The equivalent JSON schema is not available till now.

AVRO and Protobuf

Beside the JSON serializer there are both AVRO and Protobuf serializer.

Default managed types

Each serializer supports the following list of types can be used to declare the fields of the entities in Entity Framework Core:

• string
• Guid
• Guid?
• DateTime
• DateTime?
• DateTimeOffset
• DateTimeOffset?
• bool
• bool?
• char
• char?
• sbyte
• sbyte?
• byte
• byte?
• short
• short?
• ushort
• ushort?
• int
• int?
• uint
• uint?
• long
• long?
• ulong
• ulong?
• double
• double?
• float
• float?
• decimal
• decimal?
• byte[]

Code and user override

The code is based on three elements shall be available to Entity Framework Core provider for Apache Kafka™ in order to work:

• ValueContainer type: a type which encapsulate the Entity and stores needed information
• Key SerDes: the serializer of the Primary Key
• ValueContainer SerDes: the serializer of the ValueContainer

Default types

Entity Framework Core provider for Apache Kafka™ comes with some default values:

• ValueContainer class: KEFCore uses DefaultValueContainer<T> (i.e. DefaultKEFCoreSerDes.DefaultValueContainer) which stores the CLR type of Entity, the properties ordered by their index with associated CLT type, name and JSON serializaed value; the class is marked for JSON serialization and it is used from the ValueContainer SerDes;
• Key SerDes class: KEFCore uses DefaultKEFCoreSerDes.Key.JsonRaw<T> (i.e. DefaultKEFCoreSerDes.DefaultKeySerialization), the type automatically manages simple or complex Primary Key
• ValueContainer SerDes class: KEFCore uses DefaultKEFCoreSerDes.ValueContainer.JsonRaw<> (i.e. DefaultKEFCoreSerDes.DefaultValueContainerSerialization)

Both Key and ValueContainer SerDes come with two kind of data transfer mechanisms:

• Raw: it uses byte arrays for data transfer
• Buffered: they use ByteBuffer for data transfer

User override

The default serialization can be overridden with user defined ValueContainer class, Key SerDes or ValueContainer SerDes.

ValueContainer class

A custom ValueContainer class must contains enough information and shall follow the following rules:

• must implements the IValueContainer<T> interface
• must be a generic type
• must have at least a default constructor and a constructor which accept six parameters in this order:
  1. IEntityType -> the Entity will be serialized
  2. IProperty[]? -> the properties associated to the Entity at point 1
  3. object[] -> an array of values associated on each property at point 2
  4. IComplexProperty[]? -> an optional array of properties associated to the Entity at point 1 which represents the Complex properties
  5. object[]? -> an optional array of values associated on each property at point 4
  6. IComplexTypeConverterFactory? -> an optional reference to the factory where the instance will retrieve specialized serializer, see Complex type serialization

An example snippet is the follow:

Key SerDes and ValueContainer SerDes class

A custom Key SerDes class shall follow the following rules:

• must implements the ISerDes<T> interface or extend SerDes<T>
• must be a generic type
• must have a parameterless constructor
• can store serialization information using Headers of Apache Kafka™ record (this information will be used from EntityExtractor)

An example snippet is the follow based on JSON serializer:

How to use custom types

KafkaDbContext contains three properties can be used to override the default types:

• KeySerializationType: set the value of the Key SerDes type in the form CustomSerDes<>
• ValueSerializationType: set the value of the ValueContainer SerDes type in the form CustomSerDes<>
• ValueContainerType: set the value of the ValueContainer type in the form CustomValueContainer<>

IMPORTANT NOTE: the type applied in the previous properties of KafkaDbContext shall be a generic type definition, Entity Framework Core provider for Apache Kafka™ will apply the right generic type when needed.

Avro serialization

With package MASES.EntityFrameworkCore.KNet.Serialization.Avro an user can choose two different Avro serializers: The engine comes with two different encoders

• Binary
• Json

Both Key and ValueContainer SerDes, Binary or Json, come with two kind of data transfer mechanisms:

• Raw: it uses byte arrays for data transfer
• Buffered: they use ByteBuffer for data transfer

Avro schema

The following schema is the default used from the engine and can be registered in Apache Schema registry so other applications can use it to extract the data stored in the topics:

• Complex Primary Key schema:

  { "namespace": "MASES.EntityFrameworkCore.KNet.Serialization.Avro.Storage", "type": "record", "name": "AvroKeyContainer", "doc": "Represents the storage container type to be used from KEFCore for keys", "fields": [ { "name": "PrimaryKey", "doc": "Represents a primary key used from KEFCore", "type": { "type": "array", "items": [ "null", "boolean", "int", "long", "float", "double", "string" ] } } ] }

• ValueContainer schema:

  { "namespace": "MASES.EntityFrameworkCore.KNet.Serialization.Avro.Storage", "type": "record", "name": "AvroValueContainer", "doc": "Represents the storage container type to be used from KEFCore values", "fields": [ { "name": "EntityName", "doc": "Represents the entity name of the Entity in the EF Core schema of the EntityType", "type": "string" }, { "name": "ClrType", "doc": "Represents the CLR type of the Entity in the EF Core schema of the EntityType", "type": "string" }, { "name": "Data", "type": { "type": "array", "items": { "namespace": "MASES.EntityFrameworkCore.KNet.Serialization.Avro.Storage", "type": "record", "name": "PropertyDataRecord", "doc": "Represents the single container for Entity properties stored in AvroValueContainer and used from KEFCore", "fields": [ { "name": "PropertyIndex", "doc": "Represents the index of the property in the EF Core schema of the EntityType", "type": "int" }, { "name": "PropertyName", "doc": "Represents the name of the property in the EF Core schema of the EntityType", "type": "string" }, { "name": "ManagedType", "doc": "Represents the internal KEFCore type associated to the property in the EF Core schema of the EntityType", "type": "int" }, { "name": "SupportNull", "doc": "true if the ManagedType shall support null, e.g. Nullable type in .NET", "type": "boolean" }, { "name": "ClrType", "doc": "Represents the CLR type of the property in the EF Core schema of the EntityType", "type": "string" }, { "name": "Value", "type": [ "null", "boolean", "int", "long", "float", "double", "string" ] } ] } } } ] }

The extension converted this schema into code to speedup the exection of serialization/deserialization operations.

How to use Avro

KafkaDbContext contains three properties can be used to override the default types:

• KeySerializationType: set this value to AvroKEFCoreSerDes.Key.BinaryRaw<> or AvroKEFCoreSerDes.Key.JsonRaw<> or use AvroKEFCoreSerDes.DefaultKeySerialization (defaults to AvroKEFCoreSerDes.Key.BinaryRaw<>), both types automatically manages simple or complex Primary Key
• ValueSerializationType: set this value to AvroKEFCoreSerDes.ValueContainer.BinaryRaw<> or AvroKEFCoreSerDes.ValueContainer.JsonRaw<> or use AvroKEFCoreSerDes.DefaultValueContainerSerialization (defaults to AvroKEFCoreSerDes.ValueContainer.BinaryRaw<>)
• ValueContainerType: set this value to AvroValueContainerRaw<> or use AvroKEFCoreSerDes.DefaultValueContainer

An example is:

Protobuf serialization

With package MASES.EntityFrameworkCore.KNet.Serialization.Protobuf an user can choose the Protobuf serializer.

Both Key and ValueContainer SerDes come with two kind of data transfer mechanisms:

• Raw: it uses byte arrays for data transfer
• Buffered: they use ByteBuffer for data transfer

Protobuf schema

The following schema is the default used from the engine and can be registered in Apache Schema registry so other applications can use it to extract the data stored in the topics:

• Common multitype value:

  // [START declaration] syntax = "proto3"; package storage; import "google/protobuf/struct.proto"; import "google/protobuf/timestamp.proto"; // [END declaration] // [START java_declaration] option java_multiple_files = true; option java_package = "mases.entityframeworkcore.knet.serialization.protobuf"; option java_outer_classname = "GenericValue"; // [END java_declaration] // [START csharp_declaration] option csharp_namespace = "MASES.EntityFrameworkCore.KNet.Serialization.Protobuf.Storage"; // [END csharp_declaration] // [START messages] // Represent a System.Datetime message Datetime { // Represents a Timestamp value. google.protobuf.Timestamp datetime_value = 1; // Represents a Utc/Local value. bool utc_value = 2; } // Our address book file is just one of these. message GenericValue { // The kind of value. oneof kind { // Represents a null value. google.protobuf.NullValue null_value = 1; // Represents a boolean value. bool bool_value = 2; // Represents a char value. string char_value = 3; // Represents a int value. uint32 byte_value = 4; // Represents a int value. int32 sbyte_value = 5; // Represents a int value. int32 short_value = 6; // Represents a int value. uint32 ushort_value = 7; // Represents a int value. int32 int_value = 8; // Represents a int value. uint32 uint_value = 9; // Represents a long value. int64 long_value = 10; // Represents a long value. uint64 ulong_value = 11; // Represents a float value. float float_value = 12; // Represents a double value. double double_value = 13; // Represents a string value. string string_value = 14; // Represents a Guid value. bytes guid_value = 15; // Represents a Datetime value. Datetime datetime_value = 16; // Represents a Datetime value. google.protobuf.Timestamp datetimeoffset_value = 17; // Represents a decimal value. string decimal_value = 18; } int32 ManagedType = 19; bool SupportNull = 20; } // [END messages]

• Complex Primary Key schema:

  // [START declaration] syntax = "proto3"; package storage; import "GenericValue.proto"; // [END declaration] // [START java_declaration] option java_multiple_files = true; option java_package = "mases.entityframeworkcore.knet.serialization.protobuf"; option java_outer_classname = "KeyContainer"; // [END java_declaration] // [START csharp_declaration] option csharp_namespace = "MASES.EntityFrameworkCore.KNet.Serialization.Protobuf.Storage"; // [END csharp_declaration] // [START messages] message PrimaryKeyType { repeated GenericValue values = 1; } // Our address book file is just one of these. message KeyContainer { PrimaryKeyType PrimaryKey = 1; } // [END messages]

• ValueContainer schema:

  // [START declaration] syntax = "proto3"; package storage; import "GenericValue.proto"; // [END declaration] // [START java_declaration] option java_multiple_files = true; option java_package = "mases.entityframeworkcore.knet.serialization.protobuf"; option java_outer_classname = "ValueContainer"; // [END java_declaration] // [START csharp_declaration] option csharp_namespace = "MASES.EntityFrameworkCore.KNet.Serialization.Protobuf.Storage"; // [END csharp_declaration] // [START messages] message PropertyDataRecord { int32 PropertyIndex = 1; string PropertyName = 2; string ClrType = 3; GenericValue Value = 4; } // Our address book file is just one of these. message ValueContainer { string EntityName = 1; string ClrType = 2; repeated PropertyDataRecord Data = 3; } // [END messages]

The extension converted this schema into code to speedup the exection of serialization/deserialization operations.

How to use Protobuf

KafkaDbContext contains three properties can be used to override the default types:

• KeySerializationType: set this value to ProtobufKEFCoreSerDes.Key.BinaryRaw<> or use ProtobufKEFCoreSerDes.DefaultKeySerialization, the type automatically manages simple or complex Primary Key
• ValueSerializationType: set this value to ProtobufKEFCoreSerDes.ValueContainer.BinaryRaw<> or use ProtobufKEFCoreSerDes.DefaultValueContainerSerialization
• ValueContainerType: set this value to ProtobufValueContainerRaw<> or use ProtobufKEFCoreSerDes.DefaultValueContainer

An example is:

Complex type serialization

Some types can be marked as ComplexType and EF Core will manage them differently from other types associated to the model. An instance of EF Core using classic database provider will create, depends on the implementation, some extra columns to manage the content of each property of the CLR type marked as ComplexType. KEFCore works in a different way since the content stored in Kafka contains the serializable information of the Entity plus the references to other types in the model. KEFCore is based on the serializer described above and the new behavior supports conversion of ComplexType using two distinct behavior:

1. Enabling the default serialization based on Json format the complex type are threated as POCO object and the sub-system leave the responsability to the .NET Json serializer
2. Each type declared as ComplexType can have its own converter which manages the specific object using specialized classes implementing IComplexTypeConverter interface

The IComplexTypeConverter interface is very simple:

A single instance can support multiple Types (this is the type declared as ComplexType) and exposes two methods:

• Convert: converts EF Core types declared as ComplexType to something manageable from the underlying serializer, so the final output can be one of the type declared in Default managed types
• ConvertBack: converts back something manageable from the underlying serializer (i.e. one of the type declared in Default managed types) to an EF Core types declared as ComplexType

[!TIP] For implementers: the parameter is passed with ref keyword to avoid an excessive memory pressure and the converted value shall replace the received input.

[!IMPORTANT] The implementation of IComplexTypeConverter interface shall be thread-safe.

An user can build an external library and can register the custom converters within the system using the singleton service implementing IComplexTypeConverterFactory interface:

• using GetService:

  var context = new KakfaDbContext(); context.GetService<IComplexTypeConverterFactory>(); factory.Register(converter);

• using the methods available in KakfaDbContext

How to declare a complex property

Some good examples comes from https://www.learnentityframeworkcore.com/model/complex-type. An user can choose two different approaches:

• if the user is declaring its own model the ComplexType is a good choice:

  [Table("TaxInfo", Schema = "ComplexTest")] public class TaxInfo { public int TaxInfoId { get; set; } public char Code { get; set; } public decimal Percentage { get; set; } [Required] public TaxInfoExtended TaxInfoExtended { get; set; } public int ExtraValue { get; set; } // used to check index consistency since the method GetIndex of IComplexProperty is zero-based } [ComplexType] public class TaxInfoExtended { public int Code { get; set; } public decimal Percentage { get; set; } }

• however many time this kind of classes are defined in assemblies outside the control of the user (as requested from https://github.com/masesgroup/KEFCore/issues/445 the OPC-UA types are defined in specific assemblies) and EF Core offers the way to declare them programmatically:

  protected override void OnModelCreating(ModelBuilder modelBuilder) { modelBuilder.Entity<TaxInfo>(x => { x.ComplexProperty(y => y.TaxInfoExtended, y => { y.IsRequired(); }); }); base.OnModelCreating(modelBuilder); }

If the user is using the Json serialization there is no problem with the type, but with AVRO and Protobuf the type cannot be managed. A serializer shall be defined and registered:

The previous in only an example which converts, and converts back, TaxInfoExtended using a string which can be managed from underlying sub-system.