Files
windmill/ai_evals/adapters/frontend/benchmarkRunner.ts
centdix b0ddcf31e4 ci: add path-gated AI agent + ai_evals smoke workflows (#9640)
* ci: add path-gated AI agent integration tests workflow

Runs integration_tests/ai_agent_tests against real LLM providers
(Anthropic/OpenAI/Google) only when AI-agent backend code or the tests
change, since runs make paid LLM calls. Adds a conftest fixture that
skips provider-parametrized cases whose API keys are absent, so CI
exercises only the providers it has secrets for.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* ci: add path-gated ai_evals global-mode smoke workflow

Runs the global AI chat eval (global-test1) across one cheap model per
provider (Anthropic/OpenAI/Google/DeepSeek) only when the eval harness or
copilot chat code change, since runs make paid LLM calls. Builds Windmill
CE from source as the AI proxy; global tools/drafts run in the Vitest
bridge. Gates on the deterministic draft pipeline (run succeeded +
produced a draft + used write_script), not the variable LLM judge score.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* ci: run AI smokes on PR ready-for-review instead of every push

Switch the pull_request trigger from `synchronize` (every commit) to
`ready_for_review`, with a job guard skipping draft PRs, so the paid LLM
runs only fire when a PR is marked ready to merge (plus push-to-main and
manual dispatch).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(ai_evals): lazily load cli mode so non-cli evals skip the cli toolchain

The entrypoint eagerly imported modes/cli, which pulls the wmill CLI
guidance modules and their JSR deps (@cliffy/*). Global/flow/script/app
runs then crashed with "Cannot find module '@cliffy/ansi/colors'" when
the cli workspace deps were not installed. Import createCliModeRunner
dynamically inside runCliBenchmark instead.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* test(ai_agent): raise low max_completion_tokens to OpenAI's 16 minimum

OpenAI's /v1/responses rejects max_output_tokens < 16 with a 400, failing
test_low_max_tokens for openai. 16 still exercises a truncated response.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* ci: run ai_evals workflow on Node 22 for the frontend undici 8.x dep

The Vitest bridge loads frontend/node_modules/undici@8.x, which requires
Node >=22.19; Node 20 failed with "webidl.util.markAsUncloneable is not a
function" when loading vitest.config.ts.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(ai_evals): run frontend evals autonomously + give global-test1 more turns

Frontend evals (flow/script/app/global) ran the production chat prompt, which
assumes an interactive human — so cheaper models burned their turn budget
asking for confirmation, waiting for approval, or presenting a plan, sometimes
hitting maxTurns without producing a draft. Append a shared autonomy note in
baseEvalRunner (the path all frontend modes share, mirroring cli mode): act
directly on clear requests; only ask on genuinely ambiguous ones (preserving
the askUserQuestion cases). Also raise global-test1's maxTurns 8 -> 10 so a
model that over-explores still converges.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* ci(ai_evals): watch draft/prompt deps outside copilot/

The global eval runs production frontend code in-process, so the smoke's
behavior depends on files outside frontend/src/lib/components/copilot/**:
the draft model (userDraft.svelte.ts, userDraftDbSyncer.svelte.ts), script
inference (infer.ts), and the chat system prompts ($system_prompts ->
system_prompts/auto-generated). Add them to both push and PR path filters so
a change there actually triggers the smoke that gates on draft production.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix: skip direct provider tests without credentials

* feat: add ai evals skip judge flag

* fix: simplify ai evals ci gate

* fix: simplify ai evals smoke gate

* fix: handle ai eval workflow triggers

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-22 12:41:31 +02:00

143 lines
4.3 KiB
TypeScript

import { loadSelectedCases } from "../../core/cases";
import { resolveBackendValidationSettings } from "../../core/backendValidation";
import {
formatRunModelLabel,
getFrontendEvalModel,
resolveEvalModel,
} from "../../core/models";
import { buildRunResult } from "../../core/results";
import { runSuite } from "../../core/runSuite";
import type { BenchmarkRunResult, ModeRunner } from "../../core/types";
import { resolveWindmillBackendSettings } from "../../core/windmillBackendSettings";
import { emitFrontendBenchmarkProgress } from "./progress";
import { DEFAULT_JUDGE_MODEL } from "../../core/judge";
export type FrontendBenchmarkMode = "flow" | "app" | "script" | "global";
export async function runFrontendBenchmarkFromEnv(): Promise<BenchmarkRunResult> {
const mode = parseMode(process.env.WMILL_FRONTEND_AI_EVAL_MODE);
const caseIds = parseOptionalJsonStringArray(
process.env.WMILL_FRONTEND_AI_EVAL_CASE_IDS,
);
const runs = parsePositiveInteger(
process.env.WMILL_FRONTEND_AI_EVAL_RUNS,
"WMILL_FRONTEND_AI_EVAL_RUNS",
);
const emitProgress = process.env.WMILL_FRONTEND_AI_EVAL_PROGRESS === "1";
const verbose = process.env.WMILL_FRONTEND_AI_EVAL_VERBOSE === "1";
const executionOnly =
process.env.WMILL_FRONTEND_AI_EVAL_EXECUTION_ONLY === "1";
const judgeModel =
process.env.WMILL_FRONTEND_AI_EVAL_SKIP_JUDGE === "1" || executionOnly
? null
: DEFAULT_JUDGE_MODEL;
const model = resolveEvalModel(
mode,
process.env.WMILL_FRONTEND_AI_EVAL_MODEL,
);
const backendValidation = resolveBackendValidationSettings({
evalMode: mode,
requestedMode: process.env.WMILL_FRONTEND_AI_EVAL_BACKEND_VALIDATION,
});
const backendSettings = resolveWindmillBackendSettings();
const selectedCases = await loadSelectedCases(mode, caseIds);
const modeRunner = await getModeRunner(
mode,
getFrontendEvalModel(model),
backendValidation,
backendSettings,
);
const runModel = formatRunModelLabel(mode, model);
const caseResults = await runSuite({
modeRunner,
cases: selectedCases,
runs,
runModel,
judgeModel,
executionOnly,
concurrency: verbose ? 1 : undefined,
verbose,
onProgress: emitProgress
? (event) => emitFrontendBenchmarkProgress(event)
: undefined,
});
return buildRunResult({
mode,
runs,
runModel,
judgeModel,
caseResults,
});
}
async function getModeRunner(
mode: FrontendBenchmarkMode,
model: ReturnType<typeof getFrontendEvalModel>,
backendValidation: ReturnType<typeof resolveBackendValidationSettings>,
backendSettings: ReturnType<typeof resolveWindmillBackendSettings>,
): Promise<ModeRunner<any, any, any>> {
switch (mode) {
case "flow": {
const { createFlowModeRunner } = await import("../../modes/flow");
return createFlowModeRunner(model, backendValidation, backendSettings);
}
case "app": {
const { createAppModeRunner } = await import("../../modes/app");
return createAppModeRunner(model, backendSettings);
}
case "script": {
const { createScriptModeRunner } = await import("../../modes/script");
return createScriptModeRunner(
model,
backendValidation,
backendSettings,
);
}
case "global": {
const { createGlobalModeRunner } = await import("../../modes/global");
return createGlobalModeRunner(model, backendSettings);
}
}
}
function parseMode(value: string | undefined): FrontendBenchmarkMode {
if (
value === "flow" ||
value === "app" ||
value === "script" ||
value === "global"
) {
return value;
}
throw new Error(`Unsupported frontend benchmark mode: ${String(value)}`);
}
function parseOptionalJsonStringArray(value: string | undefined): string[] {
if (!value) {
return [];
}
const parsed = JSON.parse(value) as unknown;
if (
!Array.isArray(parsed) ||
parsed.some((entry) => typeof entry !== "string")
) {
throw new Error(
"WMILL_FRONTEND_AI_EVAL_CASE_IDS must be a JSON string array",
);
}
return parsed;
}
function parsePositiveInteger(
value: string | undefined,
envName: string,
): number {
const parsed = Number(value);
if (!Number.isInteger(parsed) || parsed <= 0) {
throw new Error(`${envName} must be a positive integer`);
}
return parsed;
}