What is the structure and principles of organizing the data layer in the Clean Architecture?

In Clean Architecture, the data layer is responsible for interacting with external data sources (databases, network, file system) and implements interfaces defined in the domain layer.

Main principles of organizing the data layer:

• Isolation from business logic — the data layer should not directly influence the application logic.
• Implementation of interfaces (repositories) declared in the domain layer for data access.
• Use of DTOs or data models different from domain entities to avoid dependency leaks.
• Error handling and data transformation for ease of use in the domain.

The structure of the data layer may include:

• Repositories — classes implementing data access interfaces.
• Data sources — classes working with specific technologies (Room, Retrofit, SharedPreferences).
• Mappers — convert data from source format to domain models and vice versa.

Example of a repository interface in the domain:

interface UserRepository {
    suspend fun getUser(id: String): User
}

Implementation in the data layer:

class UserRepositoryImpl(private val api: UserApi, private val userDao: UserDao) : UserRepository {
    override suspend fun getUser(id: String): User {
        val userEntity = userDao.getUserById(id) ?: api.fetchUser(id)
        return userEntity.toDomainModel()
    }
}

Thus, the data layer acts as a bridge between external sources and business logic, ensuring clean architecture and testability.