:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.
Python ab-testingai-analyticsanalyticscdpdata-warehouseexperimentsfeature-flagsjavascriptproduct-analyticspythonreactsession-replaysurveystypescriptweb-analytics
import '@testing-library/jest-dom'
import { cleanup, render, screen } from '@testing-library/react'
import { Provider } from 'kea'
import { QueryScanSummary, QueryScanWarning } from '~/queries/schema/schema-general'
import { initKeaTests } from '~/test/init'
import { QueryScanBanner } from './QueryScanBanner'
const SUMMARY: QueryScanSummary = {
rows_read: 8_400_000_000,
duration_ms: 19_000,
analysis_requested: true,
}
const FINDING: QueryScanWarning = {
kind: 'no_event_filter',
message: 'This query read every event in its date range.',
fix: 'Add an event filter naming the events this question is about.',
actionable: true,
}
const INSIGHT_SIDE_FINDING: QueryScanWarning = {
...FINDING,
kind: 'no_start_date',
fix_location: 'insight_date_range',
message: 'No date range is set for this insight or dashboard.',
fix: 'Set a date range on the insight or the dashboard.',
}
const BY_DESIGN_FINDING: QueryScanWarning = {
...FINDING,
kind: 'no_start_date',
by_design: true,
message: 'This query finds a first event ever, so it reads all your data by design.',
fix: 'Do not propose a time bound for that read.',
actionable: false,
}
describe('QueryScanBanner', () => {
beforeEach(() => {
initKeaTests()
})
afterEach(() => cleanup())
it.each([
{
label: 'nothing to advise',
findings: [],
assistantPrompt: null,
advice: [],
banner: false,
note: false,
fixer: false,
},
{
label: 'a finding the assistant can fix',
findings: [FINDING],
assistantPrompt: 'Help me get what this query is trying to find, as fast as possible.',
advice: [FINDING.message],
banner: true,
note: false,
fixer: true,
},
// The backend sends no prompt for a finding fixed on the insight, so there is nothing to hand the assistant.
{
label: 'a finding fixed on the insight',
findings: [INSIGHT_SIDE_FINDING],
assistantPrompt: null,
advice: [INSIGHT_SIDE_FINDING.message],
banner: true,
note: false,
fixer: false,
},
// A by-design finding explains the read; shown as a note, it must not read as something to fix.
{
label: 'a finding the person cannot act on',
findings: [BY_DESIGN_FINDING],
assistantPrompt: null,
advice: [BY_DESIGN_FINDING.message],
banner: false,
note: true,
fixer: false,
},
{
label: 'a by-design finding beside one to act on',
findings: [BY_DESIGN_FINDING, FINDING],
assistantPrompt: 'Help me get what this query is trying to find, as fast as possible.',
advice: [BY_DESIGN_FINDING.message, FINDING.message],
banner: true,
note: true,
fixer: true,
},
])('keeps the stat line and shows $label', ({ findings, assistantPrompt, advice, banner, note, fixer }) => {
const { container } = render(
<Provider>
<QueryScanBanner
queryScan={{
summary: SUMMARY,
findings,
cacheKey: 'cache-key',
assistantPrompt,
}}
onFixWithAI={jest.fn()}
/>
</Provider>
)
expect(screen.getByText(/Read 8,400,000,000 rows in 19.0 s/)).toBeVisible()
const shown = findings.map((finding) => finding.message).filter((message) => screen.queryByText(message))
expect(shown).toEqual(advice)
expect(!!container.querySelector('.LemonBanner')).toBe(banner)
expect(!!container.querySelector('[data-attr="query-scan-note"]')).toBe(note)
// Each finding sits at its own level: the by-design one never inside the warning.
expect(container.querySelector('.LemonBanner')?.textContent ?? '').not.toContain(BY_DESIGN_FINDING.message)
expect(!!screen.queryByText('Fix with AI')).toBe(fixer)
})
})