Context

PR #193 introduced two near-identical helpers for prefetching per-account AccountServiceOverride data into a Map<accountID, AccountServiceOverride[]> cache:

• openFanOutModal (sub-option ii) — collects single-account-bucket account IDs, fetches via Promise.all, swallows per-account errors, passes the cache to resolveBucketPaymentSeed.
• openPurchaseModal (sub-option iii) — collects all rec account IDs, fetches via Promise.all, swallows per-account errors, passes the cache to resolvePerRecPaymentSeed.

The fetch+cache shape is ~10 lines and identical aside from "which account IDs do we collect?". The seed functions themselves are different (per-bucket vs per-rec), so they don't merge — but the prefetch wrapper does.

What needs to happen

Extract a shared helper, e.g.:

// frontend/src/lib/overrides.ts
export async function fetchOverridesForAccounts(
  ids: Iterable<string>,
): Promise<Map<string, AccountServiceOverride[]>> {
  const out = new Map<string, AccountServiceOverride[]>();
  await Promise.all(
    Array.from(ids).map(async (id) => {
      try {
        out.set(id, await listAccountServiceOverrides(id));
      } catch {
        // Silent fallback — callers handle missing entries gracefully.
      }
    }),
  );
  return out;
}

Then openFanOutModal and openPurchaseModal each become a 1-line call.

Bonus: single mock target for tests

The existing tests mock api.listAccountServiceOverrides directly. After this refactor, they could either keep doing that (the helper still calls the same API) or mock the new helper directly for cleaner test fixtures.

References

• PR feat(recommendations): per-account service override pre-populated at purchase time (#111, ii+iii) #193 — both helpers introduced. The openPurchaseModal introduction comment explicitly flags the duplication and points at this follow-up.

Effort & priority

• Effort: XS (mechanical extract + test mock check)
• Priority: P3 (no functional change; only cleanup)