Files
accounted/lib/invoices/complete-invoice-rows.ts
T
6906bc4aa2 feat(compliance): InvoiceRowsCompleted behandlingshistorik event for migrated invoice rows (#2312) (#2357)
Every migrated sales invoice whose rows complete_invoice_rows writes, from
the migration wizard or the hourly row-completion pass, now leaves one
InvoiceRowsCompleted row in processing_history on a new Invoice aggregate:
the writer, the provider, the consent, the row count, and the header VAT
split before and after when the pass rewrote it (BFL 5 kap 11 §, BFNAR
2013:2 p. 9.16). One run shares one correlation id.

lib/invoices/complete-invoice-rows.ts is the one TypeScript call site for
the RPC and the one emitter: it appends only on wrote = true, records
nothing for already_filled or failed, and keeps the append best-effort
(logged, eventId null) like every other processing_history writer. The
wizard runs on the user's session client, so MigrationOptions takes a lazy
createHistoryClient for the service role. Invoice numbers stay out of the
payload (the personnummer guard would drop ten-digit ones).

Migration 20260906210100 widens the aggregate_type CHECK with Invoice and
registers the event type; pg test covers the catalog row, the aggregate,
and that the CHECK still refuses unknown aggregates.


Claude-Session: https://claude.ai/code/session_01LvMaHcTnwAfxzgYD1fGYX1

Co-authored-by: Jakob Wennberg <311770904+jakobwennberg-oss@users.noreply.github.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
2026-09-06 21:04:53 +02:00

190 lines
7.5 KiB
TypeScript

/**
* completeInvoiceRows: the one TypeScript call site for the
* complete_invoice_rows RPC (migration 20260906135730), and the one place the
* behandlingshistorik event for a completed migrated invoice is written.
*
* Two writers put rows under migrated sales invoices: the migration wizard
* (extensions/general/arcim-migration/lib/migration-orchestrator.ts, rows
* written milliseconds after the header) and the hourly row-completion pass
* (complete-invoice-lines.ts, the rows the wizard's hydration budget did not
* reach, plus the header VAT split when the stored one held no evidence).
* The RPC already gives them one write path; this wrapper gives them one
* trail. Every invoice whose rows the RPC wrote gets one InvoiceRowsCompleted
* event (BFL 5 kap 11 §, BFNAR 2013:2 p. 9.16: the behandlingshistorik has to
* say what was processed automatically, when, and by what) naming the writer,
* the provider the rows came from, the row count and, when the header split
* was rewritten, the split before and after. The pass writes no
* bokföringspost, so BFL 5 kap 5 § (rättelse) does not bind it; the migration
* that wrote the invoice header records no event of its own, so this event
* is what lets the two writers reconcile per invoice.
*
* Failure semantics: the event is written only after the RPC has answered
* wrote = true, so an invoice whose completion failed, or that another writer
* had already filled, gets no event. The append itself is best-effort, the
* convention every processing_history writer follows (the rows are
* committed; a missing change-log row is logged, not turned into a failed
* invoice that the next run would try again and find full). The caller sees
* a null eventId when that happened.
*
* PII boundary: the payload carries UUIDs, counts, amounts and enum strings
* only. Invoice numbers and provider document numbers are deliberately left
* out: a ten-digit number (2026090001) trips the personnummer guard in
* appendProcessingHistory, which would lose the event for exactly that
* invoice. The invoice id is the reference; its number is on the row.
*/
import type { SupabaseClient } from '@supabase/supabase-js'
import type { ProcessingHistoryActor } from '@/types'
import { createLogger } from '@/lib/logger'
import {
appendProcessingHistoryWithClient,
type ProcessingHistoryEventType,
} from '@/lib/processing-history/append'
const log = createLogger('invoices/complete-invoice-rows')
/** Registered in processing_event_types by migration 20260906210100. */
export const INVOICE_ROWS_COMPLETED_EVENT = 'InvoiceRowsCompleted' satisfies ProcessingHistoryEventType
type RpcClient = Pick<SupabaseClient, 'rpc'>
type HistoryClient = Pick<SupabaseClient, 'from'>
/** The six invoice columns the RPC may rewrite: all present or none. */
export interface InvoiceHeaderVatSplit {
subtotal: number
subtotal_sek: number | null
vat_amount: number
vat_amount_sek: number | null
vat_rate: number | null
vat_treatment: string
}
/**
* What the trail records of a header split. The SEK twins are left out:
* they are derived from these and the exchange rate the row already carries.
*/
export interface InvoiceHeaderVatSnapshot {
subtotal: number | null
vat_amount: number | null
vat_rate: number | null
vat_treatment: string | null
}
export interface CompleteInvoiceRowsTrail {
/** Which writer: 'migration-wizard' or 'complete-invoice-lines'. */
source: string
/** The provider the rows came from ('fortnox', 'briox', ...). */
provider: string
/** The provider consent the rows were fetched under. */
consentId: string
/** One id per run, shared by every invoice the run completed. */
correlationId: string
actor: ProcessingHistoryActor
}
export interface CompleteInvoiceRowsInput {
companyId: string
invoiceId: string
/** The invoice_items columns per row; the RPC stamps invoice_id itself. */
rows: Record<string, unknown>[]
/** The header split to apply in the same transaction, or null to leave the header alone. */
header?: InvoiceHeaderVatSplit | null
/** The stored split before the write; recorded beside the new one when the header is rewritten. */
headerBefore?: InvoiceHeaderVatSnapshot | null
trail: CompleteInvoiceRowsTrail
/**
* The client the event row is written with. processing_history has no
* INSERT policy, so this must be a service-role client: the cron's own
* client already is, the wizard (on the user's session client) passes one.
*/
historyClient: HistoryClient
}
export type CompleteInvoiceRowsResult =
/** The rows (and the header, when one was given) landed; eventId is null when the trail append failed. */
| { status: 'written'; rows: number; headerUpdated: boolean; eventId: string | null }
/** Another writer filled the invoice first; nothing was written and nothing is recorded. */
| { status: 'already_filled' }
/** The RPC errored or refused (its code, or the Postgres message). */
| { status: 'failed'; reason: string }
/** What complete_invoice_rows returns (migration 20260906135730). */
interface CompleteRowsRpcOutcome {
ok: boolean
code?: string
wrote?: boolean
rows?: number
header_updated?: boolean
}
function snapshotOf(split: InvoiceHeaderVatSnapshot | InvoiceHeaderVatSplit | null | undefined): InvoiceHeaderVatSnapshot | null {
if (!split) return null
return {
subtotal: split.subtotal,
vat_amount: split.vat_amount,
vat_rate: split.vat_rate,
vat_treatment: split.vat_treatment,
}
}
export async function completeInvoiceRows(
supabase: RpcClient,
input: CompleteInvoiceRowsInput,
): Promise<CompleteInvoiceRowsResult> {
const header = input.header ?? null
const { data, error } = await supabase.rpc('complete_invoice_rows', {
p_company_id: input.companyId,
p_invoice_id: input.invoiceId,
p_rows: input.rows,
p_header: header,
})
const outcome = (data ?? null) as CompleteRowsRpcOutcome | null
if (error || !outcome?.ok) {
return { status: 'failed', reason: error?.message ?? outcome?.code ?? 'empty RPC response' }
}
if (!outcome.wrote) return { status: 'already_filled' }
const rows = outcome.rows ?? input.rows.length
const headerUpdated = outcome.header_updated === true
const eventId = await appendCompletedEvent(input, rows, headerUpdated)
return { status: 'written', rows, headerUpdated, eventId }
}
async function appendCompletedEvent(
input: CompleteInvoiceRowsInput,
rows: number,
headerUpdated: boolean,
): Promise<string | null> {
const { trail } = input
try {
return await appendProcessingHistoryWithClient(input.historyClient, {
companyId: input.companyId,
correlationId: trail.correlationId,
aggregateType: 'Invoice',
aggregateId: input.invoiceId,
eventType: INVOICE_ROWS_COMPLETED_EVENT,
payload: {
source: trail.source,
provider: trail.provider,
consent_id: trail.consentId,
rows,
header_updated: headerUpdated,
header_before: headerUpdated ? snapshotOf(input.headerBefore) : null,
header_after: headerUpdated ? snapshotOf(input.header) : null,
},
actor: trail.actor,
occurredAt: new Date(),
})
} catch (err) {
// The rows are committed; the trail row is what is missing. Logged so
// the gap is visible, never a reason to report the invoice as failed.
log.error('InvoiceRowsCompleted append failed; rows written without their behandlingshistorik row', {
companyId: input.companyId,
invoiceId: input.invoiceId,
source: trail.source,
error: err instanceof Error ? err.message : String(err),
})
return null
}
}