feat(bank): multi-bank chequera — required bankAccountId, per-account scoping

The office keeps more than one operating account (Utilities banks in MXN,
Seguros in USD), but bank_transactions was a single implicit MXN register by
design. Adds Bank/BankAccount and makes every read and write in the module
scoped to exactly one account.

Schema:
- Bank / BankAccount. Currency is fixed per account and BankTransaction has
  no currency column of its own — a movement inherits its account's, the way
  a real bank account doesn't mix currencies.
- BankTransaction.bankAccountId, required. A movement with no known account
  isn't reconcilable against a statement.
- @@index([bankAccountId, transactionDate]): every read now filters by
  account and orders/groups by date.

Migration:
- backfill_bank_accounts.py seeds Scotiabank + "Utilities — Scotiabank (MXN)"
  and backfills all 22,669 existing rows onto it, then promotes the column to
  NOT NULL and attaches the FK. Standalone because prisma db push cannot add
  a required column to a populated table. Idempotent; re-running once a second
  account exists does not re-point rows.
- run_all.py runs it (both modes) before transform_bank.py, which now resolves
  the account by label and fails fast if it is missing.

API:
- ?bankAccountId= required on list/stats/facets/summary — not optional with an
  "all accounts" default, since summing an MXN and a USD register repeats the
  currency-collapsing mistake the billing module exists to prevent. Missing is
  400, unknown is 404.
- facets() had no account clause at all and summary() has two raw-SQL rollups;
  all three are now parameterised. Scoping only one of summary's queries would
  leave the year list and its drill-down describing different books.
- New bank/accounts + bank/banks sub-resource under a MANAGER
  bank:manage-accounts ability. currency is absent from the update DTO: booked
  movements are denominated in it, so editing would re-denominate history.
  Capture into a closed account is rejected.

Web:
- /banco gains an account picker (remembered per browser) and reads every
  figure in the selected account's currency; the "single currency (MXN)"
  doc-comment and the hardcoded MXN formatting are gone.
- New /banco/cuentas for banks and accounts. Accounts are closed, never
  deleted — the FK is required, so deleting one would destroy its register.
- /inicio's chequera card names the account it is reading instead of implying
  a single register.

Verified against dev + browser: a second USD account showed full read/write
isolation from the MXN register, whose totals were unchanged (22,669
movements, net 1,014,266.97).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

migration/transform_bank.py

@@ -10,12 +10,20 @@ Sources:
                            from the spelled-out "cantidad en letra"
   - TABLA RAMODOS       -> business_line_categories (line-of-business lookup)
 
+Bank account: SCOTHIA is the Utilities MXN chequera and nothing else — DATOS
+E/I carry no bank or currency column — so every row loads against the single
+account seeded by `backfill_bank_accounts.py`, which must have run first.
+`banks` / `bank_accounts` are NOT truncated here; only the movements are. (In
+full-rebuild mode that still clears app-captured rows on every account, the
+same whole-database truncate every transform in this pipeline does — use
+`--sync` to upsert instead.)
+
 Category link: DATOS E/I have no explicit FK to TABLA RAMODOS — the ramo is
 inferred from the CONCEPTO text, which is a fuzzy classification, not a stored
 key. So the categories are loaded but bank_transactions.categoryId is left
 NULL for now; a concept->ramo classifier is a later enhancement.
 
-Idempotent (truncate + rebuild). Run:
+Idempotent (rebuild the legacy rows). Run:
     ./.venv/bin/python transform_bank.py --env dev
 """
 
@@ -27,6 +35,7 @@ from pathlib import Path
 
 import pandas as pd
 
+from backfill_bank_accounts import UTILITIES_ACCOUNT
 from dbenv import connect, env_arg
 from sync import parse_mode
 
@@ -78,6 +87,19 @@ def main():
     print(f"[bank] target env: {env}")
     c = conn.cursor()
 
+    # Every SCOTHIA row belongs to the one Utilities MXN chequera. Resolved by
+    # label rather than created here, so this script can't silently open a
+    # second copy of the account if the backfill hasn't run.
+    c.execute("SELECT id FROM bank_accounts WHERE label = %s", (UTILITIES_ACCOUNT,))
+    row = c.fetchone()
+    if not row:
+        raise SystemExit(
+            f"missing bank account {UTILITIES_ACCOUNT!r} — run "
+            f"backfill_bank_accounts.py --env {env} first"
+        )
+    account_id = row[0]
+    print(f"[bank] account: {UTILITIES_ACCOUNT} ({account_id})")
+
     # business_line_categories (dedup TABLA RAMODOS)
     cats, seen = [], set()
     for _, r in load("tabla_ramodos").iterrows():
@@ -96,7 +118,8 @@ def main():
             skip_date += 1
             return
         rows.append((
-            str(uuid.uuid4()), td, s(r["tipo"]), s(r["num"]), s(r["concepto"]),
+            str(uuid.uuid4()), account_id,
+            td, s(r["tipo"]), s(r["num"]), s(r["concepto"]),
             amount, None,  # categoryId left NULL (see header)
             1 if truthy(r["operado"]) else 0,
             1 if (income and truthy(r["transferido"])) else 0,
@@ -110,9 +133,17 @@ def main():
     for _, r in load("datos_e").iterrows():
         add(r, -(dec(r["egreso"], Decimal(0))), income=False)
 
+    COLS = (
+        "id,bankAccountId,transactionDate,transactionType,reference,concept,amount,"
+        "categoryId,cleared,transferred,notes,amountInWords,legacySourceTable,legacyId"
+    )
+    PLACEHOLDERS = ",".join(["%s"] * 14)
+
     if sync_mode:
+        # bankAccountId is deliberately absent from the UPDATE clause: an
+        # account moved by hand in the app must not be dragged back.
         for row in rows:
-            c.execute("INSERT INTO bank_transactions (id,transactionDate,transactionType,reference,concept,amount,categoryId,cleared,transferred,notes,amountInWords,legacySourceTable,legacyId) VALUES (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s) ON DUPLICATE KEY UPDATE transactionDate=VALUES(transactionDate),transactionType=VALUES(transactionType),reference=VALUES(reference),concept=VALUES(concept),amount=VALUES(amount),cleared=VALUES(cleared),transferred=VALUES(transferred),notes=VALUES(notes),amountInWords=VALUES(amountInWords),voidedAt=NULL", row)
+            c.execute(f"INSERT INTO bank_transactions ({COLS}) VALUES ({PLACEHOLDERS}) ON DUPLICATE KEY UPDATE transactionDate=VALUES(transactionDate),transactionType=VALUES(transactionType),reference=VALUES(reference),concept=VALUES(concept),amount=VALUES(amount),cleared=VALUES(cleared),transferred=VALUES(transferred),notes=VALUES(notes),amountInWords=VALUES(amountInWords),voidedAt=NULL", row)
     else:
         c.execute("SET FOREIGN_KEY_CHECKS=0")
         for t in ("bank_transactions", "business_line_categories"):
@@ -120,7 +151,7 @@ def main():
         c.execute("SET FOREIGN_KEY_CHECKS=1")
         c.executemany("INSERT INTO business_line_categories (id,name) VALUES (%s,%s)", cats)
         c.executemany(
-            "INSERT INTO bank_transactions (id,transactionDate,transactionType,reference,concept,amount,categoryId,cleared,transferred,notes,amountInWords,legacySourceTable,legacyId) VALUES (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s)", rows)
+            f"INSERT INTO bank_transactions ({COLS}) VALUES ({PLACEHOLDERS})", rows)
     conn.commit()
 
     def count(t):
@@ -136,6 +167,15 @@ def main():
     for src, n, tot in by_src:
         print(f"       {(src or '(manual)'):10} {n:6}  sum {tot}")
     print(f"     net balance movement    : {net}")
+    # Per account, never a cross-account total: the registers are in different
+    # currencies and summing them produces a figure that never existed.
+    c.execute(
+        "SELECT a.label, a.currency, COUNT(t.id), COALESCE(SUM(t.amount), 0) "
+        "FROM bank_accounts a LEFT JOIN bank_transactions t ON t.bankAccountId = a.id "
+        "GROUP BY a.id, a.label, a.currency ORDER BY a.label"
+    )
+    for label, currency, n, total in c.fetchall():
+        print(f"       {label:34} {currency} {n:6}  neto {total}")
     print(f"  -> business_line_categories: {count('business_line_categories')}")
     print("  validation: OK")
     conn.close()