refactor: optimize whitelist table schema for performance

Migrate the `whitelist` table to `whitelist_v1` with optimized storage
and indexing. This change removes the unnecessary auto-increment `id`
column and uses `info_hash` directly as the PRIMARY KEY, improving
both storage efficiency and query performance.

Database-specific optimizations:
- **SQLite**: Use `WITHOUT ROWID` optimization since info_hash is
  already a suitable primary key (40-char hex string). This eliminates
  the hidden rowid column and reduces storage overhead.
- **MySQL**: Store info_hash as `BINARY(20)` instead of `VARCHAR(40)`,
  reducing storage from 40 bytes to 20 bytes per entry and improving
  index performance.

Changes:
- Add migration scripts for both SQLite and MySQL to:
  - Create new `whitelist_v1` table with optimized schema
  - Migrate existing data (converting hex to binary for MySQL)
  - Drop old `whitelist` table
- Update hardcoded table creation SQL in database drivers to match
  new schema
- Update all whitelist queries to use `whitelist_v1` table name
- Update MySQL driver to use binary storage with `UNHEX()`/bytes()
  for info_hash operations
- Document the new schema in migrations README with storage format
  details for both databases

The migration preserves all existing whitelist data while improving
storage efficiency and query performance for whitelist lookups.

Author: da2ce7