Adds schema

Signed-off-by: Takeshi Yoneda <[email protected]>
envoyproxy · Dec 3, 2024 · 897fb3b · 897fb3b
1 parent 6cee7bc
commit 897fb3b
Show file tree

Hide file tree

Showing 4 changed files with 116 additions and 1 deletion.
diff --git a/api/v1alpha1/api.go b/api/v1alpha1/api.go
@@ -11,6 +11,10 @@ import (
 //
 // This serves as a way to define a "unified" LLM API for a Gateway which allows downstream
 // clients to use a single schema API to interact with multiple LLM backends.
+//
+// The InputSchema is used to determine the structure of the requests that the Gateway will
+// receive. And then the Gateway will route the traffic to the appropriate LLMBackend based
+// on the output schema of the LLMBackend while doing the other necessary jobs like rate limit, etc.
 type LLMRoute struct {
 	metav1.TypeMeta   `json:",inline"`
 	metav1.ObjectMeta `json:"metadata,omitempty"`
@@ -29,6 +33,10 @@ type LLMRouteList struct {
 
 // LLMRouteSpec details the LLMRoute configuration.
 type LLMRouteSpec struct {
+	// APISchema specifies the API schema of the input that the target Gateway(s) will receive.
+	// Based on this schema, the ai-gateway will perform the necessary transformation to the
+	// output schema specified in the selected LLMBackend during the routing process.
+	APISchema LLMAPISchema `json:"inputSchema"`
 	// TargetRefs are the names of the Gateway resources this policy is being attached to.
 	// The namespace is "local", i.e. the same namespace as the LLMRoute.
 	//
@@ -69,4 +77,43 @@ type LLMBackendList struct {
 }
 
 // LLMBackendSpec details the LLMBackend configuration.
-type LLMBackendSpec struct{}
+type LLMBackendSpec struct {
+	// APISchema specifies the API schema of the output that this LLMBackend can accept.
+	// Based on this schema, the ai-gateway will perform the necessary transformation for
+	// the pair of LLMRoute.APISchema and LLMBackend.APISchema.
+	//
+	// This is required to be set.
+	APISchema LLMAPISchema `json:"outputSchema"`
+}
+
+// LLMAPISchema defines the API schema of either LLMRoute (the input) or LLMBackend (the output).
+//
+// This allows the ai-gateway to understand the input and perform the necessary transformation
+// depending on the API schema pair (input, output).
+//
+// Note that this is vendor specific, and the compatibility of the input and output API schema
+// is determined by them.
+type LLMAPISchema struct {
+	// Schema is the API schema of the LLMRoute or LLMBackend.
+	//
+	// +kubebuilder:validation:Enum=OpenAI;AWSBedrock
+	Schema APISchema `json:"schema"`
+
+	// Version is the version of the API schema.
+	// This will be used to determine the compatibility of the input and output API schema.
+	Version string `json:"version,omitempty"`
+}
+
+// APISchema defines the API schema.
+type APISchema string
+
+const (
+	// APISchemaOpenAI is the OpenAI schema.
+	//
+	// https://platform.openai.com/docs/overview
+	APISchemaOpenAI APISchema = "OpenAI"
+	// APISchemaAWSBedrock is the AWS Bedrock schema.
+	//
+	// https://docs.aws.amazon.com/bedrock/latest/APIReference/API_Operations_Amazon_Bedrock_Runtime.html
+	APISchemaAWSBedrock APISchema = "AWSBedrock"
+)
diff --git a/api/v1alpha1/zz_generated.deepcopy.go b/api/v1alpha1/zz_generated.deepcopy.go
diff --git a/manifests/charts/ai-gateway-helm/crds/aigateway.envoyproxy.io_llmbackends.yaml b/manifests/charts/ai-gateway-helm/crds/aigateway.envoyproxy.io_llmbackends.yaml
@@ -40,6 +40,31 @@ spec:
             type: object
           spec:
             description: Spec defines the details of the LLM policy.
+            properties:
+              outputSchema:
+                description: |-
+                  APISchema specifies the API schema of the output that this LLMBackend can accept.
+                  Based on this schema, the ai-gateway will perform the necessary transformation for
+                  the pair of LLMRoute.APISchema and LLMBackend.APISchema.
+
+                  This is required to be set.
+                properties:
+                  schema:
+                    description: Schema is the API schema of the LLMRoute or LLMBackend.
+                    enum:
+                    - OpenAI
+                    - AWSBedrock
+                    type: string
+                  version:
+                    description: |-
+                      Version is the version of the API schema.
+                      This will be used to determine the compatibility of the input and output API schema.
+                    type: string
+                required:
+                - schema
+                type: object
+            required:
+            - outputSchema
             type: object
         type: object
     served: true

diff --git a/manifests/charts/ai-gateway-helm/crds/aigateway.envoyproxy.io_llmroutes.yaml b/manifests/charts/ai-gateway-helm/crds/aigateway.envoyproxy.io_llmroutes.yaml
@@ -22,6 +22,10 @@ spec:
 
           This serves as a way to define a "unified" LLM API for a Gateway which allows downstream
           clients to use a single schema API to interact with multiple LLM backends.
+
+          The InputSchema is used to determine the structure of the requests that the Gateway will
+          receive. And then the Gateway will route the traffic to the appropriate LLMBackend based
+          on the output schema of the LLMBackend while doing the other necessary jobs like rate limit, etc.
         properties:
           apiVersion:
             description: |-
@@ -60,6 +64,26 @@ spec:
                   type: object
                 maxItems: 128
                 type: array
+              inputSchema:
+                description: |-
+                  APISchema specifies the API schema of the input that the target Gateway(s) will receive.
+                  Based on this schema, the ai-gateway will perform the necessary transformation to the
+                  output schema specified in the selected LLMBackend during the routing process.
+                properties:
+                  schema:
+                    description: Schema is the API schema of the LLMRoute or LLMBackend.
+                    enum:
+                    - OpenAI
+                    - AWSBedrock
+                    type: string
+                  version:
+                    description: |-
+                      Version is the version of the API schema.
+                      This will be used to determine the compatibility of the input and output API schema.
+                    type: string
+                required:
+                - schema
+                type: object
               targetRefs:
                 description: |-
                   TargetRefs are the names of the Gateway resources this policy is being attached to.
@@ -116,6 +140,8 @@ spec:
                   type: object
                 maxItems: 128
                 type: array
+            required:
+            - inputSchema
             type: object
         type: object
     served: true