PaperClipAI/tests/runner-e2e/chat-catalog-domains/internal-task-inventory.ts

import type { ChatCatalogCase } from "../chat-catalog-contract.js";
import { makeDefinition, fixtureWorker } from "./internal-definition.js";
import { checks, entity, hasLink, observation, project, sameTask, task, type Entity } from "./internal-support.js";

export function gradeOpenInventory(value: unknown) {
  const c = checks(); const o = observation(value);
  c.add("independent-inventory", !!o && Array.isArray(o.state.originalTasks), "The independently saved multi-page inventory is required.");
  if (!o || !Array.isArray(o.state.originalTasks)) return c.result();
  const originals = o.state.originalTasks as Entity[]; const p = entity(o.state, "project");
  const open = originals.filter(row => row.projectId === p.id && !["done", "cancelled"].includes(row.status ?? ""));
  const shown = open.filter(row => hasLink(o, "issues", row));
  const excluded = originals.filter(row => !open.includes(row));
  c.add("real-more-than-default-page", open.length > 500 && originals.every(row => sameTask(row, o.tasks.find(now => now.id === row.id))),
    "More than the production default 500 rows exist and the read-only answer preserves every saved task.");
  c.add("full-or-explicitly-bounded", shown.length === open.length || shown.length > 0
    && /\b(?:partial|first|limited|remaining|more results|not exhaustive|not all)\b/i.test(o.reply)
    && new RegExp(`\\b${shown.length}\\b`).test(o.reply) && new RegExp(`\\b${open.length}\\b`).test(o.reply),
    "All open work is linked, or the answer explicitly states its actual shown count and full inventory bound.");
  c.add("terminal-and-other-project-excluded", excluded.every(row => !hasLink(o, "issues", row)), "Completed, cancelled and other-project work is excluded.");
  const owners = o.state.ownerLabels as Record<string, string>;
  const positions = new Map<string, number>();
  let correctlyGrouped = !!owners;
  for (const row of shown) {
    const href = o.links.find(link => hasLink({ ...o, links: [link] }, "issues", row))!;
    const at = o.reply.indexOf(href); positions.set(row.id, at);
    if (at < 0) { correctlyGrouped = false; continue; }
    const lineStart = o.reply.lastIndexOf("\n", at) + 1;
    const lineEnd = o.reply.indexOf("\n", at);
    const line = o.reply.slice(lineStart, lineEnd < 0 ? undefined : lineEnd);
    const prior = o.reply.slice(0, lineStart).split("\n").reverse().find(text => /^\s*(?:#{1,6}\s+|\*\*)/.test(text)) ?? "";
    const owner = owners[row.assigneeAgentId ?? "unassigned"];
    if (!owner || !`${line} ${prior}`.toLowerCase().includes(owner.toLowerCase())) correctlyGrouped = false;
  }
  c.add("actual-single-owner-groups", shown.length > 0 && correctlyGrouped, "Each rendered item is grouped under its actual single assignee or the unassigned group.");
  c.add("blocked-first-within-owner", Object.keys(owners ?? {}).every(owner => {
    const group = shown.filter(row => (row.assigneeAgentId ?? "unassigned") === owner);
    const blocked = group.filter(row => row.status === "blocked").map(row => positions.get(row.id) ?? -1);
    const other = group.filter(row => row.status !== "blocked").map(row => positions.get(row.id) ?? -1);
    return !blocked.length || !other.length || Math.max(...blocked) < Math.min(...other);
  }), "Blocked work precedes other open items within every owner group.");
  c.add("read-only-audit", !o.activity.some(row => row.agentId === o.leadId && row.action !== "issue.read_marked"), "Organizing the answer creates no control-plane work changes.");
  return c.result();
}
function openInventory(): ChatCatalogCase {
  const base = makeDefinition("tasks.organize-open-work", async ctx => {
    const p = await project(ctx); const other = await project(ctx, "Experiments"); const worker = await fixtureWorker(ctx);
    const secondWorker = await fixtureWorker(ctx);
    await ctx.input.api.patch(`/api/agents/${secondWorker.id}`, { name: "Riley" });
    // Existing ownership is prerequisite state. Pause both owners before seeding
    // actionable statuses so preparing an inventory never starts provider work.
    await ctx.input.api.post(`/api/agents/${worker.id}/pause`);
    await ctx.input.api.post(`/api/agents/${secondWorker.id}/pause`);
    for (const id of [worker.id, secondWorker.id]) {
      if ((await ctx.input.api.get<Entity>(`/api/agents/${id}`)).status !== "paused") throw new Error("Inventory owner pause did not persist");
    }
    const originalTasks: Entity[] = [];
    for (let index = 0; index < 501; index++) {
      const owner = index % 3 === 0 ? worker.id : index % 3 === 1 ? secondWorker.id : null;
      originalTasks.push(await task(ctx, `Launch work ${String(index + 1).padStart(3, "0")}`, p.id, {
        status: index % 7 === 0 ? "blocked" : index % 2 === 0 ? "in_progress" : "todo", assigneeAgentId: owner,
        description: "Existing launch work. This inventory fixture records current state and must not trigger execution.",
      }));
    }
    originalTasks.push(await task(ctx, "Completed launch work", p.id, { status: "done" }), await task(ctx, "Cancelled launch work", p.id, { status: "cancelled" }),
      await task(ctx, "Unrelated experiment", other.id));
    return { project: p, originalTasks, workerIds: [worker.id, secondWorker.id], ownerLabels: { [worker.id]: "Morgan", [secondWorker.id]: "Riley", unassigned: "Unassigned" } };
  }, gradeOpenInventory);
  return { ...base, sourceFiles: [...base.sourceFiles, "chat-catalog-domains/internal-task-inventory.ts"], attemptTimeoutMs: 600_000,
    setupPolicy: { readiness: "preserve-scenario", invariantCheckIds: ["inventory-owners-paused", "lead-ready"] },
    async verifySetup(ctx, state) {
      const c = checks();
      const lead = await ctx.input.api.get<Entity & { lifecycleState: string }>(`/api/agents/${ctx.input.fixtures.agent.id}`);
      const owners = await Promise.all((state.workerIds as string[]).map(id => ctx.input.api.get<Entity & { lifecycleState: string }>(`/api/agents/${id}`)));
      c.add("inventory-owners-paused", owners.length === 2 && owners.every(row => row.status === "paused" && row.lifecycleState === "ready"),
        "The existing inventory owners are genuinely ready but paused so prerequisite task statuses admit no provider work.");
      c.add("lead-ready", lead.lifecycleState === "ready" && ["idle", "active"].includes(lead.status ?? ""), "The CEO remains ready for the canonical inventory query.");
      return c.result();
    },
    async verifyResult(ctx, state, value) {
      const o = observation(value)!; const shown = o.tasks.filter(row => hasLink(o, "issues", row));
      const representatives = new Map<string, Entity>();
      for (const row of shown) {
        const key = `${row.assigneeAgentId ?? "unassigned"}:${row.status === "blocked" ? "blocked" : "other"}`;
        if (!representatives.has(key)) representatives.set(key, row);
      }
      await base.verifyResult(ctx, state, { ...o, tasks: [...representatives.values()], projects: [], goals: [] });
    },
  };
}
export const taskInventoryCases: readonly ChatCatalogCase[] = [openInventory()];