Facet is a C# source generator that lets you define projections (DTOs, API models, etc.) directly from your domain models, without writing boilerplate.
Facetting is the process of defining focused views of a larger model at compile time.
Instead of manually writing separate DTOs, mappers, and projections, Facet allows you to declare what you want to keep, and generates everything else. It generates partial classes, records, structs, or record structs with constructors, LINQ projections, and even supports custom mappings, all at compile time, with zero runtime cost.
You can think of it like carving out a specific facet of a gem:
- The part you care about
- Leave the rest behind.
- ✅ Generate classes, records, structs, or record structs from existing types
- ✅ Define what to include, or exclude fields/properties you don't want
- ✅ Include/redact public fields
- ✅ Constructors & LINQ projection expressions
- ✅ Full mapping support with custom mapping configurations
- ✅ Auto-generate complete CRUD DTO sets with
[GenerateDtos]
- ✅ Expression transformation and mapping utilities for reusing business logic across entities and DTOs
- ✅ Preserves member and type XML documentation
Facet is modular and consists of several NuGet packages:
-
Facet: The core source generator. Generates DTOs, projections, and mapping code.
-
Facet.Extensions: Provider-agnostic extension methods for mapping and projecting (works with any LINQ provider, no EF Core dependency).
-
Facet.Mapping: Advanced static mapping configuration support with async capabilities and dependency injection for complex mapping scenarios.
-
Facet.Mapping.Expressions: Expression tree transformation utilities for transforming predicates, selectors, and business logic between source entities and their Facet projections.
-
Facet.Extensions.EFCore: Async extension methods for Entity Framework Core (requires EF Core 6+).
dotnet add package Facet
For LINQ helpers:
dotnet add package Facet.Extensions
For EF Core support:
dotnet add package Facet.Extensions.EFCore
For expression transformation utilities:
dotnet add package Facet.Mapping.Expressions
... and specify what you need.
// Exclude sensitive properties
string[] excludeFields = { "Password", "Email" };
[Facet(typeof(User), exclude: excludeFields)]
public partial class UserWithoutEmail { }
// Include only specific properties
[Facet(typeof(User), Include = new[] { "FirstName", "LastName", "Email" })]
public partial class UserContactDto { }
// Include public fields
[Facet(typeof(Entity), IncludeFields = true)]
public partial class EntityDto { }
// Include specific fields and properties
[Facet(typeof(Entity), Include = new[] { "Name", "Status" }, IncludeFields = true)]
public partial class EntitySummaryDto { }
// Make all properties nullable for query/filter scenarios
[Facet(typeof(Product), "InternalNotes", NullableProperties = true, GenerateBackTo = false)]
public partial class ProductQueryDto { }
[Facet(typeof(User))]
public partial class UserFacet { }
// Auto-generates constructor, properties, and LINQ projection
var userFacet = user.ToFacet<UserFacet>();
var userFacet = user.ToFacet<User, UserFacet>(); //Much faster
var user = userFacet.BackTo<User>();
var user = userFacet.BackTo<UserFacet, User>(); //Much faster
var users = users.SelectFacets<UserFacet>();
var users = users.SelectFacets<User, UserFacet>(); //Much faster
public class UserMapper : IFacetMapConfiguration<User, UserDto>
{
public static void Map(User source, UserDto target)
{
target.FullName = $"{source.FirstName} {source.LastName}";
target.Age = CalculateAge(source.DateOfBirth);
}
}
[Facet(typeof(User), Configuration = typeof(UserMapper))]
public partial class UserDto
{
public string FullName { get; set; }
public int Age { get; set; }
}
public class UserAsyncMapper : IFacetMapConfigurationAsync<User, UserDto>
{
public static async Task MapAsync(User source, UserDto target, CancellationToken cancellationToken = default)
{
// Async database lookup
target.ProfilePicture = await GetProfilePictureAsync(source.Id, cancellationToken);
// Async API call
target.ReputationScore = await CalculateReputationAsync(source.Email, cancellationToken);
}
}
// Usage
var userDto = await user.ToFacetAsync<User, UserDto, UserAsyncMapper>();
var userDtos = await users.ToFacetsParallelAsync<User, UserDto, UserAsyncMapper>();
public class UserAsyncMapperWithDI : IFacetMapConfigurationAsyncInstance<User, UserDto>
{
private readonly IProfilePictureService _profileService;
private readonly IReputationService _reputationService;
public UserAsyncMapperWithDI(IProfilePictureService profileService, IReputationService reputationService)
{
_profileService = profileService;
_reputationService = reputationService;
}
public async Task MapAsync(User source, UserDto target, CancellationToken cancellationToken = default)
{
// Use injected services
target.ProfilePicture = await _profileService.GetProfilePictureAsync(source.Id, cancellationToken);
target.ReputationScore = await _reputationService.CalculateReputationAsync(source.Email, cancellationToken);
}
}
// Usage with DI
var mapper = new UserAsyncMapperWithDI(profileService, reputationService);
var userDto = await user.ToFacetAsync(mapper);
var userDtos = await users.ToFacetsParallelAsync(mapper);
// Async projection directly in EF Core queries
var userDtos = await dbContext.Users
.Where(u => u.IsActive)
.ToFacetsAsync<UserDto>();
// LINQ projection for complex queries
var results = await dbContext.Products
.Where(p => p.IsAvailable)
.SelectFacet<ProductDto>()
.OrderBy(dto => dto.Name)
.ToListAsync();
[Facet(typeof(User)]
public partial class UpdateUserDto { }
[HttpPut("{id}")]
public async Task<IActionResult> UpdateUser(int id, UpdateUserDto dto)
{
var user = await context.Users.FindAsync(id);
if (user == null) return NotFound();
// Only updates properties that mutated
user.UpdateFromFacet(dto, context);
await context.SaveChangesAsync();
return NoContent();
}
// With change tracking for auditing
var result = user.UpdateFromFacetWithChanges(dto, context);
if (result.HasChanges)
{
logger.LogInformation("User {UserId} updated. Changed: {Properties}",
user.Id, string.Join(", ", result.ChangedProperties));
}
Generate standard Create, Update, Response, Query, and Upsert DTOs automatically:
// Generate all standard CRUD DTOs
[GenerateDtos(Types = DtoTypes.All, OutputType = OutputType.Record)]
public class User
{
public int Id { get; set; }
public string FirstName { get; set; }
public string LastName { get; set; }
public string Email { get; set; }
public DateTime CreatedAt { get; set; }
}
// Auto-generates:
// - CreateUserRequest (excludes Id)
// - UpdateUserRequest (includes Id)
// - UserResponse (includes all)
// - UserQuery (all properties nullable)
// - UpsertUserRequest (includes Id, for create/update operations)
Entities with Smart Exclusions
[GenerateAuditableDtos(
Types = DtoTypes.Create | DtoTypes.Update | DtoTypes.Response,
OutputType = OutputType.Record,
ExcludeProperties = new[] { "Password" })]
public class Product
{
public int Id { get; set; }
public string Name { get; set; }
public string Password { get; set; } // Excluded
public DateTime CreatedAt { get; set; } // Auto-excluded (audit)
public string CreatedBy { get; set; } // Auto-excluded (audit)
}
// Auto-excludes audit fields: CreatedAt, UpdatedAt, CreatedBy, UpdatedBy
// Different exclusions for different DTO types
[GenerateDtos(Types = DtoTypes.Response, ExcludeProperties = new[] { "Password", "InternalNotes" })]
[GenerateDtos(Types = DtoTypes.Upsert, ExcludeProperties = new[] { "Password" })]
public class Schedule
{
public int Id { get; set; }
public string Name { get; set; }
public string Password { get; set; } // Excluded from both
public string InternalNotes { get; set; } // Only excluded from Response
}
// Generates:
// - ScheduleResponse (excludes Password, InternalNotes)
// - UpsertScheduleRequest (excludes Password, includes InternalNotes)
Facet delivers competitive performance across different mapping scenarios. Here's how it compares to popular alternatives:
Library | Mean Time | Memory Allocated | Performance vs Facet |
---|---|---|---|
Facet | 15.93 ns | 136 B | Baseline |
Mapperly | 15.09 ns | 128 B | 5% faster, 6% less memory |
Mapster | 21.90 ns | 128 B | 38% slower, 6% less memory |
Library | Mean Time | Memory Allocated | Performance vs Facet |
---|---|---|---|
Mapster | 192.55 ns | 1,416 B | 10% faster, 10% less memory |
Facet | 207.32 ns | 1,568 B | Baseline |
Mapperly | 222.50 ns | 1,552 B | 7% slower, 1% less memory |
For this benchmark we used the <TSource, TTarget>
methods.
Insights:
- Single mapping: All three libraries perform similarly with sub-nanosecond differences
- Collection mapping: Mapster has a slight edge for bulk operations, while Facet and Mapperly are very close
- Memory efficiency: All libraries are within ~10% of each other for memory allocation
- Compile-time generation: Both Facet and Mapperly benefit from zero-runtime-cost source generation