chore: import maquina-comercial-pessoal history via subtree

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

git-subtree-dir: maquina-comercial-pessoal
git-subtree-mainline: f5af038e19
git-subtree-split: a9ef4d776d
This commit is contained in:
Gustavo
2026-09-28 15:50:17 -03:00
co-authored by Claude Sonnet 5
150 changed files with 33449 additions and 0 deletions
@@ -0,0 +1,100 @@
import path from 'node:path';
import { loadDotEnv, paths } from '../env.mjs';
import { openDb } from '../db.mjs';
import { extractUf } from './rfb-extractor.mjs';
import { downloadRfbDataset, validateRfbRawFiles } from './rfb-downloader.mjs';
import { createJobIO } from '../jobs/protocol.mjs';
import { randomUUID } from 'node:crypto';
import { createJobLease } from '../jobs/lease.mjs';
// Processo filho de extracao RFB: le a UF da linha em extractions e roda a
// extracao completa, reportando progresso via stdout.
await loadDotEnv();
const args = new Map(
process.argv.slice(2).map((arg) => {
const [key, value = 'true'] = arg.replace(/^--/, '').split('=');
return [key, value];
})
);
const io = createJobIO();
async function main() {
const extractionId = Number(args.get('extraction-id'));
if (!extractionId) throw new Error('extract-job: --extraction-id obrigatorio');
const db = openDb();
let lease;
try {
const extraction = db.prepare('SELECT * FROM extractions WHERE id = ?').get(extractionId);
if (!extraction) throw new Error(`extraction #${extractionId} nao encontrada`);
const uf = extraction.uf;
lease = createJobLease({ db, table: 'extractions', id: extractionId, ownerToken: args.get('owner-token') || randomUUID() });
const outCsv = path.join(paths.cnpjDir, `estabelecimentos-${uf}.csv`);
const checkpointPath = path.join(paths.cnpjDir, `checkpoint-estabelecimentos-${uf}.ndjson`);
const download = await downloadRfbDataset({
rawDir: paths.rfbRaw,
isCancelled: io.isCancelled,
onProgress: (progress) => {
const mapped = { ...progress, pct: Math.round((progress.pct || 0) * 0.45) };
io.emit(mapped);
db.prepare('UPDATE extractions SET progress_json = ? WHERE id = ?')
.run(JSON.stringify(mapped), extractionId);
}
});
if (download.cancelled) {
db.prepare("UPDATE extractions SET status = 'cancelled', finished_at = datetime('now') WHERE id = ?").run(extractionId);
io.done({ cancelled: true, snapshot: download.snapshot, downloaded: download.downloaded.length });
return;
}
await validateRfbRawFiles(paths.rfbRaw);
const result = await extractUf({
uf,
rawDir: paths.rfbRaw,
outCsv,
checkpointPath,
db,
log: (m) => io.log(m),
isCancelled: io.isCancelled,
onProgress: (p) => {
const mapped = { ...p, pct: 45 + Math.round((p.pct || 0) * 0.55) };
io.emit(mapped);
db.prepare('UPDATE extractions SET progress_json = ? WHERE id = ?')
.run(JSON.stringify(mapped), extractionId);
}
});
if (result.cancelled) {
db.prepare("UPDATE extractions SET status = 'cancelled', finished_at = datetime('now') WHERE id = ?")
.run(extractionId);
io.done({ cancelled: true });
} else {
db.prepare(
`UPDATE extractions SET status = 'done', csv_path = ?, total_rows = ?, with_phone = ?,
municipios_count = ?, finished_at = datetime('now') WHERE id = ?`
).run(result.csvPath, result.totalRows, result.withPhone, result.municipiosCount, extractionId);
io.done({
snapshot: download.snapshot,
downloaded: download.downloaded.length,
skipped: download.skipped.length,
totalRows: result.totalRows,
withPhone: result.withPhone,
municipios: result.municipiosCount
});
}
} catch (error) {
db.prepare("UPDATE extractions SET status = 'error', error = ?, finished_at = datetime('now') WHERE id = ? AND status IN ('pending','running')")
.run(String(error.message || error), extractionId);
throw error;
} finally {
lease?.close();
db.close();
}
}
main().catch((error) => {
io.fail(error);
process.exit(1);
});