feat(statements+analytics): normalise statement_type; fix analytics scoping

Groundwork for importing bank and loan statements alongside credit cards.

statement_type was whatever free text Gemini put in account_type ('Credit Card',
'credit card', 'credit_card', 'Business Card', 'ACCESS ADVANTAGE',
'multi-currency account'). The UI coped only by doing .includes("card"), which
breaks as soon as bank and loan statements arrive.

- Migration 0013 adds normalize_statement_type() + a BEFORE INSERT/UPDATE
  trigger and a CHECK constraint over credit_card|transaction|savings|loan|
  offset|investment|other. The trigger means the N8N workflow keeps working
  unchanged while it still sends free text. Raw value stays in account_type.
  Backfilled 99 existing rows.
- src/lib/statement-types.ts mirrors the vocabulary for the UI; statements page
  now filters by the real types and headlines balance vs amount due per type.

Analytics were scoped with INNER JOIN statements + s.owner_id, which silently
dropped all 180 manual/CSV transactions (statement_id IS NULL) from every
report. Switched all six routes to LEFT JOIN + COALESCE(t.owner_id, s.owner_id)
via shared fragments in src/lib/analytics-sql.ts, so the transfers/investment
exclusion that stops card-payment double counting stays consistent. Also
extended that exclusion to trip analytics, which had none.

Drive-by: /api/analytics/subscriptions was returning 500 on an unserialisable
BigInt from COUNT(*) + 1.

Verified against the live DB: monthly spend picks up the previously invisible
manual transactions (Apr 9,849.91 -> 14,286.70) and all four analytics
endpoints return 200.

prisma/migrations/0013_statement_type_normalization/migration.sql

@@ -0,0 +1,99 @@
+-- Normalise statements.statement_type to a fixed vocabulary.
+--
+-- Before this migration the column held whatever free text Gemini put in
+-- `account_type` ('Credit Card', 'credit card', 'credit_card', 'Business Card',
+-- 'ACCESS ADVANTAGE', 'multi-currency account', ...). The UI coped only by
+-- doing `.includes("card")`, which breaks as soon as bank and loan statements
+-- arrive.
+--
+-- The raw extracted text is preserved in `account_type` — this only touches
+-- `statement_type`. A BEFORE trigger normalises on write, so the N8N ingestion
+-- workflow keeps working unchanged while it still sends free text.
+--
+-- Idempotent: safe to re-run.
+
+CREATE OR REPLACE FUNCTION normalize_statement_type(raw TEXT)
+RETURNS TEXT AS $$
+DECLARE
+  v TEXT := lower(trim(coalesce(raw, '')));
+BEGIN
+  IF v = '' THEN
+    RETURN 'other';
+  END IF;
+
+  -- Already canonical
+  IF v IN ('credit_card', 'transaction', 'savings', 'loan', 'offset', 'investment', 'other') THEN
+    RETURN v;
+  END IF;
+
+  -- Offset before loan: "Mortgage Offset" is an offset account, not a loan.
+  IF v LIKE '%offset%' THEN
+    RETURN 'offset';
+  END IF;
+
+  -- Loans before transaction: "home loan account" must not match '%account%'.
+  IF v LIKE '%loan%' OR v LIKE '%mortgage%' THEN
+    RETURN 'loan';
+  END IF;
+
+  -- Cards: covers 'Credit Card', 'credit card', 'Business Card', 'Charge Card'
+  IF v LIKE '%card%' THEN
+    RETURN 'credit_card';
+  END IF;
+
+  IF v LIKE '%saving%' OR v LIKE '%term deposit%' THEN
+    RETURN 'savings';
+  END IF;
+
+  IF v LIKE '%invest%' OR v LIKE '%share%' OR v LIKE '%broker%' THEN
+    RETURN 'investment';
+  END IF;
+
+  -- Everyday transaction accounts. Bank-specific product names go here; the
+  -- generic keywords catch the rest.
+  IF v LIKE '%access advantage%'      -- ANZ
+     OR v LIKE '%complete access%'    -- Westpac
+     OR v LIKE '%smart access%'       -- CommBank
+     OR v LIKE '%netbank%'            -- CommBank
+     OR v LIKE '%classic banking%'    -- NAB
+     OR v LIKE '%multi-currency%'     -- Wise
+     OR v LIKE '%multi currency%'
+     OR v LIKE '%transaction%'
+     OR v LIKE '%everyday%'
+     OR v LIKE '%cheque%'
+     OR v LIKE '%current account%'
+     OR v LIKE '%debit%'
+  THEN
+    RETURN 'transaction';
+  END IF;
+
+  RETURN 'other';
+END;
+$$ LANGUAGE plpgsql IMMUTABLE;
+
+-- Normalise on write so the N8N workflow can keep sending raw account_type.
+CREATE OR REPLACE FUNCTION statements_normalize_type_trigger()
+RETURNS TRIGGER AS $$
+BEGIN
+  NEW.statement_type := normalize_statement_type(NEW.statement_type);
+  RETURN NEW;
+END;
+$$ LANGUAGE plpgsql;
+
+DROP TRIGGER IF EXISTS trg_statements_normalize_type ON statements;
+CREATE TRIGGER trg_statements_normalize_type
+  BEFORE INSERT OR UPDATE OF statement_type ON statements
+  FOR EACH ROW EXECUTE FUNCTION statements_normalize_type_trigger();
+
+-- Backfill existing rows.
+UPDATE statements
+SET statement_type = normalize_statement_type(statement_type)
+WHERE statement_type IS DISTINCT FROM normalize_statement_type(statement_type);
+
+-- Guard the vocabulary. Safe because the trigger runs first on every write.
+ALTER TABLE statements DROP CONSTRAINT IF EXISTS statements_statement_type_check;
+ALTER TABLE statements
+  ADD CONSTRAINT statements_statement_type_check
+  CHECK (statement_type IN ('credit_card', 'transaction', 'savings', 'loan', 'offset', 'investment', 'other'));
+
+ALTER TABLE statements ALTER COLUMN statement_type SET DEFAULT 'other';

prisma/schema.prisma

@@ -137,7 +137,7 @@ model statements {
   event_created        Boolean?       @default(false)
   tier_used            String?
   created_at           DateTime?      @default(now())
-  statement_type       String         @default("credit_card")
+  statement_type       String         @default("other")
   currency             String?        @default("AUD")
   opening_balance      Decimal?       @db.Decimal(12, 2)
   closing_balance      Decimal?       @db.Decimal(12, 2)