coalescer.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: Mohamed Boudra <boudra.moha@gmail.com>
2026-08-25 23:37:46 +02:00

152 lines
5.5 KiB
TypeScript
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
import { test, expect } from "../support/fixtures";
import { awaitAssistantMessage } from "../support/helpers/agent-stream";
import { startRunningMockAgent } from "../support/helpers/composer";
import {
sampleStreamFrames,
summarizeStreamSmoothness,
} from "../support/helpers/stream-smoothness";
/**
* Streaming smoothness, measured on the path the user actually sees.
*
* The mock provider's "bursty-stream" model emits tokens in uneven runs separated
* by idle gaps, which is how real models arrive. Sampling and the statistics live
* in ../support/helpers/stream-smoothness.ts, which also documents what the two
* numbers mean.
*
* Budgets are deliberately loose. This spec exists to catch a regression in the
* shape of the stream, not to pin exact numbers on shared CI hardware. For
* reference, on a local dev daemon the paced reveal measures a CV around 1.3–1.5
* against 2.7 when painting deltas as they arrive.
*/
const SAMPLE_WINDOW_MS = 6_000;
const CHARS_PER_FRAME_CV_BUDGET = 2;
const UPDATE_GAP_P95_BUDGET_MS = 250;
const REACT_COMMIT_P95_BUDGET_MS = 12;
const REACT_COMMIT_COUNT_BUDGET = 600;
const RUN_AGENT_STREAM_PERF = process.env.PASEO_AGENT_STREAM_PERF_E2E === "1";
const agentStreamPerfDescribe = RUN_AGENT_STREAM_PERF ? test.describe : test.describe.skip;
interface ReactCommit {
id: string;
phase: "mount" | "update" | "nested-update";
actualDuration: number;
baseDuration: number;
startTime: number;
commitTime: number;
}
agentStreamPerfDescribe("Agent stream smoothness", () => {
test("reveals bursty model output at a steady rate", async ({ page }, testInfo) => {
test.setTimeout(120_000);
await page.addInitScript(() => {
Reflect.set(globalThis, "__PASEO_RENDER_PROFILE_ENABLED__", true);
});
const agent = await startRunningMockAgent(page, {
prefix: "stream-smoothness-",
model: "bursty-stream",
prompt: "Stream bursty output for smoothness measurement.",
});
try {
await awaitAssistantMessage(page);
await page.evaluate(() => {
const reset = Reflect.get(globalThis, "__PASEO_RESET_RENDER_PROFILE__");
if (typeof reset !== "function") {
throw new Error("Render profiler did not initialize");
}
reset();
});
const samples = await sampleStreamFrames(page, { sampleWindowMs: SAMPLE_WINDOW_MS });
const report = summarizeStreamSmoothness(samples, SAMPLE_WINDOW_MS);
const reactCommits = await readStreamReactCommits(page);
const reactReport = summarizeReactCommits(reactCommits);
await testInfo.attach("agent-stream-smoothness-report", {
body: JSON.stringify(report, null, 2),
contentType: "application/json",
});
await testInfo.attach("agent-stream-react-profile", {
body: JSON.stringify(reactReport, null, 2),
contentType: "application/json",
});
console.log(`[perf] Agent stream React: ${JSON.stringify(reactReport)}`);
console.log(`[perf] Agent stream smoothness: ${JSON.stringify(report)}`);
expect(
report.growthFrames,
"the stream must actually produce text during the sample window",
).toBeGreaterThan(20);
expect(
report.charsPerFrameCv,
`characters painted per frame should be steady (cv < ${CHARS_PER_FRAME_CV_BUDGET})`,
).toBeLessThan(CHARS_PER_FRAME_CV_BUDGET);
expect(
report.updateGapP95Ms,
`the text should not stall (p95 gap < ${UPDATE_GAP_P95_BUDGET_MS}ms)`,
).toBeLessThan(UPDATE_GAP_P95_BUDGET_MS);
expect(reactReport.count, "streaming should produce measured React commits").toBeGreaterThan(
20,
);
expect(
reactReport.count,
`the reveal should cap commits below ${REACT_COMMIT_COUNT_BUDGET} per sample window`,
).toBeLessThan(REACT_COMMIT_COUNT_BUDGET);
expect(
reactReport.actualDurationP95Ms,
`React commits should stay below ${REACT_COMMIT_P95_BUDGET_MS}ms at p95`,
).toBeLessThan(REACT_COMMIT_P95_BUDGET_MS);
} finally {
await agent.cleanup();
}
});
});
async function readStreamReactCommits(
page: Parameters<typeof sampleStreamFrames>[0],
): Promise<ReactCommit[]> {
return await page.evaluate(() => {
const samples = Reflect.get(globalThis, "__PASEO_RENDER_PROFILE__");
if (!Array.isArray(samples)) {
return [];
}
return samples.filter(
(sample): sample is ReactCommit =>
typeof sample === "object" &&
sample !== null &&
typeof Reflect.get(sample, "id") === "string" &&
Reflect.get(sample, "id").startsWith("AgentStreamSection:"),
);
});
}
function summarizeReactCommits(commits: readonly ReactCommit[]) {
const durations = commits
.map((commit) => commit.actualDuration)
.sort((left, right) => left - right);
const total = durations.reduce((sum, duration) => sum + duration, 0);
return {
count: commits.length,
actualDurationTotalMs: round2(total),
actualDurationMeanMs: round2(commits.length > 0 ? total / commits.length : 0),
actualDurationP50Ms: percentile(durations, 50),
actualDurationP95Ms: percentile(durations, 95),
actualDurationMaxMs: durations.at(-1) ?? 0,
};
}
function round2(value: number): number {
return Math.round(value * 100) / 100;
}
function percentile(sorted: readonly number[], percentileValue: number): number {
if (sorted.length === 0) {
return 0;
}
return sorted[
Math.min(sorted.length - 1, Math.floor((sorted.length - 1) * (percentileValue / 100)))
];
}