1 Steve Gordon (Senior Engineer @ Elastic) @stevejgordon | stevejgordon.co.uk The New Elasticsearch .NET Client Getting Started and Behind the Scenes

2 Agenda • Introduction to Elasticsearch • .NET Client for Elasticsearch ‒ Problems with the existing client ‒ Introducing the new v8 client • Demos • Behind the scenes ‒ Building the Elasticsearch specification ‒ Code generation of the new .NET client

We build search solutions on a single stack Enterprise Search Observability Security

4 Store, Search, & Analyze Visualize & Manage Ingest Elastic Stack SOLUTIONS Kibana Elasticsearch Beats Logstash SaaS On-Prem Elastic cloud Elastic cloud Enterprise Standalone Elastic cloud On Kubernetes Elastic Agent

5 CLOUD DEMO

6 Elasticsearch Terminology

7 Basic Terminology CLUSTER A collection of one or more nodes (servers) that together hold your data and provide federated indexing and search capabilities across all nodes.

8 Basic Terminology CLUSTER A single server that is part of your cluster, stores your data, and participates in the clusters indexing and search capabilities. NODE 1 NODE 2 NODE 3

9 Basic Terminology CLUSTER A collection of documents that have somewhat similar characteristics. NODE 1 NODE 2 NODE 3 INDEX

10 Basic Terminology CLUSTER Elasticsearch provides the ability to subdivide your index into multiple pieces called shards. NODE 1 NODE 2 NODE 3 INDEX SHARD (PRIMARY) P1 SHARD (PRIMARY) P2 SHARD (PRIMARY) P3 SHARD (REPLICA) R3 SHARD (REPLICA) R1 SHARD (REPLICA) R2

11 Basic Terminology CLUSTER Elasticsearch provides the ability to subdivide your index into multiple pieces called shards. NODE 1 NODE 2 NODE 3 INDEX SHARD (PRIMARY) P1 SHARD (PRIMARY) P2 SHARD (PRIMARY) P3 SHARD (REPLICA) R3 SHARD (REPLICA) R1 SHARD (REPLICA) R2

12 Basic Terminology CLUSTER The basic unit of information that can be indexed in JSON form. NODE 1 NODE 2 NODE 3 INDEX SHARD (PRIMARY) P1 SHARD (PRIMARY) P2 SHARD (PRIMARY) P3 SHARD (REPLICA) R3 SHARD (REPLICA) R1 SHARD (REPLICA) R2 DOC DOC DOC DOC DOC DOC DOC DOC DOC DOC DOC DOC

13 13 HTTP Interface

14 The Elasticsearch API in Numbers • > 400 API endpoints • > 2000 data structures ‒ 50 query types ‒ 70 aggregation types ‒ 30 field types

15 Language Clients • .NET • Java • JavaScript • Ruby • Go • PHP • Perl • Python • Rust

16 Existing Elasticsearch .NET Client (7.x) NEST High Level Client • Methods for every API • Strongly-typed requests & responses • Aggregations • Mappings • Query DSL • Fluent syntax • Helpers Elasticsearch.Net Low Level Client • Dependency free, unopinionated client. • Handles transport • Client-side load balancing • Request parameters (query string) • Serialisation • API URL Resolution

17 Architecture of Existing .NET Client NEST Elasticsearch.NET HttpWebRequest or HttpClient Elasticsearch Server HTTP

18 Problems with the Existing Client • Hand written ‒ API is not always consistent ‒ A lot of maintenance work (400 endpoints and thousands of types!) • Legacy internalised JSON serialiser based on Utf8Json • ~12 years of historical decisions

19

20 Introducing Elastic.Clients.Elasticsearch • A new generation of the Elasticsearch client • Code Generated ‒ Based on a formal specification of the Elasticsearch API • Uses System.Text.Json serializer • Built on a common Elastic.Transport layer • Removes some of the legacy of the past to create a cleaner API The new .NET client for v8.0

21 Architecture of New .NET Client Elastic.Clients.Elasticsearch Elastic.Transport HttpWebRequest or HttpClient Elasticsearch Server HTTP Other Clients

22 DEMOS

23 Building the Elasticsearch API specification

24 REST API specification → OpenAPI? OpenAPI is too limited • Elasticsearch API is complex and not “canonical” • Would require custom extensions • Our problem is mostly about data structures, not so much URLs OpenAPI is complex • “The Schema Object is a superset of the JSON Schema Specification Draft 2020-12” 😱 • 400 endpoint, 2000 structures… in YAML/JSON 😓

25 JSON API Specification → TypeScript! TypeScript’s type system is built to represent JSON/JS • Static type checking of the API • Strong IDE support • ts-morph: a library to build TS code processors ‒ Setup, navigation, and manipulation of the TypeScript AST can be a challenge. This library wraps the TypeScript compiler API so it's simple.

27 Example: Search Request /** * @rest_spec_name search * @since 0.0.0 * @stability stable */ export interface Request extends RequestBase { path_parts: { index?: Indices } query_parameters: { allow_no_indices?: Boolean ... size?: integer from?: integer sort?: string | string[] } body: { /** @aliases aggs */ // ES uses "aggregations" in serialization aggregations?: Dictionary collapse?: FieldCollapse /** * If true, returns detailed information about score computation as part of a hit. * @server_default false */ explain?: boolean * @server_default 0 */ from?: integer ... } export type IndexName = string export type Indices = IndexName | IndexName[] Meta information Alias tag Documentation comment

28 export class Response { body: ResponseBody } export class ResponseBody { took: long timed_out: boolean _shards: ShardStatistics hits: HitsMetadata aggregations?: Dictionary _clusters?: ClusterStatistics fields?: Dictionary max_score?: double num_reduce_phases?: long profile?: Profile pit_id?: Id _scroll_id?: ScrollId suggest?: Dictionary[]> terminated_early?: boolean } Example: Search Response User-provided type

29 export class Response { body: ResponseBody } export class ResponseBody { took: long timed_out: boolean _shards: ShardStatistics hits: HitsMetadata aggregations?: Dictionary _clusters?: ClusterStatistics fields?: Dictionary max_score?: double num_reduce_phases?: long profile?: Profile pit_id?: Id _scroll_id?: ScrollId suggest?: Dictionary[]> terminated_early?: boolean } Example: Search Response export class HitsMetadata { total?: TotalHits | long hits: Hit[] max_score?: double | null } export class HitMetadata { _id: Id _index: IndexName _primary_term: long _routing: string _seq_no: SequenceNumber _source: TDocument _version: VersionNumber }

30 Example: Query /** * @variants container * @non_exhaustive * @doc_id query-dsl */ export class QueryContainer { bool?: BoolQuery boosting?: BoostingQuery /** @deprecated 7.3.0 */ common?: SingleKeyDictionary /** @since 7.13.0 */ combined_fields?: CombinedFieldsQuery constant_score?: ConstantScoreQuery dis_max?: DisMaxQuery distance_feature?: DistanceFeatureQuery exists?: ExistsQuery function_score?: FunctionScoreQuery fuzzy?: SingleKeyDictionary geo_bounding_box?: GeoBoundingBoxQuery geo_distance?: GeoDistanceQuery geo_polygon?: GeoPolygonQuery ... Container variant is used for types that contain all the variants inside the definition Properties can be tagged as deprecation since a particular version We also track versions where new properties have been added

31 Validating the Specification • Piggy-back on Elasticsearch integration tests ‒ Capture request and response JSON ‒ Does it fit in the corresponding TS type? ‒ > 5400 validation tests!

32 Validating the Specification

33 Code Generation

34 TypeScript to Code Generating code from the TypeScript AST • Too low level • Not constrained enough Transform TypeScript to a simpler schema • Tailor-made for Elastic’s specific needs • Simple unambiguous meta-model

35 Code Generation Pipeline Spec compiler schema.json TypeScript API request & response bodies specification.ts Endpoints, Request & response bodies + Rich annotations .NET Code Generator .NET Client More Code Generators Java, Go, JS, Python, Rust, Ruby, PHP clients OpenAPI OpenAPI API Docs Even more generators

36 .NET Code Generator Process Deserialise JSON Build contexts Mark and enrich contexts Build Roslyn AST Write .cs files

37 .NET Code Generator • Establish naming and namespaces for generated types • Walk type hierarchy • Identify relationships • Mark request types • Mark containers and variants • Simplify type aliases to built-in types • Determine which types require which descriptors • Mark specialised serialisation needs (Bulk etc.) Marking and Enrichment

38 .NET Code Generator • Roslyn includes a very rich API • Not a great deal of documentation • roslynquoter.azurewebsites.net Build Roslyn AST https://docs.microsoft.com/en-us/dotnet/csharp/roslyn-sdk/compiler-api-model

39 Generator Code

40 // ** REQUEST PARAMETERS var requestParametersClass = ClassDeclaration(request.RequestParametersName) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.SealedKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"RequestParameters<{request.RequestParametersName}>"))) .AddMembers(request.QueryStringParameters.Select(a => a.QueryParameterProperty()) .Where(p => p is not null).ToArray()); AddClass(requestParametersClass); var (constructors, descriptorConstructors) = CreateConstructors(); // ** REQUEST CLASS var requestClass = ClassDeclaration(request.TypeInfo.Name) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.PartialKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"PlainRequestBase<{request.RequestParametersName}>"))) .AddMembers(constructors.ToArray()) .AddMembers(request.GetCommonRequestProperties()) .AddMembers(request.GenericArguments.Where(x => x.Name == "TDocument") .Select(p => p.GenericPropertySyntax()).Where(p => p is not null).ToArray()) .AddMembers(request.QueryStringParameters.Select(p => p.QueryStringProperty()) .Where(p => p is not null).ToArray()) .AddMembers(request.BodyProperties.Select(p => p.SerializablePropertySyntax()) .Where(p => p is not null).ToArray()); ...

41 // ** REQUEST PARAMETERS var requestParametersClass = ClassDeclaration(request.RequestParametersName) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.SealedKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"RequestParameters<{request.RequestParametersName}>"))) .AddMembers(request.QueryStringParameters.Select(a => a.QueryParameterProperty()) .Where(p => p is not null).ToArray()); AddClass(requestParametersClass); var (constructors, descriptorConstructors) = CreateConstructors(); // ** REQUEST CLASS var requestClass = ClassDeclaration(request.TypeInfo.Name) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.PartialKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"PlainRequestBase<{request.RequestParametersName}>"))) .AddMembers(constructors.ToArray()) .AddMembers(request.GetCommonRequestProperties()) .AddMembers(request.GenericArguments.Where(x => x.Name == "TDocument") .Select(p => p.GenericPropertySyntax()).Where(p => p is not null).ToArray()) .AddMembers(request.QueryStringParameters.Select(p => p.QueryStringProperty()) .Where(p => p is not null).ToArray()) .AddMembers(request.BodyProperties.Select(p => p.SerializablePropertySyntax()) .Where(p => p is not null).ToArray()); ...

42 // ** REQUEST PARAMETERS var requestParametersClass = ClassDeclaration(request.RequestParametersName) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.SealedKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"RequestParameters<{request.RequestParametersName}>"))) .AddMembers(request.QueryStringParameters.Select(a => a.QueryParameterProperty()) .Where(p => p is not null).ToArray()); AddClass(requestParametersClass); var (constructors, descriptorConstructors) = CreateConstructors(); // ** REQUEST CLASS var requestClass = ClassDeclaration(request.TypeInfo.Name) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.PartialKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"PlainRequestBase<{request.RequestParametersName}>"))) .AddMembers(constructors.ToArray()) .AddMembers(request.GetCommonRequestProperties()) .AddMembers(request.GenericArguments.Where(x => x.Name == "TDocument") .Select(p => p.GenericPropertySyntax()).Where(p => p is not null).ToArray()) .AddMembers(request.QueryStringParameters.Select(p => p.QueryStringProperty()) .Where(p => p is not null).ToArray()) .AddMembers(request.BodyProperties.Select(p => p.SerializablePropertySyntax()) .Where(p => p is not null).ToArray()); ...

43 // ** REQUEST PARAMETERS var requestParametersClass = ClassDeclaration(request.RequestParametersName) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.SealedKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"RequestParameters<{request.RequestParametersName}>"))) .AddMembers(request.QueryStringParameters.Select(a => a.QueryParameterProperty()) .Where(p => p is not null).ToArray()); AddClass(requestParametersClass); var (constructors, descriptorConstructors) = CreateConstructors(); // ** REQUEST CLASS var requestClass = ClassDeclaration(request.TypeInfo.Name) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.PartialKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"PlainRequestBase<{request.RequestParametersName}>"))) .AddMembers(constructors.ToArray()) .AddMembers(request.GetCommonRequestProperties()) .AddMembers(request.GenericArguments.Where(x => x.Name == "TDocument") .Select(p => p.GenericPropertySyntax()).Where(p => p is not null).ToArray()) .AddMembers(request.QueryStringParameters.Select(p => p.QueryStringProperty()) .Where(p => p is not null).ToArray()) .AddMembers(request.BodyProperties.Select(p => p.SerializablePropertySyntax()) .Where(p => p is not null).ToArray()); ...

44 // ** REQUEST PARAMETERS var requestParametersClass = ClassDeclaration(request.RequestParametersName) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.SealedKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"RequestParameters<{request.RequestParametersName}>"))) .AddMembers(request.QueryStringParameters.Select(a => a.QueryParameterProperty()) .Where(p => p is not null).ToArray()); AddClass(requestParametersClass); var (constructors, descriptorConstructors) = CreateConstructors(); // ** REQUEST CLASS var requestClass = ClassDeclaration(request.TypeInfo.Name) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.PartialKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"PlainRequestBase<{request.RequestParametersName}>"))) .AddMembers(constructors.ToArray()) .AddMembers(request.GetCommonRequestProperties()) .AddMembers(request.GenericArguments.Where(x => x.Name == "TDocument") .Select(p => p.GenericPropertySyntax()).Where(p => p is not null).ToArray()) .AddMembers(request.QueryStringParameters.Select(p => p.QueryStringProperty()) .Where(p => p is not null).ToArray()) .AddMembers(request.BodyProperties.Select(p => p.SerializablePropertySyntax()) .Where(p => p is not null).ToArray()); ...

45 // ** REQUEST PARAMETERS var requestParametersClass = ClassDeclaration(request.RequestParametersName) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.SealedKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"RequestParameters<{request.RequestParametersName}>"))) .AddMembers(request.QueryStringParameters.Select(a => a.QueryParameterProperty()) .Where(p => p is not null).ToArray()); AddClass(requestParametersClass); var (constructors, descriptorConstructors) = CreateConstructors(); // ** REQUEST CLASS var requestClass = ClassDeclaration(request.TypeInfo.Name) .AddModifiers(Token(SyntaxKind.PublicKeyword), Token(SyntaxKind.PartialKeyword)) .AddBaseListTypes(SimpleBaseType(ParseName($"PlainRequestBase<{request.RequestParametersName}>"))) .AddMembers(constructors.ToArray()) .AddMembers(request.GetCommonRequestProperties()) .AddMembers(request.GenericArguments.Where(x => x.Name == "TDocument") .Select(p => p.GenericPropertySyntax()).Where(p => p is not null).ToArray()) .AddMembers(request.QueryStringParameters.Select(p => p.QueryStringProperty()) .Where(p => p is not null).ToArray()) .AddMembers(request.BodyProperties.Select(p => p.SerializablePropertySyntax()) .Where(p => p is not null).ToArray()); ...

46 public static PropertyDeclarationSyntax CreateSerializableProperty(PropertyV2 property, bool selfDeserialisable = false) { if (TryResolveMemberToPropertyType(property.Type, property, out var typeSyntax)) { var propertyDeclaration = PropertyDeclaration(typeSyntax.TypeSyntax, Identifier(property.CodegenName)) .AddModifiers(Token(SyntaxKind.PublicKeyword)); propertyDeclaration = propertyDeclaration.AddAttributeLists( AttributeList(SingletonSeparatedList(Attribute(IdentifierName("JsonInclude")))), AttributeList(SingletonSeparatedList(Attribute(IdentifierName("JsonPropertyName")) .AddArgumentListArguments(AttributeArgument( LiteralExpression(SyntaxKind.StringLiteralExpression, Literal(property.JsonName))))))); if (property.IsSourceProperty) propertyDeclaration = propertyDeclaration.AddAttributeLists( AttributeList(SingletonSeparatedList(Attribute(IdentifierName("SourceConverter"))))); ...

47 public static PropertyDeclarationSyntax CreateSerializableProperty(PropertyV2 property, bool selfDeserialisable = false) { if (TryResolveMemberToPropertyType(property.Type, property, out var typeSyntax)) { var propertyDeclaration = PropertyDeclaration(typeSyntax.TypeSyntax, Identifier(property.CodegenName)) .AddModifiers(Token(SyntaxKind.PublicKeyword)); propertyDeclaration = propertyDeclaration.AddAttributeLists( AttributeList(SingletonSeparatedList(Attribute(IdentifierName("JsonInclude")))), AttributeList(SingletonSeparatedList(Attribute(IdentifierName("JsonPropertyName")) .AddArgumentListArguments(AttributeArgument( LiteralExpression(SyntaxKind.StringLiteralExpression, Literal(property.JsonName))))))); if (property.IsSourceProperty) propertyDeclaration = propertyDeclaration.AddAttributeLists( AttributeList(SingletonSeparatedList(Attribute(IdentifierName("SourceConverter"))))); ...

48 ... if (property.OwningType.UsedInRequest) { propertyDeclaration = propertyDeclaration.AddAccessorListAccessors( AccessorDeclaration(SyntaxKind.GetAccessorDeclaration) .WithSemicolonToken(Token(SyntaxKind.SemicolonToken)), AccessorDeclaration(SyntaxKind.SetAccessorDeclaration) .WithSemicolonToken(Token(SyntaxKind.SemicolonToken))); } else { propertyDeclaration = propertyDeclaration.AddAccessorListAccessors( AccessorDeclaration(SyntaxKind.GetAccessorDeclaration) .WithSemicolonToken(Token(SyntaxKind.SemicolonToken)), AccessorDeclaration(SyntaxKind.InitAccessorDeclaration) .WithSemicolonToken(Token(SyntaxKind.SemicolonToken))); } return propertyDeclaration; } return null; }

49 Resources • github.com/elastic/elasticsearch-net • elastic.co/guide/en/elasticsearch/client/net-api • nuget.org/packages/NEST • nuget.org/packages/Elastic.Clients.Elasticsearch • github.com/stevejgordon/elasticsearch-examples • github.com/elastic/elasticsearch-specification • discuss.elastic.co

50 Thank you! @stevejgordon