docs: fix Phase 2 plan ORE test patterns and fixture usage

Fixes 4 HIGH severity issues identified in plan review:

1. Remove duplicate ore table creation
   - Deleted ore_data.sql fixture creation (Task 2, Step 2-3)
   - ore table already exists from migrations/002_install_ore_data.sql
   - Tests now use #[sqlx::test] without fixtures

2. Fix ORE helper function calls
   - Replace create_encrypted_json_with_index(&pool, 42, "ore64")
   - With get_ore_encrypted(&pool, 42) to select from pre-seeded table
   - Applied across Tasks 2-6, 8-9, 12 (all ORE tests)
   - ore table has 'ob' key (ORE_BLOCK), not ore64 index

3. Fix invalid test IDs
   - Replace id=91347 with id=4 for "different record" tests
   - 91347 * 10 = 913470 outside ore range (1-99) → NULL → panic
   - 4 * 10 = 40 is within ore range → valid data
   - Applied in Task 1 (inequality tests)

4. Remove ore_data fixture references
   - Removed all scripts("ore_data") fixture attributes
   - Updated tests to use migrations instead
   - Added comments explaining ore table source

Updated documentation:
- FIXTURE_SCHEMA.md: Document ore from migrations (NOT fixture)
- Added helper function documentation (get_ore_encrypted)
- Added valid id range explanation (1-9 → lookups 10-90)

Coverage: All ORE-related tests now use correct patterns

Author: tobyhede