Improve Notion ingest UX

This commit is contained in:
Andrey Avtomonov 2026-05-12 16:21:05 +02:00
parent d0f650f44a
commit 9a9e40939a
20 changed files with 615 additions and 279 deletions

View file

@ -355,6 +355,53 @@ describe('runKtxConnectionNotion', () => {
expect(io.stdout()).toContain('rootPageIds: 1');
});
it('uses inline Notion auth_token for interactive discovery', async () => {
const projectDir = join(tempDir, 'project');
const initialized = await initKtxProject({ projectDir, projectName: 'warehouse' });
await writeProjectConfig(projectDir, {
...initialized.config,
connections: {
'notion-main': {
driver: 'notion',
auth_token: 'ntn_inline_token',
crawl_mode: 'selected_roots',
root_page_ids: [PAGE_IDS.engineering],
root_database_ids: [],
root_data_source_ids: [],
max_pages_per_run: 12,
max_knowledge_creates_per_run: 2,
max_knowledge_updates_per_run: 7,
last_successful_cursor: null,
},
},
});
const api = fakeNotionApi([notionPage(PAGE_IDS.engineering, 'Engineering')]);
const createNotionApi = vi.fn((authToken: string) => {
expect(authToken).toBe('ntn_inline_token');
return api;
});
const io = makeIo();
await expect(
runKtxConnectionNotion(
{
command: 'pick',
projectDir,
connectionId: 'notion-main',
mode: 'interactive',
},
io.io,
{
createNotionApi,
renderPicker: vi.fn(async (): Promise<PickerRenderResult> => ({ kind: 'quit' })),
},
),
).resolves.toBe(0);
expect(createNotionApi).toHaveBeenCalledOnce();
expect(io.stdout()).toContain('No changes saved.');
});
it('passes partial-discovery warnings into the TUI banner state', async () => {
const projectDir = join(tempDir, 'project');
const initialized = await initKtxProject({ projectDir, projectName: 'warehouse' });

View file

@ -1,4 +1,4 @@
import { parseNotionConnectionConfig, resolveNotionAuthToken } from '@ktx/context/connections';
import { parseNotionConnectionConfig, resolveNotionConnectionAuthToken } from '@ktx/context/connections';
import { type NotionApi, type NotionBotInfo, NotionClient } from '@ktx/context/ingest';
import {
type KtxLocalProject,
@ -223,7 +223,7 @@ export async function runKtxConnectionNotion(
const project = await loadProject({ projectDir: args.projectDir });
const rawConnection = notionConnection(project, args.connectionId);
const notion = parseNotionConnectionConfig(rawConnection);
const authToken = await resolveNotionAuthToken(notion.auth_token_ref, { env: deps.env });
const authToken = await resolveNotionConnectionAuthToken(notion, { env: deps.env });
const api = deps.createNotionApi ? deps.createNotionApi(authToken) : new NotionClient(authToken);
const discovery = await discoverNotionPickerPages(api);
const tree = buildPickerTree(discovery.pages);

View file

@ -186,6 +186,91 @@ describe('runKtxIngest viz and replay', () => {
expect(io.stdout()).toContain('Connection: warehouse');
});
it('prints live viz final summaries as errors when the report has failed work units', async () => {
const projectDir = join(tempDir, 'project');
await writeWarehouseConfig(projectDir);
const io = makeIo({ isTTY: true, stdinIsTTY: true, columns: 120 });
const liveSession = {
update: vi.fn(),
close: vi.fn(),
isClosed: vi.fn(() => false),
};
const startLiveMemoryFlow = vi.fn(async (_input: MemoryFlowReplayInput, _io: unknown) => liveSession);
const runLocal = vi.fn(async (input: RunLocalIngestOptions): Promise<LocalIngestResult> => {
input.memoryFlow?.emit({ type: 'source_acquired', adapter: 'notion', trigger: 'manual_resync', fileCount: 37 });
input.memoryFlow?.update({
syncId: 'sync-notion',
plannedWorkUnits: [
{
unitKey: 'notion-cluster-1',
rawFiles: ['pages/a.md'],
peerFileCount: 0,
dependencyCount: 0,
},
],
});
input.memoryFlow?.emit({ type: 'chunks_planned', chunkCount: 1, workUnitCount: 1, evictionCount: 0 });
input.memoryFlow?.emit({
type: 'work_unit_finished',
unitKey: 'notion-cluster-1',
status: 'failed',
reason: 'notion-cluster-1 failed: {"error":"invalid_grant","error_description":"reauth related error (invalid_rapt)"}',
});
input.memoryFlow?.emit({ type: 'report_created', runId: 'live-failed' });
input.memoryFlow?.finish('done');
const failedWorkUnit = {
...localFakeBundleReport('live-failed').body.workUnits[0],
unitKey: 'notion-cluster-1',
rawFiles: ['pages/a.md'],
status: 'failed' as const,
reason: 'notion-cluster-1 failed: {"error":"invalid_grant","error_description":"reauth related error (invalid_rapt)"}',
actions: [],
touchedSlSources: [],
};
const report = localFakeBundleReport('live-failed', {
id: 'report-live-failed',
runId: 'run-live-failed',
connectionId: input.connectionId,
sourceKey: input.adapter,
body: {
workUnits: [failedWorkUnit],
failedWorkUnits: [failedWorkUnit.unitKey],
},
});
return {
result: {
jobId: 'live-failed',
runId: report.runId,
syncId: report.body.syncId,
diffSummary: report.body.diffSummary,
workUnitCount: report.body.workUnits.length,
failedWorkUnits: report.body.failedWorkUnits,
artifactsWritten: report.body.provenanceRows.length,
commitSha: report.body.commitSha,
},
report,
};
});
await expect(
runKtxIngest(
{
command: 'run',
projectDir,
connectionId: 'notion-main',
adapter: 'notion',
outputMode: 'viz',
},
io.io,
{ runLocalIngest: runLocal, startLiveMemoryFlow },
),
).resolves.toBe(1);
expect(io.stdout()).toContain('Memory-flow summary: error');
expect(io.stdout()).toContain('Notion authorization expired');
});
it('falls back to text live rendering when the TUI live session is unavailable', async () => {
const projectDir = join(tempDir, 'project');
await writeWarehouseConfig(projectDir);

View file

@ -1076,17 +1076,17 @@ describe('runKtxIngest', () => {
),
).resolves.toBe(0);
const stdout = io.stdout();
expect(stdout).toContain('[5%] Fetching source files for warehouse/historic-sql');
expect(stdout).toContain('[15%] Fetched 3 source files from historic-sql');
expect(stdout).toContain('[45%] Planned 1 work unit');
expect(stdout).toContain('[80%] Processed 1/1 work units');
expect(stdout).toContain('[100%] Ingest completed');
expect(stdout).toContain('Report: report-live-1');
expect(io.stderr()).toBe('');
const stderr = io.stderr();
expect(stderr).toContain('[5%] Fetching source files for warehouse/historic-sql');
expect(stderr).toContain('[15%] Fetched 3 source files from historic-sql');
expect(stderr).toContain('[45%] Planned 1 work unit');
expect(stderr).toContain('[80%] Processed 1/1 work units');
expect(stderr).toContain('[100%] Ingest completed');
expect(io.stdout()).toContain('Report: report-live-1');
expect(io.stdout()).not.toContain('[5%]');
});
it('writes plain TTY ingest progress and final report to stdout', async () => {
it('writes plain TTY ingest progress to stderr and final report to stdout', async () => {
const projectDir = join(tempDir, 'project');
await writeWarehouseConfig(projectDir);
const sourceDir = join(tempDir, 'source');
@ -1113,9 +1113,9 @@ describe('runKtxIngest', () => {
),
).resolves.toBe(0);
expect(io.stdout()).toContain('[5%] Fetching source files for warehouse/fake');
expect(io.stderr()).toContain('[5%] Fetching source files for warehouse/fake');
expect(io.stdout()).toContain('Report: report-live-1');
expect(io.stderr()).toBe('');
expect(io.stdout()).not.toContain('[5%]');
});
it('prints plain WorkUnit step progress during long-running local ingest', async () => {
@ -1202,13 +1202,13 @@ describe('runKtxIngest', () => {
),
).resolves.toBe(0);
const stdout = io.stdout();
expect(stdout).toContain('[45%] Planned 2 work units');
expect(stdout).toContain('[55%] Processing 1/2 work units: historic-sql-table-public-orders');
expect(stdout).toContain(
const stderr = io.stderr();
expect(stderr).toContain('[45%] Planned 2 work units');
expect(stderr).toContain('[55%] Processing 1/2 work units: historic-sql-table-public-orders');
expect(stderr).toContain(
'\r[58%] Processing work units: 0/2 complete, 1 active; latest historic-sql-table-public-orders step 7/40\u001b[K',
);
expect(stdout).toContain('[68%] Processed 1/2 work units');
expect(stderr).toContain('[68%] Processed 1/2 work units');
});
it('renders concurrent WorkUnit step progress as transient aggregate status', async () => {
@ -1294,14 +1294,14 @@ describe('runKtxIngest', () => {
),
).resolves.toBe(0);
const stdout = io.stdout();
expect(stdout).toContain(
const stderr = io.stderr();
expect(stderr).toContain(
'\r[56%] Processing work units: 0/6 complete, 6 active; latest historic-sql-table-public-suppliers step 1/40\u001b[K',
);
expect(stdout).not.toContain(
expect(stderr).not.toContain(
'\n[56%] Processing 6/6 work units: historic-sql-table-public-suppliers step 1/40\n',
);
expect(stdout).toContain('\n[100%] Ingest completed\n');
expect(stderr).toContain('\n[100%] Ingest completed\n');
});
it('passes local Looker pull-config options and agent runner into scheduled ingest for Looker scheduled ingest', async () => {

View file

@ -306,7 +306,7 @@ function createPlainIngestProgressRenderer(
if (!hasPendingTransient) {
return;
}
io.stdout.write('\n');
io.stderr.write('\n');
hasPendingTransient = false;
};
@ -315,12 +315,12 @@ function createPlainIngestProgressRenderer(
lastPercent = nextPercent;
const line = `[${nextPercent}%] ${message}`;
if (options?.transient === true) {
io.stdout.write(`\r${line}\u001b[K`);
io.stderr.write(`\r${line}\u001b[K`);
hasPendingTransient = true;
return;
}
flush();
io.stdout.write(`${line}\n`);
io.stderr.write(`${line}\n`);
};
return {
@ -437,6 +437,21 @@ function initialRunMemoryFlowInput(
};
}
function finalRunMemoryFlowInput(snapshot: MemoryFlowReplayInput, report: IngestReportSnapshot): MemoryFlowReplayInput {
const status = reportStatus(report);
return {
...snapshot,
runId: report.runId,
connectionId: report.connectionId,
adapter: report.sourceKey,
status,
syncId: report.body.syncId,
reportId: report.id,
reportPath: report.id,
errors: status === 'error' ? report.body.failedWorkUnits : snapshot.errors,
};
}
function managedDaemonOptionsForIngestRun(
args: Extract<KtxIngestArgs, { command: 'run' }>,
io: KtxIngestIo,
@ -592,7 +607,7 @@ export async function runKtxIngest(
...(memoryFlow ? { memoryFlow } : {}),
});
if (shouldUseLiveViz && memoryFlow) {
latestMemoryFlowSnapshot = memoryFlow.snapshot();
latestMemoryFlowSnapshot = finalRunMemoryFlowInput(memoryFlow.snapshot(), result.report);
liveTui?.close();
liveTui = null;
io.stdout.write(formatMemoryFlowFinalSummary(latestMemoryFlowSnapshot));