Skip to content

v0.3.2

Choose a tag to compare

View all tags
@github-actions github-actions released this 24 Apr 17:11
v0.3.2
4eb2b72

Scolta Laravel 0.3.2

Coordinated release across scolta-core, scolta-php, scolta-wp, scolta-drupal, scolta-laravel.

Ports the streaming gather and CLI wiring pattern from scolta-wp to Laravel.

Fixed

  • Peak RAM on large deployments: BuildCommand::streamContentItems() replaces Model::all() with Model::cursor() (PDO cursor, one row hydrated at a time) so the full model dataset is never resident in RAM. The sync PHP indexer path passes this generator to ContentExporter::filterItems() and then to IndexBuildOrchestrator::build(). (#6)
  • reporter: named argument crash: buildWithPhpIndexer() called $orchestrator->build($intent, $items, reporter: $reporter) but the parameter is named $progress, not $reporter. PHP 8.1+ raises TypeError: Unknown named argument $reporter. Fixed to use positional arguments. (#6)
  • Silent CLI during large builds: buildWithPhpIndexer() was passing the progress reporter but not a PSR-3 logger to build(). Now also passes Laravel's PSR-3 logger so memory telemetry and phase markers are visible in php artisan scolta:build output. (#6)
  • Lint: Fixed concat_space, unary_operator_spaces, not_operator_with_successor_space, blank_line_before_statement, unused import, and import ordering violations in MemoryBudgetCommand.php and ScoltaServiceProvider.php. (#5)

Added

  • Flexible memory budget and chunk size: artisan scolta:build now accepts --memory-budget=<budget> with profile names or raw byte values (256M, 1G), and a new --chunk-size=<n> flag for pages-per-chunk independent of the memory profile. Both are also configurable via scolta.memory_budget.chunk_size in config/scolta.php and the SCOLTA_CHUNK_SIZE env var. ProcessIndexChunk queue job accepts the $chunkSize parameter and passes it to MemoryBudget::fromOptions().
  • BuildCommand::gatherItemCount(): int: Uses Model::count() (one SELECT COUNT per model class) to get the total without loading any model fields. (#6)
  • BuildCommand::streamContentItems(): \Generator: Uses Model::cursor() (PDO cursor, one row hydrated at a time) instead of Model::all() so the full model dataset is never resident in RAM. (#6)

Changed

  • CI now pulls scolta-php at @dev.

Installation

composer require tag1/scolta-laravel ^0.3.2

Verification

After updating, confirm:

  • Search result titles link to real URLs, not #
  • php artisan scolta:build prints chunk-cadence progress throughout the run
  • Peak RAM on large corpora is substantially reduced vs 0.3.1

Feedback

Issues: https://github.com/tag1consulting/scolta-laravel/issues

Acknowledgements

Thanks to Hank VanZile for surfacing the 0.3.1 regressions via real-world 44k-page testing.

Full changelog

See CHANGELOG.md at tag v0.3.2.

Assets 3
Loading