mirror of
https://github.com/windmill-labs/windmill.git
synced 2026-10-15 00:02:56 +00:00
* feat: windmill-chat sdk for chat-mode flows in external frontends and raw apps Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018aQiZNAU8g17kWkyTryS5J * fix: keep streamed answers until persisted, finish turns after history fallback Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018aQiZNAU8g17kWkyTryS5J * feat: ai sdk transport and assistant-ui runtime for windmill-chat Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix: finish a turn from the flow result until its answer row lands, hash chat ids without crypto.subtle Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix: judge a turn answered by a persisted assistant row, wherever it was fetched Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix: attribute a turn's answer to its own jobs, keep a local turn when switching conversations Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix: mirror local history on every change, attribute failure-handler answers to the turn Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix: new chat per token string in the React hook, idle after destroy, no reorder on view Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix: recreate the hook's chat on any credential change, namespace local history per user Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix: send the latest inputs from the React hook Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> --------- Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
179 lines
7.4 KiB
TypeScript
179 lines
7.4 KiB
TypeScript
import { describe, expect, test } from 'bun:test'
|
|
import type { UIMessage, UIMessageChunk } from 'ai'
|
|
import { createWindmillChatTransport, toUIMessages } from '../src/ai-sdk'
|
|
import type { ChatMessage } from '../src/types'
|
|
import { fetchMock, json, ndjson, sse, text, type Route } from './support'
|
|
|
|
const FLOW = 'f/chat/agent'
|
|
const run: Route = (c) =>
|
|
c.method === 'POST' && c.url.pathname === `/api/w/ws/jobs/run/f/${FLOW}` ? text('job-1') : undefined
|
|
const streamPath = '/api/w/ws/jobs_u/getupdate_sse/job-1'
|
|
|
|
const userMessage = (text: string): UIMessage => ({ id: 'u1', role: 'user', parts: [{ type: 'text', text }] })
|
|
|
|
async function collect(stream: ReadableStream<UIMessageChunk>): Promise<UIMessageChunk[]> {
|
|
const chunks: UIMessageChunk[] = []
|
|
const reader = stream.getReader()
|
|
while (true) {
|
|
const { value, done } = await reader.read()
|
|
if (done) return chunks
|
|
chunks.push(value)
|
|
}
|
|
}
|
|
|
|
describe('createWindmillChatTransport', () => {
|
|
test('maps an agent turn to AI SDK chunks and derives the conversation from the chat id', async () => {
|
|
const { fetch, calls } = fetchMock(run, (c) =>
|
|
c.url.pathname === streamPath
|
|
? sse([
|
|
{
|
|
type: 'update',
|
|
new_result_stream: ndjson(
|
|
{ type: 'reasoning_token_delta', content: 'think' },
|
|
{ type: 'token_delta', content: 'Let me ' },
|
|
{ type: 'tool_call', call_id: 'c1', function_name: 'lookup' },
|
|
{ type: 'tool_call_arguments', call_id: 'c1', function_name: 'lookup', arguments: '{"q":1}' },
|
|
{ type: 'tool_execution', call_id: 'c1', function_name: 'lookup' }
|
|
),
|
|
stream_offset: 5
|
|
},
|
|
{
|
|
type: 'update',
|
|
new_result_stream: ndjson(
|
|
{ type: 'tool_result', call_id: 'c1', function_name: 'lookup', result: '{"answer":42}', success: true },
|
|
{ type: 'token_delta', content: '42' }
|
|
),
|
|
stream_offset: 7,
|
|
completed: true,
|
|
only_result: { output: '42', messages: [] }
|
|
}
|
|
])
|
|
: undefined
|
|
)
|
|
const transport = createWindmillChatTransport({
|
|
baseUrl: 'http://wm.test',
|
|
workspace: 'ws',
|
|
flowPath: FLOW,
|
|
token: 'tok',
|
|
inputs: { tone: 'kind' },
|
|
fetch
|
|
})
|
|
const chunks = await collect(
|
|
await transport.sendMessages({
|
|
trigger: 'submit-message',
|
|
chatId: 'chat-abc',
|
|
messageId: undefined,
|
|
messages: [userMessage('what is it?')],
|
|
abortSignal: undefined,
|
|
body: { locale: 'fr' }
|
|
})
|
|
)
|
|
|
|
const runCall = calls.find((c) => c.method === 'POST')!
|
|
expect(runCall.body).toEqual({ tone: 'kind', locale: 'fr', user_message: 'what is it?' })
|
|
expect(runCall.url.searchParams.get('memory_id')).toBe(transport.conversationId('chat-abc'))
|
|
expect(transport.conversationId('chat-abc')).toMatch(/^[0-9a-f]{8}-[0-9a-f]{4}-5[0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$/)
|
|
expect(transport.conversationId('chat-abc')).toBe(transport.conversationId('chat-abc'))
|
|
expect(transport.conversationId('chat-abd')).not.toBe(transport.conversationId('chat-abc'))
|
|
expect(transport.conversationId('4D6E5C8C-2C3B-4E1A-9F31-0B2E6F1C9D10')).toBe(
|
|
'4d6e5c8c-2c3b-4e1a-9f31-0b2e6f1c9d10'
|
|
)
|
|
|
|
const shape = chunks.map((c) => ('delta' in c ? `${c.type}:${c.delta}` : c.type))
|
|
expect(shape).toEqual([
|
|
'start',
|
|
'reasoning-start',
|
|
'reasoning-delta:think',
|
|
'text-start',
|
|
'text-delta:Let me ',
|
|
'reasoning-end',
|
|
'text-end',
|
|
'tool-input-start',
|
|
'tool-input-available',
|
|
'tool-output-available',
|
|
'text-start',
|
|
'text-delta:42',
|
|
'text-end',
|
|
'finish'
|
|
])
|
|
expect(chunks.find((c) => c.type === 'tool-input-available')).toMatchObject({
|
|
toolCallId: 'c1',
|
|
toolName: 'lookup',
|
|
input: { q: 1 },
|
|
dynamic: true
|
|
})
|
|
expect(chunks.find((c) => c.type === 'tool-output-available')).toMatchObject({ output: { answer: 42 } })
|
|
// The job finished, so there is nothing to reconnect to.
|
|
expect(await transport.reconnectToStream({ chatId: 'chat-abc' })).toBeNull()
|
|
})
|
|
|
|
test('answers from the flow result when nothing streamed, and reports a failed flow as an error', async () => {
|
|
let turn = 0
|
|
const { fetch } = fetchMock(
|
|
run,
|
|
(c) => {
|
|
if (c.url.pathname !== streamPath) return undefined
|
|
turn++
|
|
return sse([
|
|
turn === 1
|
|
? { type: 'update', completed: true, only_result: { windmill_chat_answer: 'From a script' } }
|
|
: { type: 'update', completed: true, only_result: { error: { name: 'ExecutionErr', message: 'boom' } } }
|
|
])
|
|
},
|
|
(c) =>
|
|
c.url.pathname === '/api/w/ws/jobs_u/completed/get_result_maybe/job-1'
|
|
? json({ completed: true, success: false })
|
|
: undefined
|
|
)
|
|
const transport = createWindmillChatTransport({ baseUrl: 'http://wm.test', workspace: 'ws', flowPath: FLOW, fetch })
|
|
const send = () =>
|
|
transport.sendMessages({
|
|
trigger: 'submit-message',
|
|
chatId: 'c',
|
|
messageId: undefined,
|
|
messages: [userMessage('hi')],
|
|
abortSignal: undefined
|
|
})
|
|
const first = await collect(await send())
|
|
expect(first.map((c) => ('delta' in c ? c.delta : c.type))).toEqual(['start', 'text-start', 'From a script', 'text-end', 'finish'])
|
|
const second = await collect(await send())
|
|
expect(second.map((c) => c.type)).toEqual(['start', 'error'])
|
|
expect(second[1]).toMatchObject({ errorText: 'ExecutionErr: boom' })
|
|
})
|
|
|
|
test('refuses attachments with a clear error', async () => {
|
|
const transport = createWindmillChatTransport({ baseUrl: 'http://wm.test', workspace: 'ws', flowPath: FLOW, fetch: fetchMock().fetch })
|
|
await expect(
|
|
transport.sendMessages({
|
|
trigger: 'submit-message',
|
|
chatId: 'c',
|
|
messageId: undefined,
|
|
messages: [{ id: 'u', role: 'user', parts: [{ type: 'file', mediaType: 'image/png', url: 'data:...' }] }],
|
|
abortSignal: undefined
|
|
})
|
|
).rejects.toThrow('attachments are not supported')
|
|
})
|
|
})
|
|
|
|
describe('toUIMessages', () => {
|
|
test('folds a turn into one assistant message with reasoning, tool and text parts', () => {
|
|
const base = { success: true, createdAt: '2026-01-01T00:00:00Z', pending: false }
|
|
const messages: ChatMessage[] = [
|
|
{ ...base, id: 'u1', role: 'user', content: 'hi' },
|
|
{ ...base, id: 't1', role: 'tool', content: 'Used lookup tool', tool: { callId: 'c1', name: 'lookup', status: 'success', arguments: '{"q":1}', result: '42' } },
|
|
{ ...base, id: 'a1', role: 'assistant', content: 'The answer is 42', reasoning: 'hmm' },
|
|
{ ...base, id: 'u2', role: 'user', content: 'thanks' },
|
|
{ ...base, id: 't2', role: 'tool', content: 'Error executing lookup', success: false, tool: { name: 'lookup', status: 'error' } }
|
|
]
|
|
const ui = toUIMessages(messages)
|
|
expect(ui.map((m) => [m.id, m.role, m.parts.map((p) => p.type)])).toEqual([
|
|
['u1', 'user', ['text']],
|
|
['t1', 'assistant', ['dynamic-tool', 'reasoning', 'text']],
|
|
['u2', 'user', ['text']],
|
|
['t2', 'assistant', ['dynamic-tool']]
|
|
])
|
|
expect(ui[1].parts[0]).toMatchObject({ toolCallId: 'c1', toolName: 'lookup', state: 'output-available', input: { q: 1 }, output: 42 })
|
|
expect(ui[3].parts[0]).toMatchObject({ state: 'output-error', errorText: 'Error executing lookup' })
|
|
})
|
|
})
|