Index Strategies

Index strategies represent the culmination of things needed to actually index data. A strategy accepts a Data Source, and uses that source to loop through the records and save the data.

Different data sources may require different values to be set to them in-order to function. Perhaps it's necessary to authenticate with a secret, or maybe you're importing a CSV and need to map columns to model properties. Either way, this is generally a required step to accomplish what's needed. The Index Strategy will take a data source will all of these set properties, and then use it to index the data.

Example - Index data from a CSV import

Let's say you have a CSV file that you want to import to the database. To accomplish this, you can make use of the CSV Data Source, attach that data source to an index strategy, and then call that strategy to instruct the system to import, or index the data in the database.

For the sake of this example, let's assume we have a very basic content model that includes content, a title, and an author.

class Example_Model implements Has_Title, Has_Content, Has_Author{
  use With_Title
  use With_Content
  use With_Author
}

Now, we can associate that model with a CSV import by mapping each column to the various model setters, like so:

// This would probably come from a file_get_contents call.
$csv_data = 'id,author,title,content
1,Alex,"When Twitter Loses, WordPress Wins",Content 1
2,Kara,What It’s Like Working in NICU,Content 2
3,Courtney,How to train your golden to do anything,Content 3
4,Jonathon,Forget capital gains – get dividends,Content 4
5,Kellie,I. Love. You.,Content 5';

$data_source = (new CSV())
                 ->set_model(Example_Model::class)
                 ->set_data_source_adapter(
                     (new Data_Source_Adapter()
                        ->map_column('id', 'set_id', Types::Integer)
                        ->map_column('author','set_author', Types::String)
                        ->map_column('title', 'set_title', Types::String)
                        ->map_column('content', 'set_content', Types::String)
                     )
                 )
                 ->set_csv($csv_data)

Finally, we can combine that data source with a data source, and when the time is right use it to index the data.

$strategy = (new Index_Strategy())
              ->set_data_source($data_source);

// When we're ready, we can use this data source. Perhaps this is called in a hook of some sort.
// Calling this would initiate a loop that would go through all CSV files and index the data.
$strategy->index_data();

// Or we can index a single record from the CSV, if we prefer. This would index "When Twitter Loses, WordPress Wins" by "Alex"
$strategy->index_item(1)

Suggested Implementation - Use a Builder

An Index Strategy Builder works kind of like a "template" for an index strategy. They allow you to pre-build common patterns among different integrations, so you don't have to repeat a lot of the setup every time. We could expand on the example above and actually make a builder to make it easier to import any CSV that matches this format in the future.

You absolutely don't have to do this - but this helps illustrate how some of the core implementers in Adiungo enables you to set this kind of thing up.

class Example_CSV_Importer implements Has_Index_Strategy, Has_Data_Source{
  use With_Index_Strategy;
  use With_Data_Source;
  
  public function __construct(string $csv_data){
    $this->set_data_source((new CSV())
                 ->set_model(Example_Model::class)
                 ->set_data_source_adapter(
                     (new Data_Source_Adapter()
                        ->map_column('id', 'set_id', Types::Integer)
                        ->map_column('author','set_author', Types::String)
                        ->map_column('title', 'set_title', Types::String)
                        ->map_column('content', 'set_content', Types::String)
                     )
                 )
                 ->set_csv($csv_data));
   $this->set_index_strategy((new Index_Strategy())->set_data_source($this->get_data_source()));
  }
}

Now, when the time comes, you can build this class, and potentially change anything about the data source or index strategy if needed. Think of this like a "template" that provides sane defaults for your example.

// This would probably come from a file_get_contents call.
$csv_data = 'id,author,title,content
1,Alex,"When Twitter Loses, WordPress Wins",Content 1
2,Kara,What It’s Like Working in NICU,Content 2
3,Courtney,How to train your golden to do anything,Content 3
4,Jonathon,Forget capital gains – get dividends,Content 4
5,Kellie,I. Love. You.,Content 5';

// When we're ready, we can use this data source. Perhaps this is called in a hook of some sort.
// Calling this would initiate a loop that would go through all CSV files and index the data.
(new Example_CSV_Importer($csv_data))->get_index_strategy()->index_data();

// Or we can index a single record from the CSV, if we prefer. This would index "When Twitter Loses, WordPress Wins" by "Alex"
(new Example_CSV_Importer($csv_data))->get_index_strategy()->index_item(1);

// Finally, you can override or change things. Perhaps you have a special csv where the "author" column is called "user"
$importer = (new Example_CSV_Importer($csv_data));

$importer->get_data_source()->get_data_source_adapter()->map_column('user','set_author',Types::String);

$importer->get_index_strategy()->index_data();