Prisma

Core Principles

• Always declare explicit types for variables and functions. Avoid using 'any'
• Use PascalCase for classes/interfaces, camelCase for variables/functions, kebab-case for files
• Write functions under 20 lines with single responsibility
• Always use type-safe Prisma client operations

Schema Design

• Employ domain-driven model naming conventions
• Utilize decorators like @id, @unique, and @relation
• Implement soft deletes using deletedAt timestamps
• Maintain normalized, DRY schemas
• Define proper relationships between models
• Use appropriate field types and constraints

Client Usage

• Leverage transactions for multi-step operations
• Apply middleware for logging, soft deletes, and auditing
• Handle optional relations explicitly
• Use select and include for efficient queries
• Implement pagination for large datasets

Error Management

• Catch specific errors:
  • PrismaClientKnownRequestError
  • PrismaClientValidationError
• Provide contextual, user-friendly messages
• Log detailed debugging information
• Handle unique constraint violations gracefully

Architecture

• Separate data access from business logic
• Implement repository patterns
• Use dependency injection
• Follow SOLID principles
• Prefer composition over inheritance

Performance

• Use select to fetch only needed fields
• Implement proper indexing in schema
• Use batch operations for bulk updates
• Avoid N+1 queries with proper includes
• Use connection pooling in production

Testing

• Use in-memory databases for testing
• Implement comprehensive scenario coverage
• Mock Prisma client for unit tests
• Use database transactions for test isolation

Security

• Implement input validation
• Use Row Level Security patterns
• Rely on Prisma's built-in SQL injection protection
• Validate data at both schema and application level