mirror of
https://github.com/garrytan/gstack.git
synced 2026-10-31 22:14:25 +01:00
v1.91.299.0 fix(owner-brief): Captain brief resolves rec like gstack-gate render when a gate carries decisions (#3546) (#3549)
* fix(owner-brief): the Captain brief resolves rec through parseReply, so carried decisions match gstack-gate render (#3546) Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com> * fix(owner-brief): the Captain brief keeps answers already recorded at this gate, as answer --reply rec does Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com> * fix(owner-brief): the do-it line says every choice when rec keeps earlier answers Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com> --------- Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
This commit is contained in:
1 parent
415e343b52
commit
4c745eec10
6 files changed
+117
-14
No files matched your search
@@ -4,7 +4,7 @@
|
||||
* every choice rec takes shown in words, gate ids only on the agent line, the
|
||||
* rec_hash binding, and the incomplete and answered states.
|
||||
*/
|
||||
import { afterEach, describe, expect, test } from 'bun:test';
|
||||
import { afterAll, afterEach, beforeAll, describe, expect, test } from 'bun:test';
|
||||
import { spawnSync } from 'node:child_process';
|
||||
import * as fs from 'node:fs';
|
||||
import * as os from 'node:os';
|
||||
@@ -13,6 +13,8 @@ import { capyRelayLine, renderCaptainBrief } from '../lib/captain-brief';
|
||||
import { parseReply, recHash, type GateList } from '../lib/gate-list';
|
||||
import { findSection, headings } from '../lib/markdown-sections';
|
||||
import { loadBriefInputs } from '../lib/owner-brief';
|
||||
import { capyRunner, phaseVia, readRunJson, writePlan } from './helpers/autoplan-capy';
|
||||
import { SMALL_PLAN } from './helpers/autoplan-run-fixture';
|
||||
|
||||
const ROOT = path.resolve(import.meta.dir, '..');
|
||||
const dirs: string[] = [];
|
||||
@@ -151,3 +153,89 @@ describe('gstack-owner-brief render --captain', () => {
|
||||
expect(capyRelayLine('/r', {})).toBeUndefined();
|
||||
});
|
||||
});
|
||||
|
||||
describe('carried decisions at a cap gate (#3546)', () => {
|
||||
const r = capyRunner();
|
||||
afterAll(() => r.cleanup());
|
||||
const SECTIONS = '\n## Before → after\n\nBefore: retries hammer. After: they back off.\n\n## Why it matters\n\nOutages.\n\n## What caused this work\n\nIncident 7.\n\n## Big decisions\n\n1. Exponential backoff.\n';
|
||||
const PLAN = SMALL_PLAN + SECTIONS;
|
||||
const WORDS: Record<string, string> = { 'keep your direction': 'keep the plan as written', 'accept the change': "make the reviewer's change" };
|
||||
|
||||
function cappedRun(): string {
|
||||
const dir = path.join(r.host.base, 'carried');
|
||||
const plan = writePlan(dir, PLAN);
|
||||
const out = path.join(dir, 'run');
|
||||
expect(r.ap(['next', '--out', out, '--plan', plan, '--no-ui', '--max-rounds', '1']).status).toBe(0);
|
||||
const challenge = (title: string) => ({ severity: 'Medium', title, fix: `change: ${title}`, user_challenge: true, changes: 'scope' });
|
||||
phaseVia(r, out, 'ceo', { native: [challenge('Use a queue table'), challenge('Add metrics')], outside: [challenge('Drop jitter')] });
|
||||
phaseVia(r, out, 'eng', { native: [], outside: [] });
|
||||
const ds = JSON.parse(fs.readFileSync(path.join(out, 'gate.json'), 'utf8')).items.filter((i: any) => i.id.startsWith('d'));
|
||||
expect(ds.map((i: any) => i.recommended)).toEqual(['a', 'a', 'a']);
|
||||
expect(r.ap(['answer', '--by', 'human', '--out', out, '--gate-rev', '1', '--reply', `${ds.map((i: any) => `${i.id}b`).join(' ')} p1b`]).status).toBe(0);
|
||||
const v2 = path.join(dir, 'plan-v2.md');
|
||||
fs.writeFileSync(v2, PLAN.replace('cap 30 s', 'cap 60 s, dead-letter after 8 attempts'));
|
||||
const rev = r.ap(['revise', '--out', out, '--plan', v2, '--incorporates', readRunJson(out).amendments.map((a: any) => a.decision).join(',')]);
|
||||
expect(rev.stdout).toMatch(/^REVISED: plan=[0-9a-f]{12} reopened=none capped=ceo,eng$/m);
|
||||
return out;
|
||||
}
|
||||
const cli = (bin: string, args: string[]) => spawnSync('bun', [path.join(ROOT, 'bin', bin), ...args], { encoding: 'utf8', timeout: 60_000 }).stdout ?? '';
|
||||
function renders(out: string) {
|
||||
const gate = JSON.parse(fs.readFileSync(path.join(out, 'gate.json'), 'utf8')) as GateList;
|
||||
const gateText = cli('gstack-gate', ['render', path.join(out, 'gate.json'), '--all']);
|
||||
const briefText = cli('gstack-owner-brief', ['render', '--captain', '--out', out]);
|
||||
const gateRec = JSON.parse(/\`rec\` resolves to: (\{.*\})$/m.exec(gateText)![1]!).choices as Record<string, string>;
|
||||
const briefRec = Object.fromEntries(/rec resolves to ([a-z0-9 ]+);/.exec(briefText)![1]!.split(' ').map(t => [t.slice(0, -1), t.slice(-1)]));
|
||||
return { gate, gateText, briefText, gateRec, briefRec };
|
||||
}
|
||||
let out = '';
|
||||
beforeAll(() => { out = cappedRun(); });
|
||||
|
||||
test('the Captain brief resolves rec to the carried choices, as gstack-gate render does, with the same choice labels', () => {
|
||||
const { gate, gateText, briefText, gateRec, briefRec } = renders(out);
|
||||
expect(gate.gate_rev).toBe(2);
|
||||
expect(gate.items.some(i => i.cap?.kind === 'delta')).toBe(true);
|
||||
const carried = gate.items.filter(i => i.carried);
|
||||
expect(carried.map(i => [i.recommended, i.carried!.chosen])).toEqual([['a', 'b'], ['a', 'b'], ['a', 'b']]);
|
||||
expect(briefRec).toEqual(gateRec);
|
||||
expect(Object.fromEntries(carried.map(i => [i.id, gateRec[i.id]]))).toEqual({ d1: 'b', d2: 'b', d3: 'b' });
|
||||
gate.items.filter(i => i.id !== 'p1').forEach((item, n) => {
|
||||
const text = item.options.find(o => o.key === gateRec[item.id])!.text;
|
||||
expect(gateText).toContain(` ${item.id}${gateRec[item.id]}) ${text}`);
|
||||
expect(briefText).toContain(` ${n + 1}. ${item.title.replace(/\s*\[[^\]]*\]\s*$/, '')}: ${WORDS[text] ?? text}`);
|
||||
});
|
||||
expect(briefText).toContain("Use a queue table: make the reviewer's change (your answer at gate_rev 1, carried forward).");
|
||||
expect(briefText).toContain('rec keeps the 3 you already answered and takes the recommended answer for the rest.');
|
||||
expect(briefText).toContain('Reply **rec** to keep the answers you already gave and take every other recommendation (recommended)');
|
||||
});
|
||||
|
||||
test('an answer already given at this gate is what rec keeps, as gstack-autoplan answer --reply rec records it', () => {
|
||||
const copy = path.join(r.host.base, 'prior');
|
||||
fs.cpSync(out, copy, { recursive: true });
|
||||
const partial = r.ap(['answer', '--by', 'human', '--out', copy, '--gate-rev', '2', '--reply', 'd2a']);
|
||||
expect(partial.stdout).toMatch(/^GATE_PENDING: /m);
|
||||
const { briefText, briefRec } = renders(copy);
|
||||
expect(briefRec).toMatchObject({ d1: 'b', d2: 'a', d3: 'b' });
|
||||
expect(briefText).toContain('Add metrics: keep the plan as written (a reviewer suggested changing it) (your answer at this gate).');
|
||||
const answered = r.ap(['answer', '--by', 'human', '--out', copy, '--gate-rev', '2', '--reply', 'rec']);
|
||||
expect(answered.stdout).toMatch(/^APPROVED: gate_rev=2/m);
|
||||
const recorded = fs.readFileSync(path.join(copy, 'decisions.jsonl'), 'utf8').split('\n').filter(Boolean).map(l => JSON.parse(l)).filter(d => d.gate_rev === 2);
|
||||
expect(Object.fromEntries(recorded.map(d => [d.label.toLowerCase(), d.chosen]))).toEqual(briefRec);
|
||||
});
|
||||
|
||||
test('negative control: without carried decisions (gate items and their decision rows) both resolve rec to the recommendations', () => {
|
||||
const copy = path.join(r.host.base, 'uncarried');
|
||||
fs.cpSync(out, copy, { recursive: true });
|
||||
const gate = JSON.parse(fs.readFileSync(path.join(copy, 'gate.json'), 'utf8')) as GateList;
|
||||
expect(Object.fromEntries(gate.items.map(i => [i.id, i.recommended]))).not.toEqual(renders(out).gateRec);
|
||||
for (const item of gate.items) delete item.carried;
|
||||
fs.writeFileSync(path.join(copy, 'gate.json'), JSON.stringify(gate));
|
||||
const rows = fs.readFileSync(path.join(copy, 'decisions.jsonl'), 'utf8').split('\n').filter(Boolean).map(l => JSON.parse(l));
|
||||
fs.writeFileSync(path.join(copy, 'decisions.jsonl'), rows.map(d => JSON.stringify(d.carried_from ? { ...d, status: 'pending', chosen: undefined, answered_at: undefined, reply: undefined, carried_from: undefined } : d)).join('\n') + '\n');
|
||||
const { briefText, gateRec, briefRec } = renders(copy);
|
||||
expect(briefRec).toEqual(gateRec);
|
||||
expect(gateRec).toEqual(Object.fromEntries(gate.items.map(i => [i.id, i.recommended])));
|
||||
expect(briefText).toContain('Use a queue table: keep the plan as written (a reviewer suggested changing it).');
|
||||
expect(briefText).not.toContain('carried forward');
|
||||
expect(briefText).toContain('rec takes the recommended answer for each.');
|
||||
});
|
||||
});
|
||||
Reference in new issue
Block a user