okr/tests/innboks-convert.test.mjs
2026-07-16 20:06:43 +02:00

140 lines
5.4 KiB
JavaScript

// innboks-convert.test.mjs
// Step 11 (SC format): konverterings-adaptere txt/md/docx/eml/pdf -> markdown.
// Verifiserer:
// - .txt/.md -> markdown alltid (zero-dep-sti, node:-builtins)
// - eksterne lenker NOEYTRALISERES ved konvertering (inline/ref-def/autolink/
// HTML-anker/bilde): lenketeksten bevares, scheme-maalet droppes -- slik at
// strict-gaten (B3/F2) aldri felles av legitim konvertert output
// - .docx/.eml -> heading-STRUKTUR bevart ('# '-linjer, atx -- IKKE byte-snapshot);
// .pdf -> tekst ekstrahert (flat, dokumentert v1-caveat M5)
// - convert-twice-identical: samme binaerfil konvertert 2x gir byte-identisk
// markdown (konverterings-determinismen idempotensen hviler paa)
// - ukjent extension (.xlsx) -> null + norsk notice, INGEN throw
// Binaer-adapterne kjoerer conditional { skip: !engines } (plugin kan vaere
// installert uten node_modules -- npm-deps er dokumentert prerequisite).
// Moenster: node:test conditional skip; fixtures under tests/fixtures/inbox-sample/.
import { test } from 'node:test';
import assert from 'node:assert/strict';
import { mkdtempSync, writeFileSync, rmSync } from 'node:fs';
import { tmpdir } from 'node:os';
import { join, dirname } from 'node:path';
import { fileURLToPath } from 'node:url';
import { convert } from '../lib/convert/index.mjs';
const ROOT = join(dirname(fileURLToPath(import.meta.url)), '..');
const SAMPLE = join(ROOT, 'tests', 'fixtures', 'inbox-sample');
// Er binaer-konverterings-motorene installert? (npm-deps = dokumentert prerequisite.)
async function enginesAvailable() {
try {
await import('mammoth');
await import('turndown');
await import('postal-mime');
await import('unpdf');
return true;
} catch {
return false;
}
}
const engines = await enginesAvailable();
// Temp-katalog for haandlagde input-filer. Ryddes alltid.
function withTmp(fn) {
const dir = mkdtempSync(join(tmpdir(), 'innboks-convert-'));
try {
return fn(dir);
} finally {
rmSync(dir, { recursive: true, force: true });
}
}
test('.txt konverteres alltid (zero-dep): innhold bevart som markdown', async () => {
const md = await convert(join(SAMPLE, 'dok-a.txt'));
assert.equal(typeof md, 'string');
assert.ok(md.length > 0);
});
test('.md passthrough: markdown-struktur bevart', async () => {
await withTmp(async (dir) => {
const p = join(dir, 'notat.md');
writeFileSync(p, '# Tittel\n\nAvsnitt.\n\n## Under\n\nMer.\n');
const md = await convert(p);
assert.match(md, /^# Tittel$/m);
assert.match(md, /^## Under$/m);
});
});
test('eksterne lenker noeytraliseres: inline/ref-def/autolink/HTML/bilde', async () => {
await withTmp(async (dir) => {
const p = join(dir, 'fiendtlig.txt');
writeFileSync(
p,
'# Notat\n\n'
+ 'Se [rapporten](https://evil.example/exfil) og [vedlegget][r].\n\n'
+ '[r]: https://evil.example/ref\n\n'
+ '<https://evil.example/auto>\n\n'
+ '<a href="file:///etc/passwd">passord</a>\n\n'
+ '![skjermbilde](https://evil.example/pixel.png)\n\n'
+ 'Intern relasjon: [Notat](/dokumenter/notat.md) beholdes.\n',
);
const md = await convert(p);
// Lenketekst bevart:
assert.match(md, /rapporten/);
assert.match(md, /vedlegget/);
assert.match(md, /passord/);
assert.match(md, /skjermbilde/);
// Ingen lenke-FORM med eksterne maal igjen (inline, ref-def, autolink, HTML):
assert.doesNotMatch(md, /\]\(https?:/);
assert.doesNotMatch(md, /^\s{0,3}\[[^\]]+\]:/m);
assert.doesNotMatch(md, /<https?:/);
assert.doesNotMatch(md, /<a\s/i);
assert.doesNotMatch(md, /!\[/);
assert.doesNotMatch(md, /file:\/\//);
// Trygg bundle-root-relativ .md-lenke beholdes (relasjons-formatet):
assert.match(md, /\[Notat\]\(\/dokumenter\/notat\.md\)/);
});
});
test('.docx -> markdown med atx-heading-struktur (ikke byte-snapshot)', { skip: !engines }, async () => {
const md = await convert(join(SAMPLE, 'dok.docx'));
assert.equal(typeof md, 'string');
// M5/F4: {headingStyle:'atx'} er pinnet -> '#'-headings, aldri setext.
assert.match(md, /^# /m);
assert.match(md, /^## /m);
assert.doesNotMatch(md, /^=+$/m);
});
test('.eml -> markdown: subject som heading + body-tekst (kun body, vedlegg = non-goal)', { skip: !engines }, async () => {
const md = await convert(join(SAMPLE, 'dok.eml'));
assert.match(md, /^# Statusoppdatering T2/m);
assert.match(md, /Fremdrift paa maaltallene/);
});
test('.pdf -> tekst ekstrahert (flat markdown, isEvalSupported:false-sti)', { skip: !engines }, async () => {
const md = await convert(join(SAMPLE, 'dok.pdf'));
assert.equal(typeof md, 'string');
assert.match(md, /Tildelingsbrev 2026/);
});
test('convert-twice-identical: docx og pdf gir byte-identisk markdown', { skip: !engines }, async () => {
for (const f of ['dok.docx', 'dok.pdf']) {
const first = await convert(join(SAMPLE, f));
const second = await convert(join(SAMPLE, f));
assert.equal(first, second, `ikke-deterministisk konvertering: ${f}`);
}
});
test('ukjent extension (.xlsx) -> null + norsk notice, ingen throw', async () => {
await withTmp(async (dir) => {
const p = join(dir, 'regneark.xlsx');
writeFileSync(p, 'ikke egentlig xlsx');
const notices = [];
const md = await convert(p, { onNotice: (msg) => notices.push(msg) });
assert.equal(md, null);
assert.equal(notices.length, 1);
assert.match(notices[0], /hopper over/);
assert.match(notices[0], /regneark\.xlsx/);
});
});