mirror of
https://github.com/CopilotKit/CopilotKit.git
synced 2026-09-14 16:26:20 +08:00
5cd77233b7
Brings ms-agent-python to one-to-one parity with langgraph-python (the D5
north star). Playwright e2e suite goes from 49/108 (~26%) → 164/178 (~92%),
33 of 37 cells fully green.
Manifest parity:
- Drop 4 MAF-only cells with no LGP analog: agentic-chat-reasoning,
hitl-in-chat-booking, shared-state-write, reasoning-default-render.
Reasoning is handled by reasoning-default + reasoning-custom (LGP);
booking pill folds into hitl-in-chat; shared-state-write was a TODO stub.
- Rename byoc-hashbrown → declarative-hashbrown and byoc-json-render →
declarative-json-render. Demo dir, API route dir, and frontend agent id
follow LGP's naming. Python module files retain the legacy `byoc_*`
prefix and FastAPI paths stay `/byoc-hashbrown` / `/byoc-json-render`
(matches LGP's "module name retains legacy graph id" convention).
- Port LGP `_shared/`, `_shared/interrupt-fallback-slots.ts`, and
`demos/layout.tsx` for one-to-one parity.
Cells ported verbatim from LGP (page + spec):
- agentic-chat, auth, beautiful-chat, chat-customization-css, chat-slots,
declarative-gen-ui, declarative-hashbrown, declarative-json-render,
frontend-tools, frontend-tools-async, gen-ui-agent, gen-ui-interrupt,
gen-ui-tool-based, headless-complete, headless-simple, hitl-in-app,
hitl-in-chat, shared-state-read, shared-state-read-write,
shared-state-streaming, subagents, tool-rendering, plus all four
tool-rendering* variants, a2ui-fixed-schema, agent-config, mcp-apps,
multimodal, open-gen-ui, open-gen-ui-advanced, prebuilt-popup,
prebuilt-sidebar, readonly-state-agent-context, reasoning-default,
reasoning-custom, voice.
Backend infrastructure:
- Swap shared `OpenAIChatClient` (Responses API) → `OpenAIChatCompletionClient`
(ChatCompletions). Root cause of the cross-cell post-tool ChatClientException
family: Responses API is stateful and only sends NEW items per leg,
relying on `previous_response_id` for history. aimock has no view of
that server-side state, so second-leg requests arrived without the
user message — fixture matchers keyed on `userMessage` couldn't fire
and the run fell through to real OpenAI. ChatCompletions sends full
history every leg, matching the LGP wire shape.
- Bump @ag-ui/client ^0.0.43 → ^0.0.53 (matches google-adk/LGP). Fixes
the REASONING_* Zod discriminator trap on the catch-all agent.
- Regenerate package-lock.json in isolation outside the pnpm monorepo so
npm-arborist doesn't resolve transitives against pnpm's hoisted
symlinks (avoid 40+ `../../../node_modules/.pnpm/...` paths in the
lockfile that break `npm ci` inside Docker).
- Add `yaml` (^2.8.4) for the new `src/app/demos/layout.tsx` that reads
manifest.yaml for per-cell page titles (LGP parity).
New / re-added MAF agent backends with LGP-equivalent behavior:
- reasoning_agent.py (uses Responses API explicitly — the only chat
client that emits AG-UI REASONING_MESSAGE_* events; rest of the
integration stays on ChatCompletions).
- tool_rendering_agent.py (non-reasoning sibling of the existing
reasoning_chain variant; shares tool surface via direct imports so
they can never drift apart; routes the three catchall cells to a
non-reasoning backend so the default renderer spec stops failing on
leaked reasoning blocks).
- gen_ui_agent.py — `set_steps` tool + `steps` state schema +
`predict_state_config` mirrors LGP's StateStreamingMiddleware shape.
- shared_state_streaming.py — `write_document` tool with
`predict_state_config` that streams the `document` arg into
`state.document` per-token.
- readonly_state_agent_context.py — minimal agent that consumes
frontend-provided `useAgentContext` entries; no tools.
- headless_complete_agent.py — three deterministic tools (`get_weather`,
`get_stock_price`, `get_revenue_chart`) mounted at /headless-complete
on the mcp-apps runtime (was routing to catch-all sales agent, which
returned seeded-random weather instead of the deterministic 68°F the
test asserts on).
Wiring:
- copilotkit/route.ts: register the new agents, drop the stale
shared-state-write entry, route all three tool-rendering variants to
the non-reasoning backend (the reasoning-chain cell keeps its own
dedicated path), register reasoning-default + reasoning-custom on
/reasoning, register gen-ui-agent on /gen-ui-agent,
shared-state-streaming on /shared-state-streaming,
readonly-state-agent-context on its dedicated path.
- copilotkit-mcp-apps/route.ts: register headless-complete agent (was
missing — the strict useAgent runtime sync in the newer
@copilotkit/react-core surfaced the gap).
- copilotkit-declarative-hashbrown/route.ts + copilotkit-declarative-json-render/route.ts:
new dedicated runtimes; agent IDs and runtime URLs follow LGP.
- copilotkit-declarative-gen-ui/route.ts: drop non-LGP `openGenerativeUI:
false` for parity.
A2UI tool rename — `render_a2ui` → `_design_a2ui_surface`:
- Ported LGP's `tools/generate_a2ui.py` (LGP renamed the secondary-LLM
tool to `_design_a2ui_surface` to avoid the A2UI middleware's bypass;
shared d5-all.json fixtures key the response on this name).
- Renamed every `render_a2ui` occurrence in src/agents/{a2ui_dynamic,
agent,beautiful_chat}.py and `tools/__init__.py`.
- Updated 4 declarative-gen-ui aimock fixtures to pass `context` arg in
the first-leg `generate_a2ui` tool call (agent_framework doesn't
auto-inject AgentSession into our @tool function so `session=None` and
the secondary-LLM `user_content` was defaulting to a catch-all string
containing "KPI dashboard" — every pill matched the KPI fixture).
Aimock router patch persisted alongside the integration changes:
hasToolResult matcher restricted to scan only messages after the last
user message (was global). The patch lives in F:/projects/cpk/aimock —
upstream PR pending.
Test infrastructure:
- playwright.config.ts: cap local workers at 4 + retries at 1. CI keeps
workers=1, retries=2. `agent_framework.Agent` is reused across requests
and the shared OpenAI HTTP client serialises concurrent SSE streams;
>4 workers makes 30s timeouts inevitable on a few cells. Confirmed
with hard data: workers=1 = 164 passed (16.8 min), workers=4+retries=1
= 164 passed (7.2 min), workers=undefined = 159 passed. Same green
set, ~2x faster. Long-term upstream fix is per-request Agent
instantiation in agent_framework_ag_ui.
Remaining 14 failures across 4 cells documented per-cell in the Notion
D5 sweep doc (declarative-gen-ui A2UI surface mounting, multimodal
attachment forwarding, tool-rendering-default-catchall multi-pill chain,
tool-rendering-reasoning-chain multi-leg chains). Each has a specific
next-pass action.
187 lines
7.7 KiB
TypeScript
187 lines
7.7 KiB
TypeScript
import { test, expect } from "@playwright/test";
|
||
|
||
// QA reference: qa/frontend-tools-async.md
|
||
// Demo source: src/app/demos/frontend-tools-async/{page.tsx, notes-card.tsx}
|
||
//
|
||
// The demo registers ONE async frontend tool via `useFrontendTool`:
|
||
// `query_notes(keyword: string)`. The handler sleeps 500ms (simulated local
|
||
// DB latency) then returns up to 5 matches from an in-memory 7-note DB.
|
||
// A custom `render` mounts `NotesCard` which exposes:
|
||
// - `data-testid="notes-card"` (outer container)
|
||
// - `data-testid="notes-keyword"` (heading: `Matching "<keyword>"`)
|
||
// - `data-testid="notes-list"` (the <ul> of matches)
|
||
// - `data-testid="note-n1"` … `note-n7` per-note rows
|
||
//
|
||
// Genuine-pass strategy: the deterministic aimock fixtures match each pill's
|
||
// verbatim prompt with a dedicated `query_notes(keyword=…)` tool call so the
|
||
// async handler runs against the real client-side NOTES_DB. The card's
|
||
// `keyword` heading is then the keyword we asserted in the fixture, and the
|
||
// `notes-list` rows reflect the actual handler-filtered results — proving
|
||
// the async tool round-trip end-to-end.
|
||
|
||
test.describe("Frontend Tools (async query_notes)", () => {
|
||
test.setTimeout(120_000);
|
||
|
||
test.beforeEach(async ({ page }) => {
|
||
await page.goto("/demos/frontend-tools-async");
|
||
});
|
||
|
||
test("page loads with composer and 3 pills", async ({ page }) => {
|
||
await expect(page.getByPlaceholder("Type a message")).toBeVisible();
|
||
await expect(
|
||
page.getByRole("button", { name: /Find project-planning notes/i }),
|
||
).toBeVisible({ timeout: 15_000 });
|
||
await expect(
|
||
page.getByRole("button", { name: /Search for 'auth'/i }),
|
||
).toBeVisible({ timeout: 15_000 });
|
||
await expect(
|
||
page.getByRole("button", { name: /What do I have about reading\?/i }),
|
||
).toBeVisible({ timeout: 15_000 });
|
||
});
|
||
|
||
test("project-planning pill → Notes DB card with project-planning notes", async ({
|
||
page,
|
||
}) => {
|
||
await page
|
||
.getByRole("button", { name: /Find project-planning notes/i })
|
||
.click();
|
||
|
||
const notesCard = page.locator('[data-testid="notes-card"]').first();
|
||
await expect(notesCard).toBeVisible({ timeout: 60_000 });
|
||
|
||
// The keyword heading proves the async handler resolved against the
|
||
// fixture-emitted `query_notes(keyword="project planning")` call.
|
||
await expect(notesCard.locator('[data-testid="notes-keyword"]')).toHaveText(
|
||
/Matching\s+["“]project planning["”]/i,
|
||
{ timeout: 30_000 },
|
||
);
|
||
|
||
// The async handler matches notes n1 ("Q2 project planning kickoff")
|
||
// and n5 ("Project planning retrospective notes") from NOTES_DB.
|
||
const list = notesCard.locator('[data-testid="notes-list"]');
|
||
await expect(list).toBeVisible({ timeout: 30_000 });
|
||
await expect(notesCard.locator('[data-testid="note-n1"]')).toBeVisible();
|
||
await expect(notesCard.locator('[data-testid="note-n5"]')).toBeVisible();
|
||
|
||
// Anti-regression: the generic-plan boilerplate from the cross-cell
|
||
// catch-all fixture must NOT appear. If it does, the d5-all.json
|
||
// fixture lost match priority to feature-parity.json's "plan" entry.
|
||
await expect(
|
||
page.getByText("Research the topic, Outline key points"),
|
||
).toHaveCount(0);
|
||
});
|
||
|
||
test("auth pill → Notes DB card with auth-related notes", async ({
|
||
page,
|
||
}) => {
|
||
await page.getByRole("button", { name: /Search for 'auth'/i }).click();
|
||
|
||
const notesCard = page.locator('[data-testid="notes-card"]').first();
|
||
await expect(notesCard).toBeVisible({ timeout: 60_000 });
|
||
|
||
await expect(notesCard.locator('[data-testid="notes-keyword"]')).toHaveText(
|
||
/Matching\s+["“]auth["”]/i,
|
||
{ timeout: 30_000 },
|
||
);
|
||
|
||
// The async handler matches note n2 ("Planning: migrate auth to
|
||
// passkeys") on the "auth" tag.
|
||
const list = notesCard.locator('[data-testid="notes-list"]');
|
||
await expect(list).toBeVisible({ timeout: 30_000 });
|
||
await expect(notesCard.locator('[data-testid="note-n2"]')).toBeVisible();
|
||
|
||
// Anti-regression: the showcase-assistant catch-all from
|
||
// feature-parity.json must NOT have intercepted this prompt.
|
||
await expect(page.getByText("I'm your showcase assistant")).toHaveCount(0);
|
||
});
|
||
|
||
test("reading pill → Notes DB card with Book recommendations + locked narration", async ({
|
||
page,
|
||
}) => {
|
||
await page
|
||
.getByRole("button", { name: /What do I have about reading\?/i })
|
||
.click();
|
||
|
||
const notesCard = page.locator('[data-testid="notes-card"]').first();
|
||
await expect(notesCard).toBeVisible({ timeout: 60_000 });
|
||
|
||
// Keyword heading + match count + per-note testid + content +
|
||
// tag chip — the full canonical shape per spec test #4.
|
||
await expect(notesCard.locator('[data-testid="notes-keyword"]')).toHaveText(
|
||
/Matching\s+["“]reading["”]/i,
|
||
{ timeout: 30_000 },
|
||
);
|
||
await expect(notesCard.getByText("1 match", { exact: false })).toBeVisible({
|
||
timeout: 30_000,
|
||
});
|
||
|
||
const note = notesCard.locator('[data-testid="note-n4"]');
|
||
await expect(note).toBeVisible({ timeout: 30_000 });
|
||
await expect(note.getByText("Book recommendations")).toBeVisible();
|
||
await expect(note.getByText(/Thinking Fast and Slow/i)).toBeVisible();
|
||
await expect(
|
||
note.getByText(/The Design of Everyday Things/i),
|
||
).toBeVisible();
|
||
await expect(note.getByText("reading", { exact: true })).toBeVisible();
|
||
|
||
// Locked narration leading phrase — proves the deterministic 2nd-turn
|
||
// fixture wired correctly through the async tool result.
|
||
await expect(
|
||
page
|
||
.locator('[data-testid="copilot-assistant-message"]')
|
||
.filter({
|
||
hasText:
|
||
'You have a note titled "Book recommendations" that is tagged with "reading',
|
||
})
|
||
.first(),
|
||
).toBeVisible({ timeout: 60_000 });
|
||
});
|
||
|
||
// Regression for the aimock multi-pill bug:
|
||
// The three frontend-tools-async fixtures used `hasToolResult: false/true`
|
||
// gates to split first-turn (emit `query_notes`) vs. follow-up (narration).
|
||
// After the user clicked a tool-using pill earlier in the same thread, the
|
||
// first-turn fixture was skipped (the thread already had a prior tool
|
||
// result), the follow-up fixture fired immediately with just narration,
|
||
// and the Notes DB card never rendered. Fix: chain via `toolCallId`, drop
|
||
// the gates. This test drives all three pills in a single thread and
|
||
// asserts every pill renders its own Notes DB card.
|
||
test("sequential pills in one thread each render their own Notes DB card", async ({
|
||
page,
|
||
}) => {
|
||
// Three pills × async-handler latency × LLM mock chain; the existing
|
||
// describe-level 120s is not enough once we drive all three in one test.
|
||
test.setTimeout(240_000);
|
||
|
||
const cards = page.locator('[data-testid="notes-card"]');
|
||
|
||
await page
|
||
.getByRole("button", { name: /Find project-planning notes/i })
|
||
.click();
|
||
await expect.poll(() => cards.count(), { timeout: 60_000 }).toBe(1);
|
||
await expect(
|
||
page.locator('[data-testid="notes-keyword"]', {
|
||
hasText: /Matching\s+[""“]project planning[""”]/i,
|
||
}),
|
||
).toBeVisible({ timeout: 60_000 });
|
||
|
||
await page.getByRole("button", { name: /Search for 'auth'/i }).click();
|
||
await expect.poll(() => cards.count(), { timeout: 60_000 }).toBe(2);
|
||
await expect(
|
||
page.locator('[data-testid="notes-keyword"]', {
|
||
hasText: /Matching\s+[""“]auth[""”]/i,
|
||
}),
|
||
).toBeVisible({ timeout: 60_000 });
|
||
|
||
await page
|
||
.getByRole("button", { name: /What do I have about reading\?/i })
|
||
.click();
|
||
await expect.poll(() => cards.count(), { timeout: 60_000 }).toBe(3);
|
||
await expect(
|
||
page.locator('[data-testid="notes-keyword"]', {
|
||
hasText: /Matching\s+[""“]reading[""”]/i,
|
||
}),
|
||
).toBeVisible({ timeout: 60_000 });
|
||
});
|
||
});
|