: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 { useValues } from 'kea'
import { useCallback } from 'react'
import type { NotebookComponentRunStatus, NotebookComponentRunStatusResolver } from 'lib/components/MarkdownNotebook'
import type { NotebookComponentBlockNode } from 'lib/components/MarkdownNotebook/types'
import { NotebookNodeType } from '../types'
import { MARKDOWN_TAG_TO_NOTEBOOK_NODE_TYPE } from './markdownNotebookRegistry'
import {
NotebookNodeRunTerminalStatus,
notebookNodeStalenessLogic,
NotebookStaleReason,
} from './notebookNodeStalenessLogic'
// Only the revamped code cells report run state. Every other block — markdown, insights, images —
// has nothing to run, so it stays grey.
const RUNNABLE_NODE_TYPES: (NotebookNodeType | undefined)[] = [NotebookNodeType.SQLV2, NotebookNodeType.PythonV2]
const TERMINAL_STATUSES: NotebookNodeRunTerminalStatus[] = ['done', 'failed', 'interrupted']
const persistedRunStatus = (value: unknown): NotebookNodeRunTerminalStatus | undefined =>
TERMINAL_STATUSES.find((status) => status === value)
export function resolveNotebookComponentRunStatus(
node: NotebookComponentBlockNode,
staleNodeIds: Record<string, NotebookStaleReason>,
nodeRunStatuses: Record<string, NotebookNodeRunTerminalStatus>
): NotebookComponentRunStatus {
if (!RUNNABLE_NODE_TYPES.includes(MARKDOWN_TAG_TO_NOTEBOOK_NODE_TYPE[node.tagName])) {
return 'idle'
}
// A cell pins its own id the first time it runs; until then the block id is a content
// fingerprint, which is what the rest of the notebook keys the cell by too.
const nodeId = typeof node.props.nodeId === 'string' ? node.props.nodeId : node.id
// The session's own record of the run wins. A cell that produced a result persists how that
// run ended too, so a reload keeps the color — an interrupted run leaves a partial result
// behind just like a completed one, so the result alone can't stand in for the outcome.
const runStatus = nodeRunStatuses[nodeId] ?? persistedRunStatus(node.props.runStatus)
// A cell that didn't finish has nothing trustworthy to be out of date, so its own outcome
// outranks an upstream re-run.
if (runStatus === 'failed' || runStatus === 'interrupted') {
return 'error'
}
if (staleNodeIds[nodeId]) {
return 'stale'
}
return runStatus === 'done' ? 'success' : 'idle'
}
export function useNotebookComponentRunStatusResolver(shortId: string): NotebookComponentRunStatusResolver {
const { staleNodeIds, nodeRunStatuses } = useValues(notebookNodeStalenessLogic({ shortId }))
return useCallback(
(node: NotebookComponentBlockNode) => resolveNotebookComponentRunStatus(node, staleNodeIds, nodeRunStatuses),
[staleNodeIds, nodeRunStatuses]
)
}