Merge pull request #244 from bjwswang/charts-dev

feat: create embedder&llm for worker based on model type

api/v1alpha1/common.go

@@ -25,6 +25,38 @@ const (
 	Finalizer = Group + "/finalizer"
 )
 
+type ProviderType string
+
+const (
+	ProviderTypeUnknown  ProviderType = "unknown"
+	ProviderType3rdParty ProviderType = "3rd_party"
+	ProviderTypeWorker   ProviderType = "worker"
+)
+
+// Provider defines how to prvoide the service
+type Provider struct {
+	// Enpoint defines connection info
+	Enpoint *Endpoint `json:"endpoint,omitempty"`
+
+	// Worker defines the worker info
+	// Means this LLM is provided by a arcadia worker
+	Worker *TypedObjectReference `json:"worker,omitempty"`
+}
+
+// GetType returnes the type of this provider
+func (p Provider) GetType() ProviderType {
+	// if endpoint provided,then 3rd_party
+	if p.Enpoint != nil {
+		return ProviderType3rdParty
+	}
+	// if worker provided,then worker
+	if p.Worker != nil {
+		return ProviderTypeWorker
+	}
+
+	return ProviderTypeUnknown
+}
+
 // After we use k8s.io/api v1.26.0, we can remove this types to use corev1.TypedObjectReference
 // that types is introduced in https://github.com/kubernetes/kubernetes/pull/113186
 

api/v1alpha1/embedder_types.go

@@ -32,8 +32,8 @@ type EmbedderSpec struct {
 	// ServiceType indicates the source type of embedding service
 	ServiceType embeddings.EmbeddingType `json:"serviceType,omitempty"`
 
-	// Enpoint defines connection info
-	Enpoint *Endpoint `json:"endpoint,omitempty"`
+	// Provider defines the provider info which provide this embedder service
+	Provider `json:"provider,omitempty"`
 }
 
 // EmbeddingsStatus defines the observed state of Embedder

api/v1alpha1/llm_types.go

@@ -29,8 +29,8 @@ type LLMSpec struct {
 	// Type defines the type of llm
 	Type llms.LLMType `json:"type"`
 
-	// Enpoint defines connection info
-	Enpoint *Endpoint `json:"endpoint,omitempty"`
+	// Provider defines the provider info which provide this llm service
+	Provider `json:"provider,omitempty"`
 }
 
 // LLMStatus defines the observed state of LLM

api/v1alpha1/model.go

@@ -18,19 +18,35 @@ package v1alpha1
 
 import (
 	"fmt"
+	"strings"
 )
 
 const (
-	// LabelModelTypes keeps the spec.types field
-	LabelModelTypes    = Group + "/model-types"
+	// LabelModelEmbedding indicates this is a embedding model
+	LabelModelEmbedding = Group + "/embedding"
+	// LabelModelLLM indicates this is a llm model
+	LabelModelLLM      = Group + "/llm"
 	LabelModelFullPath = Group + "/full-path"
 )
 
-func (model Model) ModelTypes() string {
-	if model.Spec.Types == "" {
-		return "unknown"
+// IsLLMModel checks whether this model is a llm model
+func (model Model) IsLLMModel() bool {
+	for _, t := range strings.Split(model.Spec.Types, ",") {
+		if strings.ToLower(t) == "llm" {
+			return true
+		}
 	}
-	return model.Spec.Types
+	return false
+}
+
+// IsLLMModel checks whether this model is a embedding model
+func (model Model) IsEmbeddingModel() bool {
+	for _, t := range strings.Split(model.Spec.Types, ",") {
+		if strings.ToLower(t) == "embedding" {
+			return true
+		}
+	}
+	return false
 }
 
 // FullPath with bucket and object path

api/v1alpha1/worker.go

@@ -17,8 +17,14 @@ limitations under the License.
 package v1alpha1
 
 import (
+	"crypto/sha256"
+	"encoding/hex"
+
 	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+
+	"github.com/kubeagi/arcadia/pkg/embeddings"
+	"github.com/kubeagi/arcadia/pkg/llms"
 )
 
 type WorkerType string
@@ -72,3 +78,65 @@ func (worker Worker) ErrorCondition(msg string) Condition {
 		LastSuccessfulTime: metav1.Now(),
 	}
 }
+
+func (worker Worker) generateUniqueName() string {
+	// Create a new SHA-256 hasher
+	hasher := sha256.New()
+
+	// Write the input string to the hasher
+	hasher.Write([]byte(worker.Name))
+
+	// Calculate the hash sum
+	hashSum := hasher.Sum(nil)
+
+	// Convert the hash sum to a hexadecimal string
+	hashString := hex.EncodeToString(hashSum)
+
+	return hashString[:10]
+}
+
+func (worker Worker) BuildEmbedder() *Embedder {
+	return &Embedder{
+		ObjectMeta: metav1.ObjectMeta{
+			Namespace: worker.Namespace,
+			Name:      worker.generateUniqueName() + "-worker",
+		},
+		Spec: EmbedderSpec{
+			CommonSpec: CommonSpec{
+				Creator:     worker.Spec.Creator,
+				Description: "Embedder created by Worker(OpenAI compatible)",
+			},
+			ServiceType: embeddings.OpenAI,
+			Provider: Provider{
+				Worker: &TypedObjectReference{
+					Kind:      "Worker",
+					Namespace: &worker.Namespace,
+					Name:      worker.Name,
+				},
+			},
+		},
+	}
+}
+
+func (worker Worker) BuildLLM() *LLM {
+	return &LLM{
+		ObjectMeta: metav1.ObjectMeta{
+			Namespace: worker.Namespace,
+			Name:      worker.generateUniqueName() + "-worker",
+		},
+		Spec: LLMSpec{
+			CommonSpec: CommonSpec{
+				Creator:     worker.Spec.Creator,
+				Description: "LLM created by Worker(OpenAI compatible)",
+			},
+			Type: llms.OpenAI,
+			Provider: Provider{
+				Worker: &TypedObjectReference{
+					Kind:      "Worker",
+					Namespace: &worker.Namespace,
+					Name:      worker.Name,
+				},
+			},
+		},
+	}
+}

config/crd/bases/arcadia.kubeagi.k8s.com.cn_embedders.yaml

@@ -48,13 +48,48 @@ spec:
               displayName:
                 description: DisplayName defines datasource display name
                 type: string
-              endpoint:
-                description: Enpoint defines connection info
+              provider:
+                description: Provider defines the provider info which provide this
+                  embedder service
                 properties:
-                  authSecret:
-                    description: AuthSecret if the chart repository requires auth
-                      authentication, set the username and password to secret, with
-                      the field user and password respectively.
+                  endpoint:
+                    description: Enpoint defines connection info
+                    properties:
+                      authSecret:
+                        description: AuthSecret if the chart repository requires auth
+                          authentication, set the username and password to secret,
+                          with the field user and password respectively.
+                        properties:
+                          apiGroup:
+                            description: APIGroup is the group for the resource being
+                              referenced. If APIGroup is not specified, the specified
+                              Kind must be in the core API group. For any other third-party
+                              types, APIGroup is required.
+                            type: string
+                          kind:
+                            description: Kind is the type of resource being referenced
+                            type: string
+                          name:
+                            description: Name is the name of resource being referenced
+                            type: string
+                          namespace:
+                            description: Namespace is the namespace of resource being
+                              referenced
+                            type: string
+                        required:
+                        - kind
+                        - name
+                        type: object
+                      insecure:
+                        description: Insecure if the endpoint needs a secure connection
+                        type: boolean
+                      url:
+                        description: URL chart repository address
+                        type: string
+                    type: object
+                  worker:
+                    description: Worker defines the worker info Means this LLM is
+                      provided by a arcadia worker
                     properties:
                       apiGroup:
                         description: APIGroup is the group for the resource being
@@ -76,12 +111,6 @@ spec:
                     - kind
                     - name
                     type: object
-                  insecure:
-                    description: Insecure if the endpoint needs a secure connection
-                    type: boolean
-                  url:
-                    description: URL chart repository address
-                    type: string
                 type: object
               serviceType:
                 description: ServiceType indicates the source type of embedding service

config/crd/bases/arcadia.kubeagi.k8s.com.cn_llms.yaml

@@ -44,13 +44,48 @@ spec:
               displayName:
                 description: DisplayName defines datasource display name
                 type: string
-              endpoint:
-                description: Enpoint defines connection info
+              provider:
+                description: Provider defines the provider info which provide this
+                  llm service
                 properties:
-                  authSecret:
-                    description: AuthSecret if the chart repository requires auth
-                      authentication, set the username and password to secret, with
-                      the field user and password respectively.
+                  endpoint:
+                    description: Enpoint defines connection info
+                    properties:
+                      authSecret:
+                        description: AuthSecret if the chart repository requires auth
+                          authentication, set the username and password to secret,
+                          with the field user and password respectively.
+                        properties:
+                          apiGroup:
+                            description: APIGroup is the group for the resource being
+                              referenced. If APIGroup is not specified, the specified
+                              Kind must be in the core API group. For any other third-party
+                              types, APIGroup is required.
+                            type: string
+                          kind:
+                            description: Kind is the type of resource being referenced
+                            type: string
+                          name:
+                            description: Name is the name of resource being referenced
+                            type: string
+                          namespace:
+                            description: Namespace is the namespace of resource being
+                              referenced
+                            type: string
+                        required:
+                        - kind
+                        - name
+                        type: object
+                      insecure:
+                        description: Insecure if the endpoint needs a secure connection
+                        type: boolean
+                      url:
+                        description: URL chart repository address
+                        type: string
+                    type: object
+                  worker:
+                    description: Worker defines the worker info Means this LLM is
+                      provided by a arcadia worker
                     properties:
                       apiGroup:
                         description: APIGroup is the group for the resource being
@@ -72,12 +107,6 @@ spec:
                     - kind
                     - name
                     type: object
-                  insecure:
-                    description: Insecure if the endpoint needs a secure connection
-                    type: boolean
-                  url:
-                    description: URL chart repository address
-                    type: string
                 type: object
               type:
                 description: Type defines the type of llm