feat(ms-ai-architect): Spor 1 — unified migrate-corpus applier (backup+atomisk+advisor-fence+per-felt-idempotent, TDD) [skip-docs]
This commit is contained in:
parent
c13b7505e0
commit
cec98b5cb5
2 changed files with 512 additions and 0 deletions
281
tests/kb-update/test-migrate-apply.test.mjs
Normal file
281
tests/kb-update/test-migrate-apply.test.mjs
Normal file
|
|
@ -0,0 +1,281 @@
|
|||
// tests/kb-update/test-migrate-apply.test.mjs
|
||||
// Integration tests for scripts/kb-update/migrate-corpus.mjs — the unified Spor 1
|
||||
// applier (Step 7). Hermetic: every case runs against a fresh mkdtemp corpus via the
|
||||
// injected `--root`/`backupRoot`, so nothing ever writes to the real skills/ tree.
|
||||
//
|
||||
// Covers the Step-7 acceptance set: idempotency, advisor-exclusion, throw-guard,
|
||||
// atomic (no *.tmp.* leftover + import-grep), backup+restore round-trip, dry-run
|
||||
// default byte-identity, body-after-`---` identity + one `## Innhold`, post-write
|
||||
// parse assertion (500B truncation), and normalize→insert ordering.
|
||||
|
||||
import { test } from 'node:test';
|
||||
import assert from 'node:assert/strict';
|
||||
import {
|
||||
mkdtempSync,
|
||||
mkdirSync,
|
||||
rmSync,
|
||||
writeFileSync,
|
||||
readFileSync,
|
||||
readdirSync,
|
||||
existsSync,
|
||||
} from 'node:fs';
|
||||
import { tmpdir } from 'node:os';
|
||||
import { join, dirname } from 'node:path';
|
||||
import { fileURLToPath } from 'node:url';
|
||||
import {
|
||||
planFileMutation,
|
||||
migrateCorpus,
|
||||
} from '../../scripts/kb-update/migrate-corpus.mjs';
|
||||
import {
|
||||
parseTypeHeader,
|
||||
parseSourceHeader,
|
||||
parseVerifiedHeader,
|
||||
} from '../../scripts/kb-update/lib/kb-headers.mjs';
|
||||
|
||||
const HERE = dirname(fileURLToPath(import.meta.url));
|
||||
const SRC = join(HERE, '..', '..', 'scripts', 'kb-update', 'migrate-corpus.mjs');
|
||||
|
||||
function withTmp(fn) {
|
||||
const dir = mkdtempSync(join(tmpdir(), 'migrate-test-'));
|
||||
try {
|
||||
return fn(dir);
|
||||
} finally {
|
||||
rmSync(dir, { recursive: true, force: true });
|
||||
}
|
||||
}
|
||||
|
||||
/** Materialize a corpus under `root` from a {relpath: content} map. */
|
||||
function makeCorpus(root, files) {
|
||||
for (const [rel, content] of Object.entries(files)) {
|
||||
const path = join(root, rel);
|
||||
mkdirSync(dirname(path), { recursive: true });
|
||||
writeFileSync(path, content, 'utf8');
|
||||
}
|
||||
}
|
||||
|
||||
/** Read every file under `root` into a {relpath: content} map (posix keys). */
|
||||
function readAll(root) {
|
||||
const out = {};
|
||||
function walk(dir, prefix) {
|
||||
for (const entry of readdirSync(dir, { withFileTypes: true })) {
|
||||
const rel = prefix ? `${prefix}/${entry.name}` : entry.name;
|
||||
if (entry.isDirectory()) walk(join(dir, entry.name), rel);
|
||||
else if (entry.isFile()) out[rel] = readFileSync(join(dir, entry.name), 'utf8');
|
||||
}
|
||||
}
|
||||
walk(root, '');
|
||||
return out;
|
||||
}
|
||||
|
||||
/** Canonical `skills/…` manifest key for a corpus-relative path. */
|
||||
const key = (rel) => `skills/${rel}`;
|
||||
|
||||
const REF_SMALL = `# Azure AI Foundry Overview
|
||||
|
||||
**Category:** Engineering
|
||||
**Status:** Active
|
||||
|
||||
---
|
||||
|
||||
## Overview
|
||||
|
||||
Foundry is a platform.
|
||||
|
||||
## Details
|
||||
|
||||
More detail here.
|
||||
`;
|
||||
|
||||
const LARGE_SECTIONS = Array.from({ length: 8 }, (_, i) =>
|
||||
`## Section ${i + 1}\n\n${'body paragraph line.\n'.repeat(14)}`,
|
||||
).join('\n');
|
||||
const REF_LARGE = `# Big Reference
|
||||
|
||||
**Category:** Engineering
|
||||
**Status:** Active
|
||||
|
||||
---
|
||||
|
||||
${LARGE_SECTIONS}`;
|
||||
|
||||
const REF_STALE_VERIFIED = `# Stale Verified File
|
||||
|
||||
**Category:** Engineering
|
||||
**Status:** Active
|
||||
**Verified:** MCP 2026-06
|
||||
|
||||
---
|
||||
|
||||
## Overview
|
||||
|
||||
Body.
|
||||
`;
|
||||
|
||||
test('planFileMutation — reference entry stamps Type + Source; body after --- byte-identical', () => {
|
||||
const entry = { type: 'reference', source: 'https://learn.microsoft.com/azure/ai-foundry' };
|
||||
const plan = planFileMutation(key('ms-ai-engineering/references/a.md'), REF_SMALL, entry);
|
||||
assert.equal(plan.changed, true);
|
||||
assert.equal(parseTypeHeader(plan.content), 'reference');
|
||||
assert.equal(parseSourceHeader(plan.content), entry.source);
|
||||
// Body from the first `---` rule onward is untouched (header-only insertion; small file → no TOC).
|
||||
const marker = '\n---\n';
|
||||
assert.equal(plan.content.slice(plan.content.indexOf(marker)), REF_SMALL.slice(REF_SMALL.indexOf(marker)));
|
||||
});
|
||||
|
||||
test('planFileMutation — deferred reference (no source) gets Type only, never Source', () => {
|
||||
const entry = { type: 'reference', deferred: true, deferReason: 'no-ms-citation' };
|
||||
const plan = planFileMutation(key('ms-ai-governance/references/d.md'), REF_SMALL, entry);
|
||||
assert.equal(parseTypeHeader(plan.content), 'reference');
|
||||
assert.equal(parseSourceHeader(plan.content), null);
|
||||
assert.equal(plan.expectSource, false);
|
||||
});
|
||||
|
||||
test('planFileMutation — throw-guard fires on an advisor target path', () => {
|
||||
const entry = { type: 'reference', source: 'https://learn.microsoft.com/x' };
|
||||
assert.throws(
|
||||
() => planFileMutation('skills/ms-ai-advisor/references/architecture/cost-models.md', REF_SMALL, entry),
|
||||
/advisor/i,
|
||||
);
|
||||
});
|
||||
|
||||
test('planFileMutation — normalize then insert applied in order on a **Verified:** MCP fixture', () => {
|
||||
const entry = { type: 'reference', source: 'https://learn.microsoft.com/azure/x' };
|
||||
const plan = planFileMutation(key('ms-ai-security/references/v.md'), REF_STALE_VERIFIED, entry);
|
||||
// Stale header Verified stripped (normalize ran) AND Type inserted (insert ran).
|
||||
assert.equal(parseVerifiedHeader(plan.content), null);
|
||||
assert.equal(parseTypeHeader(plan.content), 'reference');
|
||||
assert.equal(parseSourceHeader(plan.content), entry.source);
|
||||
});
|
||||
|
||||
test('planFileMutation — already-migrated content is a no-op (idempotent)', () => {
|
||||
const entry = { type: 'reference', source: 'https://learn.microsoft.com/azure/ai-foundry' };
|
||||
const once = planFileMutation(key('ms-ai-engineering/references/a.md'), REF_SMALL, entry);
|
||||
const twice = planFileMutation(key('ms-ai-engineering/references/a.md'), once.content, entry);
|
||||
assert.equal(twice.changed, false);
|
||||
assert.equal(twice.content, once.content);
|
||||
});
|
||||
|
||||
test('migrateCorpus — dry-run (default) leaves the corpus byte-identical on disk', () => {
|
||||
withTmp((tmp) => {
|
||||
const root = join(tmp, 'skills');
|
||||
makeCorpus(root, { 'ms-ai-engineering/references/a.md': REF_SMALL });
|
||||
const before = readAll(root);
|
||||
const manifest = { [key('ms-ai-engineering/references/a.md')]: { type: 'reference', source: 'https://learn.microsoft.com/x' } };
|
||||
const report = migrateCorpus({ root, backupRoot: join(tmp, '.kb-backup'), manifest, write: false });
|
||||
assert.equal(report.write, false);
|
||||
assert.ok(report.counts.changed > 0, 'dry-run still reports the pending change');
|
||||
assert.equal(report.backupPath, null);
|
||||
assert.deepEqual(readAll(root), before);
|
||||
});
|
||||
});
|
||||
|
||||
test('migrateCorpus — --write stamps non-advisor files and leaves advisor byte-identical', () => {
|
||||
withTmp((tmp) => {
|
||||
const root = join(tmp, 'skills');
|
||||
makeCorpus(root, {
|
||||
'ms-ai-engineering/references/a.md': REF_SMALL,
|
||||
'ms-ai-advisor/references/architecture/cost-models.md': REF_SMALL,
|
||||
});
|
||||
const advisorBefore = readFileSync(join(root, 'ms-ai-advisor/references/architecture/cost-models.md'), 'utf8');
|
||||
const manifest = {
|
||||
[key('ms-ai-engineering/references/a.md')]: { type: 'reference', source: 'https://learn.microsoft.com/x' },
|
||||
// Advisor entry is present in the real manifest (deferred), but the walker must never reach it.
|
||||
[key('ms-ai-advisor/references/architecture/cost-models.md')]: { type: 'reference', deferred: true, deferReason: 'advisor-scope' },
|
||||
};
|
||||
const report = migrateCorpus({ root, backupRoot: join(tmp, '.kb-backup'), manifest, write: true });
|
||||
assert.equal(report.counts.written, 1);
|
||||
// Non-advisor mutated.
|
||||
const eng = readFileSync(join(root, 'ms-ai-engineering/references/a.md'), 'utf8');
|
||||
assert.equal(parseTypeHeader(eng), 'reference');
|
||||
assert.equal(parseSourceHeader(eng), 'https://learn.microsoft.com/x');
|
||||
// Advisor untouched — byte-identical.
|
||||
assert.equal(readFileSync(join(root, 'ms-ai-advisor/references/architecture/cost-models.md'), 'utf8'), advisorBefore);
|
||||
});
|
||||
});
|
||||
|
||||
test('migrateCorpus — second --write run writes 0 files (idempotent on disk)', () => {
|
||||
withTmp((tmp) => {
|
||||
const root = join(tmp, 'skills');
|
||||
makeCorpus(root, { 'ms-ai-engineering/references/a.md': REF_SMALL });
|
||||
const manifest = { [key('ms-ai-engineering/references/a.md')]: { type: 'reference', source: 'https://learn.microsoft.com/x' } };
|
||||
const opts = { root, backupRoot: join(tmp, '.kb-backup'), manifest, write: true };
|
||||
const first = migrateCorpus(opts);
|
||||
assert.equal(first.counts.written, 1);
|
||||
const second = migrateCorpus(opts);
|
||||
assert.equal(second.counts.written, 0);
|
||||
assert.equal(second.counts.changed, 0);
|
||||
});
|
||||
});
|
||||
|
||||
test('migrateCorpus — atomic write leaves no *.tmp.* leftover', () => {
|
||||
withTmp((tmp) => {
|
||||
const root = join(tmp, 'skills');
|
||||
makeCorpus(root, { 'ms-ai-engineering/references/a.md': REF_SMALL });
|
||||
const manifest = { [key('ms-ai-engineering/references/a.md')]: { type: 'reference', source: 'https://learn.microsoft.com/x' } };
|
||||
migrateCorpus({ root, backupRoot: join(tmp, '.kb-backup'), manifest, write: true });
|
||||
const names = Object.keys(readAll(root));
|
||||
assert.equal(names.some((n) => /\.tmp\./.test(n)), false, 'no atomic-write temp file survived');
|
||||
});
|
||||
});
|
||||
|
||||
test('migrate-corpus source — uses atomicWriteSync, never bare writeFileSync', () => {
|
||||
const src = readFileSync(SRC, 'utf8');
|
||||
assert.match(src, /atomicWriteSync/, 'imports/uses the atomic writer');
|
||||
assert.doesNotMatch(src, /\bwriteFileSync\b/, 'never calls writeFileSync directly for corpus writes');
|
||||
assert.match(src, /ms-ai-advisor/, 'carries the advisor fence');
|
||||
});
|
||||
|
||||
test('migrateCorpus — backup exists and restore() round-trips byte-identity', () => {
|
||||
withTmp((tmp) => {
|
||||
const root = join(tmp, 'skills');
|
||||
makeCorpus(root, {
|
||||
'ms-ai-engineering/references/a.md': REF_SMALL,
|
||||
'ms-ai-security/references/b.md': REF_STALE_VERIFIED,
|
||||
});
|
||||
const original = readAll(root);
|
||||
const manifest = {
|
||||
[key('ms-ai-engineering/references/a.md')]: { type: 'reference', source: 'https://learn.microsoft.com/x' },
|
||||
[key('ms-ai-security/references/b.md')]: { type: 'reference', source: 'https://learn.microsoft.com/y' },
|
||||
};
|
||||
const report = migrateCorpus({ root, backupRoot: join(tmp, '.kb-backup'), manifest, write: true });
|
||||
assert.ok(report.backupPath && existsSync(report.backupPath), 'backup snapshot exists');
|
||||
assert.equal(typeof report.restore, 'function');
|
||||
// Corpus is mutated now…
|
||||
assert.notDeepEqual(readAll(root), original);
|
||||
// …restore brings it back byte-identical.
|
||||
report.restore();
|
||||
assert.deepEqual(readAll(root), original);
|
||||
});
|
||||
});
|
||||
|
||||
test('migrateCorpus — large file gains exactly one ## Innhold, sections preserved', () => {
|
||||
withTmp((tmp) => {
|
||||
const root = join(tmp, 'skills');
|
||||
makeCorpus(root, { 'ms-ai-engineering/references/big.md': REF_LARGE });
|
||||
const manifest = { [key('ms-ai-engineering/references/big.md')]: { type: 'reference', source: 'https://learn.microsoft.com/x' } };
|
||||
migrateCorpus({ root, backupRoot: join(tmp, '.kb-backup'), manifest, write: true });
|
||||
const out = readFileSync(join(root, 'ms-ai-engineering/references/big.md'), 'utf8');
|
||||
assert.equal((out.match(/^## Innhold$/gm) || []).length, 1, 'exactly one TOC heading');
|
||||
assert.equal(parseTypeHeader(out), 'reference');
|
||||
// All original sections survive.
|
||||
for (let i = 1; i <= 8; i++) assert.match(out, new RegExp(`^## Section ${i}$`, 'm'));
|
||||
});
|
||||
});
|
||||
|
||||
test('migrateCorpus — post-write parse assertion catches a deep-preamble (>500B) file', () => {
|
||||
withTmp((tmp) => {
|
||||
const root = join(tmp, 'skills');
|
||||
// A title long enough that the inserted **Type:** line lands past the 500-byte
|
||||
// scan window → parseTypeHeader returns null on re-read → the applier must throw.
|
||||
const deep = `# ${'A'.repeat(490)}\n\n**Category:** Engineering\n\n## Section\n\nBody.\n`;
|
||||
makeCorpus(root, { 'ms-ai-engineering/references/deep.md': deep });
|
||||
const original = readAll(root);
|
||||
const manifest = { [key('ms-ai-engineering/references/deep.md')]: { type: 'reference', source: 'https://learn.microsoft.com/x' } };
|
||||
assert.throws(
|
||||
() => migrateCorpus({ root, backupRoot: join(tmp, '.kb-backup'), manifest, write: true }),
|
||||
/truncat|500|header/i,
|
||||
);
|
||||
// The applier restored the corpus on the failed assertion — no partial mutation left behind.
|
||||
assert.deepEqual(readAll(root), original);
|
||||
});
|
||||
});
|
||||
Loading…
Add table
Add a link
Reference in a new issue