vaened / php-search-engine Goto Github PK

The Search Library provides a fundamental and flexible foundation for building search engines in applications. This library is not tied to any specific ORM implementation; rather, it serves as the core upon which more complex search solutions can be built in other libraries that integrate with specific ORMs.

class ProductSearchEngine extends SearchEngine  
{  
    use Flaggable, Indexable;  
  
    public function __construct(  
        private readonly IndexRepository $index,  
        private readonly FlagFiltrator   $filter  
    ) {  
        $this->apply(Criterias\ProductStatus::active());  
    }  
  
    protected function filterer(): Filterer  
    {  
        return $this->filter;  
    }  
  
    protected function indexer(): indexer  
    {  
        return $this->index;  
    }  
}

It offers a robust structure for handling common search functionalities, such as custom filters, text searches, and range management. Its modular design ensures efficient integration and the ability to extend to cover more sophisticated search needs as project requirements evolve.

PHP Search Engine requires PHP 8.1. To get the latest version, simply require the project using Composer:

composer require vaened/php-search-engine

This library provides a foundational framework for creating search engines. It includes the AbstractSearchEngine class as a base for custom search implementations and two key traits: Flaggable for filtering based on flags, and Indexable for search operations using indices. These components work together to offer a flexible and extensible solution for managing search criteria.

The library provides a base class called AbstractSearchEngine which offers fundamental functionality typically extended by other libraries for specific implementations. However, it's not mandatory to extend this class—custom implementations of SearchEngine can be created from scratch.

The key feature of AbstractSearchEngine is the apply method:

protected function apply(Scope|Expression|Filter ...$criterias): void
{
    $this->criterias = array_merge($this->criterias, $criterias);
}

This method progressively builds search criteria based on applied filters, ultimately constructing the Criteria object.

The criteria method is responsible for creating a Criteria instance that encapsulates all the search criteria:

protected function criteria(int $page = 1, ?int $perPage = null): Criteria
{
    $perPage ??= $this->perPage;
    $offset  = ($page - 1) * $perPage;

    return new Criteria($this->criterias, $this->order, $perPage, $offset);
}

This method calculates the pagination offset and initializes the Criteria object with the accumulated search criteria, order, and pagination details.

The Flaggable trait adds the filter method to the SearchEngine, allowing you to apply filters based on specific flags. This method takes a FlagBag containing a set of flags and applies the corresponding filters. Each flag becomes a search criterion that is added to the engine.

public function filter(FlagBag $flags): self
{
    $this->filterer()
         ->only($flags)
         ->each(
             fn(Scope|Expression|Filter $criteria) => $this->apply($criteria)
         );

    return $this;
}

The Indexable trait provides the search method, which allows you to perform searches based on a specific index. It uses an index defined by the BackedEnum and a queryString to search for the associated criterion in the index repository. If a valid criterion is found, it is applied to the search engine.

public function search(BackedEnum $index, ?string $queryString): static
{
    if (
        null !== $queryString &&
        null !== (
        $criteria = $this->indexer()->search($index, $queryString)
        )
    ) {
        $this->apply($criteria);
    }

    return $this;
}

Filters represent conditions applied during the search process. These filters are crucial for defining and refining search criteria according to specific requirements. Filters can be simple, such as a direct value check, or more complex, combining multiple conditions to restrict results in detail.

Example of internal definition

interface Filter  
{  
    public function field(): FilterField;  
  
    public function operator(): FilterOperator;  
  
    public function value(): FilterValue;  
}

Flags are represented by classes that define binary filters, meaning they do not require additional parameters to determine whether a condition is met or not. These filters act as boolean indicators that apply predefined conditions without needing extra input values. Flags return a collection of filters based on the flag parameter provided.

Example of internal implementation

public function only(FlagBag $flags): ArrayList
{
    return $this->flags()
                ->only($flags)
                ->map(static fn(callable $criteria) => apply($criteria));
}

Indexes are represented by classes that manage specific search indices. Each index is associated with a search criterion that is applied to a given value. The primary function of an Index is to return a single filter based on the provided index and value. This structure allows for optimized and precise searches using predefined indices.

Example of internal implementation

public function search(BackedEnum $index, string $queryString): null|Scope|Expression|Filter
{
    $criteria = $this->indexes()->get($index);

    if (null === $criteria) {
        return null;
    }

    return apply($criteria, [$queryString]);
}

In the Search Library, each filter is treated as a separate PHP class with its own specific logic. This approach promotes order and cohesion in the design of the search system. By encapsulating filtering logic in separate classes, the code becomes easier to maintain and extend.

For instance, instead of having scattered filtering functions, classes like PersonId and PersonName are defined, each containing methods such as equals(value) or startsWith(value). This not only enhances code clarity but also enables more efficient reuse of filters in different contexts. By keeping filtering logic within dedicated classes, each filter has a single, well-defined responsibility, contributing to a cleaner and more organized design.

This library is licensed under the MIT License. For more information, please see the license file.