Files
PaperClipAI/ui/src/pages/DesignSandboxTasks.tsx
T
DottaandPaperclip b31ec0a2dc feat(ui): link routines to a filtered task list (#18317)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work.
> - Routines create tasks that people need to inspect and manage.
> - The routines list sends its View all runs link to Audit.
> - Tasks already supports routine origin filters in its API, but the
page has no control for them.
> - This pull request adds a Routine tasks checkbox and links to that
selection.
> - People can inspect routine work in the normal task list, search, and
board views.

## Linked Issues or Issue Description

**What existing behavior does this improve?**

The routines list link and the Tasks filters.

**Current behavior**

View all runs opens Audit. People cannot select routine-created tasks in
the Tasks filter menu.

**Proposed behavior**

View all tasks opens `/issues?view=all&originKind=routine_execution`.
Routine tasks is selected. People can toggle or clear it from Tasks.

**Reason and benefit**

People can use normal task controls to inspect work created by routines.
The filter includes manually started routine runs. It excludes manual
tasks and other automatic work.

**Breaking changes**

The routines list link changes its label and destination. No API,
schema, authorization, or routine execution contract changes.

## What Changed

- Add Routine tasks under Visibility in the current, streamlined, and
legacy Tasks filters. Keep Live runs only in the same group and use Show
agent-created tasks with proper capitalization where that control is
available.
- Keep the selection in the URL through reload, search, task navigation,
and checkbox changes.
- Apply the existing API origin filter before pagination. Include
routine tasks even when other automatic tasks are hidden.
- Keep source filtering in the shared task filter module and collection
owners. Reuse the existing company-scoped issue-list contract. Hide the
live-run control in the design sandbox, which does not load live-run
IDs.
- Update regression coverage, product docs, feature/access maps, and the
toolbar preview.

## Verification

- All 58 final-head checks are complete with no failures for
`f46c2e409b20fa372d166f4721601a0471894fb0`, including successful `ci /
verify` and `ci / e2e` gates. Greptile and Paperclip Review rate this
commit 5/5 with no actionable findings or unresolved review threads. [CI
run](https://github.com/paperclipai/paperclip/actions/runs/41078093110).
- Local checks pass: full repo typecheck/build for the feature, UI
typecheck/build after the final refinements, token gates, feature-map
checks, and Storybook catalog generation.
- Regression coverage exercises both real list implementations, filter
grouping and selection, search and all board columns, URL/pagination
state, and mixed pending-decision batches. The Visibility update passed
all 89 relevant tests across separate runs. The final sandbox correction
passes all 10 filter-menu tests, including three cases that reproduced
the bug before the fix. The earlier broad UI run passed 10,043 tests
across 845 files.
- Browser proof used the actual local app with disposable PostgreSQL
data: three tasks created through the routine-run API and one manual
task. Followed View all tasks; verified the selected Routine tasks
checkbox under Visibility, toggling, reload, search, and board view.
Checked current, streamlined, and legacy filter presentations.
Screenshots show the link, routine-only results, and requested
capitalization/grouping.
- Failure accounting: the initial broad UI run found two preference
expectations that needed the new false default; those were corrected
before the full UI pass. Scrolling tests hit five-second timeouts in
combined runs and passed in isolated reruns without timeout changes. A
duplicate full local UI run was stopped after the earlier final-commit
CI UI lanes passed, with no failures reported in that stopped run.
- Full local `pnpm test:run` could not complete because native
PostgreSQL initialization exhausted the host semaphore pool. A direct
initdb probe confirmed `could not create semaphores: No space left on
device`. Browser verification used a disposable container; no host
settings or other workspace processes were changed.
- Review found that the newly exposed live-run control could empty the
design sandbox, which has no live-run IDs. The sandbox now disables that
unsupported control; its regression test failed before the fix and
passes after it.

## Risks

- Low risk. The change uses the existing routine origin and read
permissions. A routine task remains automatic work, including when a
person starts it manually.
- The routine checkbox is exposed only on the Tasks collection. Other
filter consumers retain their controls and normalize the new field to
false.
- Routine histories and Audit run links inside routine details keep
their current behavior.

## Model Used

- OpenAI GPT-6, through Codex, assisted with implementation and
verification. The exact runtime model ID and context window are not
exposed in this session. Capabilities used: repository editing, command
execution, tests, and browser interaction.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-10-30 10:28:58 -05:00

106 lines
8.1 KiB
TypeScript

import { useEffect } from "react";
import { useInfiniteQuery, useMutation, useQuery, useQueryClient } from "@tanstack/react-query";
import { ListChecks, SearchX } from "lucide-react";
import { accessApi } from "@/api/access";
import { agentsApi } from "@/api/agents";
import { useAccountIdentity } from "@/api/companies-query";
import { issuesApi } from "@/api/issues";
import { projectsApi } from "@/api/projects";
import { EmptyState } from "@/components/EmptyState";
import { KanbanBoard } from "@/components/KanbanBoard";
import { TaskList, type TaskListItem } from "@/components/TaskList";
import { TaskListToolbar } from "@/components/TaskListToolbar";
import { TruncatedTextMarqueeProvider } from "@/components/TruncatedText";
import { Button } from "@/components/ui/button";
import { useBreadcrumbs } from "@/context/BreadcrumbContext";
import { useCompany } from "@/context/CompanyContext";
import { useDialogActions } from "@/context/DialogContext";
import { useTaskListCollection } from "@/hooks/useTaskListCollection";
import { buildCompanyUserLabelMap } from "@/lib/company-members";
import { queryKeys } from "@/lib/queryKeys";
import { taskListViewStorageKey } from "@/lib/task-list-views";
import type { Issue } from "@paperclipai/shared";
const pageSize = 200;
/** Real API adapter for the Design proposals / Pages / Tasks composition.
* All visual and collection behavior comes from the reviewed shared subjects. */
export function DesignSandboxTasks() {
const { selectedCompanyId: companyId } = useCompany();
const { userId } = useAccountIdentity();
const { openNewIssue } = useDialogActions();
const { setBreadcrumbs } = useBreadcrumbs();
const client = useQueryClient();
useEffect(() => { setBreadcrumbs([{ label: "Tasks" }]); }, [setBreadcrumbs]);
const tasks = useInfiniteQuery({
queryKey: [...queryKeys.issues.list(companyId!), "design-sandbox", userId],
queryFn: ({ pageParam, signal }) => issuesApi.list(companyId!, {
limit: pageSize, offset: pageParam, includeAgentCreatedTasks: true,
includeRoutineExecutions: true, sortField: "id", sortDir: "asc",
}, { signal }),
initialPageParam: 0,
getNextPageParam: (last, _pages, offset) => last.length === pageSize ? offset + pageSize : undefined,
enabled: !!companyId,
});
const agents = useQuery({ queryKey: queryKeys.agents.list(companyId!), queryFn: () => agentsApi.list(companyId!), enabled: !!companyId });
const projects = useQuery({ queryKey: queryKeys.projects.list(companyId!, { includeArchived: true }), queryFn: () => projectsApi.list(companyId!, { includeArchived: true }), enabled: !!companyId });
const directory = useQuery({ queryKey: queryKeys.access.companyUserDirectory(companyId!), queryFn: () => accessApi.listUserDirectory(companyId!), enabled: !!companyId });
const labels = useQuery({ queryKey: queryKeys.issues.labels(companyId!), queryFn: () => issuesApi.listLabels(companyId!), enabled: !!companyId });
// The ordinary organization list does not include viewer read context.
// Ask the server for the signed-in user's unread set instead of guessing
// unread state from dates or reusing Storybook fixture flags.
const unread = useQuery({
queryKey: [...queryKeys.issues.list(companyId!), "design-sandbox-unread", userId],
queryFn: async ({ signal }) => {
const ids = new Set<string>();
let page: Issue[];
let offset = 0;
do {
page = await issuesApi.list(companyId!, { unreadForUserId: "me", includeAgentCreatedTasks: true,
includeRoutineExecutions: true, limit: pageSize, offset, sortField: "id", sortDir: "asc" }, { signal });
for (const issue of page) ids.add(issue.id);
offset += pageSize;
} while (page.length === pageSize);
return ids;
},
enabled: !!companyId && !!userId,
});
const users = buildCompanyUserLabelMap(directory.data?.users);
const agentNames = new Map(agents.data?.map(agent => [agent.id, agent.name]));
const projectById = new Map(projects.data?.map(project => [project.id, project]));
const items: TaskListItem[] = [...new Map((tasks.data?.pages.flat() ?? []).map(issue => [issue.id, issue])).values()].map(issue => ({
issue: { ...issue, project: issue.project ?? projectById.get(issue.projectId ?? "") },
unread: unread.data?.has(issue.id) ?? false,
createdBy: issue.createdByAgentId ? agentNames.get(issue.createdByAgentId) ?? "Unknown agent" : users.get(issue.createdByUserId ?? "") ?? (issue.createdByUserId === "local-board" ? "Board" : "Unknown user"),
assignee: issue.assigneeAgentId ? agentNames.get(issue.assigneeAgentId) ?? "Unknown agent" : users.get(issue.assigneeUserId ?? "") ?? (issue.assigneeUserId === "local-board" ? "Board" : null),
}));
const collection = useTaskListCollection({ items, companyId, userId,
viewStorageKey: taskListViewStorageKey("design-integration-sandbox", userId ?? "local-board"),
});
const update = useMutation({
mutationFn: ({ id, data }: { id: string; data: Record<string, unknown> }) => issuesApi.update(id, data),
onSuccess: () => { void client.invalidateQueries({ queryKey: queryKeys.issues.list(companyId!) }); },
});
const error = tasks.error ?? agents.error ?? projects.error ?? directory.error ?? labels.error ?? unread.error ?? update.error;
const empty = collection.companyItems.length === 0
? <EmptyState icon={ListChecks} message="No tasks in this organization" action="New Task" onAction={() => openNewIssue()} />
: <EmptyState icon={SearchX} message="No tasks match this view" action="Clear filters" hideActionIcon onAction={collection.clear} />;
if (!companyId) return <EmptyState icon={ListChecks} message="Select an organization to view tasks." />;
return <TruncatedTextMarqueeProvider><div className="task-row-design-preview flex flex-col gap-8" data-testid="design-sandbox-tasks">
<TaskListToolbar view={collection.view} onViewChange={collection.setView} query={collection.query} onQueryChange={collection.setQuery}
layout={collection.layout} onLayoutChange={collection.setLayout} allExpanded={collection.allExpanded} onToggleExpanded={collection.toggleExpanded}
onNewTask={() => openNewIssue()} filters={{ state: collection.filters, onChange: collection.updateFilters, activeFilterCount: collection.activeFilterCount,
enableLiveRunsFilter: false,
agents: agents.data ?? [], projects: projects.data ?? [], labels: labels.data ?? [], currentUserId: userId, creators: collection.creators }} />
{error ? <div role="alert" className="flex items-center gap-2 text-sm text-destructive"><span>{error instanceof Error ? error.message : "Could not load tasks"}</span><Button variant="outline" size="sm" onClick={() => { update.reset(); void client.invalidateQueries({ queryKey: queryKeys.issues.list(companyId) }); void agents.refetch(); void projects.refetch(); void directory.refetch(); void labels.refetch(); }}>Retry</Button></div> : null}
{tasks.isPending ? <p role="status" className="text-sm text-muted-foreground">Loading tasks…</p> : <>
{collection.narrowed ? <p role="status" className="design-task-results">{collection.matchingItems.length} matching tasks{collection.layout === "list" && collection.visibleItems.length > collection.matchingItems.length ? " · Parent tasks included for context" : ""}</p> : null}
<div hidden={collection.layout !== "list"} className="design-task-list-panel"><TaskList key={companyId} items={collection.visibleItems} allItems={collection.companyItems} collapsedIds={collection.collapsedIds} onCollapsedIdsChange={collection.setCollapsedIds} emptyState={empty} /></div>
{collection.layout === "board" ? collection.matchingItems.length ? <KanbanBoard showOverflowFades issues={collection.matchingItems.map(item => item.issue)} agents={agents.data} onUpdateIssue={(id, data) => update.mutate({ id, data })} /> : empty : null}
{tasks.hasNextPage ? <div className="flex items-center gap-3 text-sm text-muted-foreground"><span>Search, views and filters cover the {items.length} loaded tasks.</span><Button variant="outline" size="sm" disabled={tasks.isFetchingNextPage} onClick={() => void tasks.fetchNextPage()}>{tasks.isFetchingNextPage ? "Loading…" : "Load more tasks"}</Button></div> : null}
</>}
</div></TruncatedTextMarqueeProvider>;
}