* fix(db): reconcile prod-orphaned migrations blocking Supabase branching
Prod's schema_migrations carries three versions with no committed file on
main, leaving the default Supabase branch in MIGRATIONS_FAILED and stopping
preview branches from being created:
20260707113729 add_transactions_enrichment (adopted from #927)
20260708120000 ledger_stats_committed_at_lag (adopted from #935)
20260708130000 ledger_deep_context (adopted from #935)
Adopt the byte-identical SQL under the exact apply-time versions, plus the
matching pg-tests and fixtures for the two RPCs so pg-real stays green:
20260708120000 switches get_ledger_usage_stats' median_booking_lag_days to
committed_at, so the existing test now asserts the new behavior. Idempotent
(ADD COLUMN IF NOT EXISTS / CREATE OR REPLACE FUNCTION): no-op on prod,
clean on fresh replays, no-op on #927/#935's next rebase. The knowledge-page
UI/lib/i18n stay in #935.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(deps): pin @anthropic-ai/bedrock-sdk to 0.29.1
0.32.0 (grouped dependabot bump #884) broke Bedrock streaming in prod: empty stream / "request ended without sending any chunks", taking down the in-app AI assistant and invoice OCR. Local dev ran the stale 0.29.1 in node_modules, so it only failed on deploys built fresh from the lockfile. Revert to the six-week-stable 0.29.1; creds/region were never the cause (proven AKIA key + eu-west-1).
Guard against an accidental re-bump three ways: exact pin (no caret), a dependabot ignore, and a pinned-dep check in scripts/checks/no-new-antipatterns.mjs (check:guards). Unpin only once 0.32.x streaming is verified against Bedrock. See DECISIONS.md.
Found by the switch-on check (calling gnubok_get_agent_briefing on real
prod data): counterparty patterns for reverse-charge foreign SaaS
(Google/ngrok/Supabase) reported dominant_account 2614 (reverse-charge
output VAT) instead of 5420 (software expense). The dominant_account CTE in
get_ledger_usage_stats excluded only 19xx, so on a reverse-charge booking
(expense + 2645 + 2614 + 1930) the three non-bank accounts tie at equal
counts and the account_number ascending tiebreak picks the low VAT number.
Migration 20260708110000 CREATE OR REPLACEs the function to also exclude
26xx (always moms in BAS, never characterizes a counterparty). Loan/tax
counterparties booking to 23xx/24xx/25xx/27xx stay eligible. supplier_patterns
is unaffected (it aggregates supplier_invoice_items.account_number, expense
only). Regression pg test asserts 5420 over 2614 and was confirmed to fail on
the old function.
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* feat(mcp): ledger-context resource with per-company booking patterns
Adds Accounted://ledger/context: derived account usage, counterparty
booking patterns with explicit confidence share (0.7 floor), explicit
mapping rules kept separate as authoritative, observed VAT profile, and
conventions. Backed by a SECURITY INVOKER get_ledger_usage_stats RPC so
group-bys run SQL-side, and surfaced as a top-5 digest stanza on
gnubok_get_agent_briefing so one call still bootstraps a session.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* feat(mcp): fold source-quality prereqs into the ledger-context RPC
Merchant-name normalization at the aggregation path (the splinter fix):
new normalize_counterparty_key() SQL function mirroring
normalizeCounterpartyName() so KORTKÖP/SWISH/date-suffixed labels merge
into one counterparty key, which also makes the categorization_templates
join exact. New supplier_patterns section (per-supplier dominant expense
account + VAT treatment from supplier invoices; credit notes and reversed
invoices excluded). account_usage excludes storno lines (they re-inflate
the account a correction moved away from); the counterparty CTE keeps
corrections because the transaction relink self-heals. Pattern confidence
is now count-grounded evidence {seen_12m, agree, share, last_booked}
instead of a bare ratio, and the digest frames it as historical frequency,
never auto-book permission.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(agent-context): use roundOre for the share ratio (antipattern ratchet)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(agent-context): defensive storno filter on counterparty CTE, fail-loud secondary reads
Review follow-ups: the counterparty CTE now excludes source_type='storno'
defensively (no live code path links a transaction to a storno, but legacy
rows may predate reverseEntry's unlink; a linked storno would count the
reversed category as precedent). Corrections stay included: they are the
live booking after relink. Secondary reads (rules, templates, settings)
now throw instead of silently reading as empty data: an agent must never
be told 'no rules' when the truth is 'read failed'.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
