Files
accounted/extensions/general/enable-banking/lib/trigger-sync.ts
T
Mattsson b68c082ef5 feat(bank-sync): close the F2 report: gap backfill, consent and paused chip states, agent-triggered sync (#2165)
* fix(bank-sync): cron backfills the gap since the last successful sync

The daily incremental sync always asked the bank for the last 7 days. Any
pause longer than that (a lapsed subscription paid again, a consent renewed
after expiry, an outage) silently lost the days in between: the connection
came back, looked healthy, and the missing transactions never arrived.

The lookback now widens to cover the gap since last_synced_at plus one day
of overlap, capped at the 90-day PSD2 limit, and a gap of a month or more
asks for strategy=longest like the manual sync route does. Dedup via
external_id makes the overlap harmless. First syncs keep their 90-day path.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QdcDV7CngLkWUvfKSsxFhS

* feat(bank-sync): chip warns seven days before a bank consent expires

The transactions-page chip only reacted once a connection was already dead
(expired/error) or had gone stale. A consent that is about to end looked
healthy until the morning it stopped syncing. New "expiring" state when a
live connection's consent_expires is within seven days, the same threshold
as the consent-expiry email in the sync cron. Precedence: attention,
expiring, stale, healthy.

getChipState moves to lib/transactions/bank-sync-chip-state.ts so the
precedence is unit-tested; the component keeps the rendering only.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QdcDV7CngLkWUvfKSsxFhS

* feat(bank-sync): chip says paused when the subscription lapsed

The daily cron filters connections by the bank_sync capability, so a
company whose trial or subscription ended keeps status=active rows with a
frozen last_synced_at. The chip read that as "stale, check the connection",
which sends the user to re-authorise a connection that is perfectly alive.
56 of 191 active connections on prod were in this state on 2026-09-01.

New "paused" state, ranked above everything else, when the company lacks
bank_sync: hosted points at billing, self-host at the connector key, the
same split BankSyncNowButton already makes. getChipState takes an options
object so the clock stays out of render (react-hooks/purity).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QdcDV7CngLkWUvfKSsxFhS

* feat(api): agent-triggerable bank sync in v1 and MCP

Closes the first wish in the F2 report: an integration could read bank
data but never refresh it. New POST /api/v1/companies/{id}/bank-connections/
{connectionId}/sync and MCP gnubok_sync_bank, both on a shared runner
(extensions/general/enable-banking/lib/trigger-sync.ts).

Cost is bounded structurally, not by policy: the window is never
caller-controlled (the cron's gap-aware 7 to 90 day lookback), a connection
synced within 15 minutes answers BANK_SYNC_COOLDOWN with next_allowed_at
(429 + Retry-After on v1; synced=false in-band on MCP so the agent reads on
instead of retrying), and a failing connection is throttled per process by
attempt time. A dead session is flipped to expired with a remediation that
hands the user the connect link: no API call revives a consent.

Gated on bank_sync like gnubok_connect_bank; scope transactions:write.
Registry, scope map, load-routes, spec snapshot and the generated
accounted-api skill updated; five BANK_SYNC_* / BANK_SESSION_EXPIRED codes
added to the structured-error registry. The web Synka-nu route is left as
is (see DECISIONS.md).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QdcDV7CngLkWUvfKSsxFhS

* test(bank-sync): use the options object in the remaining chip-state calls

Four multi-line calls still passed the clock positionally after
getChipState moved to an options object; tsc flagged them (vitest did not,
the extra argument was ignored at runtime).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QdcDV7CngLkWUvfKSsxFhS

* fix(api): address skeptic findings on the agent-triggered bank sync

Three refutations from the pre-publish skeptic pass:

1. Core imported the extension. The v1 sync route pulled the runner
   straight from @/extensions, which the core-build gate rejects and which
   left a live bank endpoint on zero-extension builds. The route now
   resolves it through the registry's services channel against a contract
   in lib/bank-sync/trigger-sync-contract.ts (same pattern as the
   Skatteverket read service) and answers EXTENSION_DISABLED when the
   extension is absent.

2. The idempotency cache stored the handler-level 429. A same-key retry
   after Retry-After, which is the documented retry, replayed the stale
   cooldown as a 400 for the cache's 24-hour TTL. withApiV1 no longer
   caches 429 responses; regression test added. The endpoint's pitfall no
   longer claims Idempotency-Key is mandatory (it was never enforced).

3. Two cron tests read the clock twice and failed whenever a millisecond
   passed between the reads. They now pin the clock with fake timers.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QdcDV7CngLkWUvfKSsxFhS

* fix(bank-sync): durable cooldown lease and review wording

Resolves the PR #2165 review findings in one pass.

Superagent P1: the attempt throttle was a process-local Map, so two agent
calls on different serverless instances (or a retry after a cold start on
a failing connection) could each bill an Enable Banking call, contradicting
the one-sync-per-15-minutes promise. New bank_connections.sync_lease_until
(migration 20260902150000), claimed with one conditional UPDATE before the
bank is called; Postgres row locking makes exactly one claimer win, the
rest answer BANK_SYNC_COOLDOWN. The lease stays for the full window on
success and failure. Tests cover the claim order, a failed attempt seen
from a second instance, a lost race, and an expired lease.

CodeRabbit: the =1 plural branch now reads "in 1 day" / "om 1 dag"
(daysUntilConsentExpiry rounds a partial day up, so "tomorrow" could be
today); the cooldown pitfall on the v1 endpoint, the MCP description and
the in-band cooldown instruction now say a cooldown can follow a failed
attempt and tell the agent to compare last_synced_at before deciding.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0125TMQQBjBBZG9YxP7wQWub

* fix(bank-sync): lease claim as a literal filter for the schema guard

CI's no-phantom-columns guard counts runtime-built query expressions and
its ceiling is exact; the templated `.or('sync_lease_until.is.null,...')`
claim added one. The column now defaults to epoch (NOT NULL), so "never
claimed" is just "expired long ago" and the atomic claim is a single
literal `.lte('sync_lease_until', now)` the guard can check. Migration is
unshipped (same PR), so it is edited in place.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0125TMQQBjBBZG9YxP7wQWub

* fix(bank-sync): runner verifies company membership before the lease

Superagent (round 3): the MCP path reached the shared runner without a
membership check of its own. Both callers do enforce it upstream
(withApiV1's company resolution and resolveMcpCompanyContext in the MCP
dispatcher), but the runner writes transactions and bills a bank call, so
it now checks company_members itself, before the cooldown and the lease
claim, and answers NOT_FOUND for a non-member. The viewer check that was
buried inside the sync block moves up with it.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0125TMQQBjBBZG9YxP7wQWub

---------

Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
2026-09-02 17:17:41 +02:00

298 lines
11 KiB
TypeScript

/**
* Agent-triggered bank sync: the shared runner behind the v1 REST endpoint
* POST /companies/{id}/bank-connections/{connectionId}/sync and the MCP tool
* gnubok_sync_bank.
*
* Deliberately narrower than the cookie-session "Synka nu" route in
* index.ts: the window is never caller-controlled (the gap-aware incremental
* lookback from cron-lookback.ts, 7 to 90 days), and a connection that
* synced OR was attempted within SYNC_COOLDOWN_MS answers with a cooldown
* instead of another paid Enable Banking round-trip. The attempt guard is a
* durable lease on bank_connections.sync_lease_until, claimed with one
* conditional UPDATE, so it holds across serverless instances and cold
* starts. An unattended agent loop therefore costs at most one sync per
* connection per cooldown window, regardless of how often it asks.
*
* Failures are reported as codes, never thrown, so each surface maps them
* to its own envelope (structured-errors.ts BANK_SYNC_*). A dead PSD2
* session is flipped to 'expired' here exactly like the web route does:
* nothing an API call can do revives it, only BankID in a browser.
*
* Core cannot import this module (CI guard): the v1 route reaches it via
* the extension's registered `services`, against the contract in
* lib/bank-sync/trigger-sync-contract.ts.
*/
import type { SupabaseClient } from '@supabase/supabase-js'
import { syncAccountTransactions, type SyncOptions } from './sync'
import {
SessionExpiredError,
REAUTH_REQUIRED_MESSAGE,
SYNC_FAILED_MESSAGE,
} from './api-client'
import { incrementalLookbackDays } from './cron-lookback'
import { updateBalancesFromSync } from '@/lib/cash-accounts/service'
import { eventBus } from '@/lib/events/bus'
import {
SYNC_COOLDOWN_MS,
type TriggerSyncInput,
type TriggerSyncResult,
} from '@/lib/bank-sync/trigger-sync-contract'
import type { StoredAccount } from '../types'
import type { Transaction } from '@/types'
export { SYNC_COOLDOWN_MS }
export type { TriggerSyncInput, TriggerSyncResult }
function isUuid(value: string): boolean {
return /^[0-9a-f]{8}-[0-9a-f]{4}-[1-5][0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$/i.test(value)
}
function cooldownResult(connectionId: string, nextAllowed: number, now: number): TriggerSyncResult {
return {
ok: false,
code: 'BANK_SYNC_COOLDOWN',
connection_id: connectionId,
next_allowed_at: new Date(nextAllowed).toISOString(),
retry_after_seconds: Math.max(1, Math.ceil((nextAllowed - now) / 1000)),
}
}
export async function triggerConnectionSync(
supabase: SupabaseClient,
input: TriggerSyncInput,
): Promise<TriggerSyncResult> {
const { companyId, userId, connectionId, log } = input
const now = input.now ?? Date.now()
if (!isUuid(connectionId)) {
return { ok: false, code: 'NOT_FOUND', connection_id: connectionId }
}
const { data: connection, error: connectionError } = await supabase
.from('bank_connections')
.select(
'id, company_id, bank_name, status, accounts_data, last_synced_at, error_message, sync_lease_until',
)
.eq('id', connectionId)
.eq('company_id', companyId)
.maybeSingle()
if (connectionError) throw connectionError
if (!connection) {
return { ok: false, code: 'NOT_FOUND', connection_id: connectionId }
}
// 'error' is retryable (a transient upstream failure parks the row there
// while the session is alive); 'expired' and the pending states are not:
// they need the browser flow.
if (connection.status !== 'active' && connection.status !== 'error') {
return {
ok: false,
code: 'BANK_SYNC_NOT_ACTIVE',
connection_id: connectionId,
status: connection.status,
}
}
// Membership is enforced by both callers before we run (withApiV1's
// company resolution, the MCP dispatcher's resolveMcpCompanyContext), but
// this runner writes transactions and bills a bank call, so it checks the
// caller's membership itself as well: a service-role client with the
// wrong userId must never get past this point. Viewers get raw inserts
// only, exactly like the web route.
const { data: membership, error: membershipError } = await supabase
.from('company_members')
.select('role')
.eq('company_id', companyId)
.eq('user_id', userId)
.maybeSingle()
if (membershipError) throw membershipError
if (!membership) {
return { ok: false, code: 'NOT_FOUND', connection_id: connectionId }
}
const isViewer = (membership as { role?: string }).role === 'viewer'
// A successful sync (ours, the web button's or the cron's) within the
// window: the data is fresh, say so without touching the bank.
const lastSynced = connection.last_synced_at
? new Date(connection.last_synced_at as string).getTime()
: null
if (lastSynced !== null && now - lastSynced < SYNC_COOLDOWN_MS) {
return cooldownResult(connectionId, lastSynced + SYNC_COOLDOWN_MS, now)
}
// A lease still held from a recent ATTEMPT (success or failure): cheap
// read-side answer before the write below.
const heldLease = connection.sync_lease_until
? new Date(connection.sync_lease_until as string).getTime()
: null
if (heldLease !== null && heldLease > now) {
return cooldownResult(connectionId, heldLease, now)
}
const allAccounts = ((connection.accounts_data as StoredAccount[] | null) ?? []).map((a) => ({
...a,
}))
const accounts = allAccounts.filter((a) => a.enabled !== false)
if (accounts.length === 0) {
return { ok: false, code: 'BANK_SYNC_NO_ACCOUNTS', connection_id: connectionId }
}
// Durable, atomic cooldown claim. One conditional UPDATE: the lease is
// taken only if the current one has expired (the column defaults to epoch,
// so "never claimed" needs no NULL branch), and Postgres row locking
// serialises concurrent claimers, so two agent calls landing on different
// serverless instances (or retries of a failing connection after a cold
// start) can never both reach the bank. The lease stays for the full
// window whether the sync succeeds or fails: that IS the throttle.
const nowIso = new Date(now).toISOString()
const leaseUntil = now + SYNC_COOLDOWN_MS
const { data: claimed, error: claimError } = await supabase
.from('bank_connections')
.update({ sync_lease_until: new Date(leaseUntil).toISOString() })
.eq('id', connectionId)
.eq('company_id', companyId)
.lte('sync_lease_until', nowIso)
.select('id')
if (claimError) throw claimError
if (!claimed || claimed.length === 0) {
// Lost the race: another caller claimed between our read and this write.
// Its lease started at most a moment ago, so ours is the honest estimate.
log.info('agent-triggered bank sync: lease held by a concurrent caller', { connectionId })
return cooldownResult(connectionId, leaseUntil, now)
}
const lookbackDays = incrementalLookbackDays(connection.last_synced_at as string | null, now)
const toDate = new Date(now).toISOString().split('T')[0]
const fromDate = new Date(now - lookbackDays * 24 * 60 * 60 * 1000).toISOString().split('T')[0]
const syncStartedAt = new Date(now).toISOString()
try {
// Same SIE-overlap guard as the web route and the cron: never
// auto-categorise into a range a completed SIE import already covers.
const { data: sieOverlap } = await supabase
.from('sie_imports')
.select('id')
.eq('company_id', companyId)
.eq('status', 'completed')
.gte('fiscal_year_end', fromDate)
.limit(1)
.maybeSingle()
const syncOptions: SyncOptions = {
...(sieOverlap ? { skipAutoCategorization: true } : {}),
...(isViewer ? { rawInsertOnly: true } : {}),
...(lookbackDays >= 30 ? { strategy: 'longest' as const } : {}),
}
const results = await Promise.all(
accounts.map((account) =>
syncAccountTransactions(
supabase,
companyId,
userId,
connection.id as string,
account,
fromDate,
toDate,
undefined,
syncOptions,
),
),
)
const imported = results.reduce((sum, r) => sum + r.imported, 0)
const duplicates = results.reduce((sum, r) => sum + r.duplicates, 0)
const syncedAt = new Date().toISOString()
await updateBalancesFromSync(
supabase,
companyId,
connection.id as string,
allAccounts.map((a) => ({
external_uid: a.uid,
balance: a.balance,
available_balance: a.available_balance,
balance_updated_at: a.balance_updated_at,
})),
)
await supabase
.from('bank_connections')
.update({
accounts_data: allAccounts,
last_synced_at: syncedAt,
...(connection.status === 'error' ? { status: 'active' } : {}),
...(connection.status === 'error' || connection.error_message ? { error_message: null } : {}),
})
.eq('id', connection.id)
.eq('company_id', companyId)
if (imported > 0) {
const { data: syncedTransactions } = await supabase
.from('transactions')
.select('*')
.eq('company_id', companyId)
.eq('bank_connection_id', connection.id)
.gte('created_at', syncStartedAt)
.order('created_at', { ascending: false })
.limit(imported)
if (syncedTransactions && syncedTransactions.length > 0) {
await eventBus.emit({
type: 'transaction.synced',
payload: { transactions: syncedTransactions as Transaction[], userId, companyId },
})
}
}
log.info('agent-triggered bank sync completed', {
connectionId,
imported,
duplicates,
lookbackDays,
})
return {
ok: true,
connection_id: connection.id as string,
bank: (connection.bank_name as string | null) ?? null,
imported,
duplicates,
from_date: fromDate,
to_date: toDate,
last_synced_at: syncedAt,
}
} catch (error) {
if (error instanceof SessionExpiredError) {
log.warn('agent-triggered bank sync: session expired', { connectionId })
await supabase
.from('bank_connections')
.update({ status: 'expired', error_message: REAUTH_REQUIRED_MESSAGE })
.eq('id', connection.id)
.eq('company_id', companyId)
return {
ok: false,
code: 'BANK_SESSION_EXPIRED',
connection_id: connectionId,
status: 'expired',
}
}
log.error('agent-triggered bank sync failed', {
connectionId,
message: error instanceof Error ? error.message : String(error),
name: error instanceof Error ? error.name : undefined,
})
if (connection.status === 'error') {
await supabase
.from('bank_connections')
.update({ error_message: SYNC_FAILED_MESSAGE })
.eq('id', connection.id)
.eq('company_id', companyId)
}
return {
ok: false,
code: 'BANK_SYNC_FAILED',
connection_id: connectionId,
status: connection.status as string,
}
}
}