Scout Database Engine Laravel Package

Getting Started

1. Install: composer require lunarphp/scout-database-engine
2. Configure Scout: In config/scout.php, set driver to database:

   'driver' => env('SCOUT_DRIVER', 'database'),
   

3. Publish config (optional): php artisan vendor:publish --tag=scout-config to inspect defaults (mostly unchanged, just ensures driver=database).
4. First use: Make sure your searchable model uses Lunarphp\ScoutDatabaseEngine\Traits\SearchableTrait and defines toSearchableArray(). Then run:

   php artisan scout:import "App\Models\YourModel"
   

   Search via YourModel::search('query')->get().

Implementation Patterns

• Indexing strategy:

  • Override toSearchableArray() in your model to control which columns/indexes are built (e.g., combine title + content into a single body string for better relevance).
  • Leverage Scout’s queue option (Config::set('scout.queue', true)) for background indexing during sync operations.
  • Use withoutSyncingToSearch() for bulk operations (e.g., seeding, migrations) to avoid overhead.

• Searching patterns:

  • Use Scout’s high-level API: Model::search('term')->where('status', 'published')->get().
  • Filter on indexed columns only—non-indexed fields won’t be searched or filtered reliably.
  • Combine multiple search constraints with where() and orWhere() (e.g., for tag-based filtering).

• Scaling locally: For dev environments, this engine avoids complex local Docker setups (Elasticsearch, etc.), enabling consistent search behavior across local, staging, and CI without external services.

Gotchas and Tips

• Full-text limitations: Only uses MySQL LIKE %term% and MATCH ... AGAINST (for MyISAM/InnoDB) depending on config—not for production-scale relevance ranking. Avoid complex regex or accent-insensitive queries without DB-specific tweaks.
• Index column limits: ScoutDatabaseEngine indexes only what’s returned from toSearchableArray(). If you forget to include a column there, filtering/sorting on it won’t work as expected.
• Case sensitivity: On MySQL, case-insensitivity depends on collation (e.g., utf8mb4_general_ci). Use lower() in toSearchableArray() if you want strict case normalization.
• Performance caveats:
  • Large tables: Add DB-level indexes on indexed columns used in where()/orderBy() after syncing.
  • sync indexing (default) blocks requests until indexed—consider queue + scout:import for large datasets.
• Debugging: Enable Scout logging (LOG_LEVEL=debug + SCOUT_LOG=true) to see raw SQL queries generated (check storage/logs).
• Extension point: Override toDatabaseSearchString() in your model for custom query logic (e.g., MySQL-specific REGEXP, or JSON path extraction).