Self-hosted, Self-Improving, Federated Everything App — dev management, agent orchestration, persistent mind, digital twin, second brain, and creative pipeline from a single dashboard over Tailscale to port the user into the OS
JavaScript ai-agentsdashboarddeveloper-toolsdigital-twinquantified-selfsecond-brainself-hostedtailscale
/**
* Author personas — file-backed store (escape-hatch / test backend).
*
* Persists to data/authors.json (array, atomicWrite). PostgreSQL (db.js) is the
* default; this backend is reachable only via the MEMORY_BACKEND=file escape
* hatch or NODE_ENV=test — same posture as the memory backend. The dispatcher
* in index.js picks between this and db.js.
*
* All mutation semantics live in logic.js so this backend and the PG backend
* can't drift; this module only does load/find/persist.
*/
import { join } from 'path';
import { randomUUID } from 'crypto';
import { PATHS, readJSONFile, atomicWrite, ensureDir } from '../../lib/fileUtils.js';
import { ServerError } from '../../lib/errorHandler.js';
import { sanitizeAuthor, buildAuthorRecord, applyAuthorPatch, mergeAuthorRecord } from './logic.js';
import {
maybeJournalBeforeOverwrite, setSyncBaseHash, contentHashForRecord, flushBaseHashes, deleteSyncBaseHash,
withBaseHashFlushBatch,
} from '../../lib/conflictJournal.js';
const AUTHORS_FILE = join(PATHS.data, 'authors.json');
async function loadAll() {
// Authors are durable user records; never save an empty fallback over unreadable bytes.
const raw = await readJSONFile(AUTHORS_FILE, [], { strict: true });
return (Array.isArray(raw) ? raw : []).map(sanitizeAuthor).filter(Boolean);
}
async function saveAll(authors) {
await ensureDir(PATHS.data);
await atomicWrite(AUTHORS_FILE, authors);
}
const byName = (a, b) => (a.name || '').localeCompare(b.name || '');
export async function listAuthors({ includeDeleted = false } = {}) {
const all = await loadAll();
return (includeDeleted ? all : all.filter((a) => !a.deleted)).sort(byName);
}
export async function getAuthor(id, { includeDeleted = false } = {}) {
const all = await loadAll();
const found = all.find((a) => a.id === id);
if (!found) return null;
return includeDeleted || !found.deleted ? found : null;
}
/** Live author ids (or all when includeDeleted) — used by tombstone GC sweeps. */
export async function listAuthorIds({ includeDeleted = false } = {}) {
const all = await loadAll();
return (includeDeleted ? all : all.filter((a) => !a.deleted)).map((a) => a.id);
}
export async function createAuthor(input) {
const author = buildAuthorRecord(input, { id: `auth-${randomUUID()}`, now: new Date().toISOString() });
if (!author) throw new ServerError('Invalid author payload', { status: 400, code: 'VALIDATION' });
const all = await loadAll();
all.push(author);
await saveAll(all);
console.log(`✍️ Created author persona: ${author.id} (${author.name})`);
return author;
}
export async function updateAuthor(id, patch) {
const all = await loadAll();
const idx = all.findIndex((a) => a.id === id);
if (idx < 0 || all[idx].deleted) throw new ServerError('Author not found', { status: 404, code: 'NOT_FOUND' });
const next = applyAuthorPatch(all[idx], patch);
if (!next) throw new ServerError('Invalid author payload', { status: 400, code: 'VALIDATION' });
all[idx] = next;
await saveAll(all);
return next;
}
export async function deleteAuthor(id) {
const all = await loadAll();
const idx = all.findIndex((a) => a.id === id);
if (idx < 0 || all[idx].deleted) throw new ServerError('Author not found', { status: 404, code: 'NOT_FOUND' });
const now = new Date().toISOString();
all[idx] = { ...all[idx], deleted: true, deletedAt: now, updatedAt: now };
await saveAll(all);
return { id };
}
/**
* File-backend mirror of db.js `mergeAuthorsFromSync` — LWW-per-id (tombstone-
* aware) via the shared `mergeAuthorRecord` decision so the two backends can't
* drift. Single load → per-record merge → single save (the file is small and
* the app is single-instance, so one read-modify-write covers the batch).
*/
export async function mergeAuthorsFromSync(remoteAuthors, { source = { via: 'sync', peerId: null } } = {}) {
if (!Array.isArray(remoteAuthors)) return { applied: false, count: 0 };
const all = await loadAll();
const byId = new Map(all.map((a) => [a.id, a]));
let changed = 0;
for (const remote of remoteAuthors) {
const local = byId.get(remote?.id) || null;
const { next, inserted, remoteWins, changed: didChange } = mergeAuthorRecord(local, remote);
if (!next) continue;
if (inserted) {
byId.set(next.id, next);
await setSyncBaseHash('author', next.id, contentHashForRecord('author', next));
changed += 1;
continue;
}
// local wins, OR remote won but is byte-identical to local (already agree —
// nothing to journal or advance). See db.js for why authors skip the
// every-remoteWins journaling mediaCollections does.
if (!remoteWins || !didChange) continue;
await maybeJournalBeforeOverwrite({ kind: 'author', id: next.id, local, remote: next, source });
byId.set(next.id, next);
await setSyncBaseHash('author', next.id, contentHashForRecord('author', next));
changed += 1;
}
if (changed > 0) await saveAll([...byId.values()]);
await flushBaseHashes();
if (changed === 0) return { applied: false, count: 0 };
return { applied: true, count: changed };
}
/**
* Hard-remove tombstoned authors whose deletedAt is older than the cutoff.
* Mirrors db.js `pruneTombstonedAuthors`; evicts each pruned author's base hash.
*/
export async function pruneTombstonedAuthors(olderThanMs) {
if (!Number.isFinite(olderThanMs)) return { pruned: 0 };
const all = await loadAll();
const survivors = [];
const pruned = [];
for (const a of all) {
const ms = a.deleted ? Date.parse(a.deletedAt || '') : NaN;
if (a.deleted && Number.isFinite(ms) && ms < olderThanMs) pruned.push(a.id);
else survivors.push(a);
}
if (pruned.length === 0) return { pruned: 0 };
await saveAll(survivors);
// Coalesce the per-author base-hash evictions into ONE disk write.
await withBaseHashFlushBatch(async () => {
for (const id of pruned) await deleteSyncBaseHash('author', id);
});
return { pruned: pruned.length };
}