Remove "objectID" usage of this repository

[curquiza]

This repository was highly built based on the Algolia's search bundle.

Algolia, to define the unique field of the documents uses the objectID field. Which is not the case of MeiliSearch. MeiliSearch uses the primary key: it infers it, or the user can define it.
More about the primary key of MeiliSerach: https://docs.meilisearch.com/learn/core_concepts/documents.html#primary-field

It means having a code base with objectID does not make sense for a MeiliSearch repository in terms of naming (related to #58) but also in terms of usage!
Indeed it leads us to some issues in the past, especially because the search() method is badly designed for MeiliSearch. See:

• Service returns bad entities for correct search #56
• Entity mapping not working correctly? #23

⚠️ Because you cannot know the primary key (so the unique field) when getting the search response by MeiliSearch we might think about a way to avoid using this information

[conradfr]

I've been burned by the hardcoded 'objectID' in this bundle which was making the indexing task fail in meilisearch (which is done async so the error was not easily apparent) because the primary key had been set as 'id' during the index creation (on another language / lib).

if would be great to remove the hardcoded value (look up the index function in engine.php) and allow the primary key to be set in the config file indices.

[norkunas]

Sorry for the bad experience, I also had some problems with this. It's on my todo list