@salesforce/lds-adapters-service-einsteinllm 1.414.1 → 1.415.0

package/src/raml/api.raml

@@ -435,6 +435,10 @@ types:
         description: Additional configuration to fine tune model parameters
         type: EinsteinLlmAdditionalConfigInputRepresentation # Hand-rolled update from 'object'
         required: false # TODO Hand-rolled W-9314597
+      includeMetrics:
+        description: Whether to return metrics
+        type: boolean
+        required: false
       inputParams:
         description: Dictionary of input parameters to generate prompt from prompt
           template.
@@ -481,6 +485,11 @@ types:
         description: Any provider specific attributes will be included as part of
           this object
         type: WrappedMap | nil # Hand-rolled union 'nil'
+      performance:
+        description: Performance metrics including overall latency and detailed execution
+          breakdown
+        required: false
+        type: EinsteinPromptTemplatePerformanceRepresentation
       prompt:
         description: Prompt used for the generation
         type: string | nil # Hand-rolled union 'nil'
@@ -520,6 +529,10 @@ types:
         required: false # TODO Hand-rolled W-17552098
         items:
           type: EinsteinPromptTemplateAttachmentRepresentation
+      tokenUsage:
+        description: token usage information
+        required: false
+        type: EinsteinPromptTemplateTokenUsageRepresentation
   EinsteinLlmGenerationCitationRepresentation:
     description: Output of a Einstein LLM Generation Citation output
     type: object
@@ -984,6 +997,121 @@ types:
       versionId:
         description: Prompt Template Version Id
         type: string
+  EinsteinPromptTemplatePerformanceRepresentation:
+    description: Performance metrics including overall latency and detailed execution
+      breakdown
+    type: object
+    properties:
+      execution:
+        description: Detailed latency breakdown for all execution steps
+        type: EinsteinPromptTemplateLatencyExecutionRepresentation
+      latency:
+        description: Overall latency metrics for the entire operation
+        type: EinsteinPromptTemplateLatencyRepresentation
+  EinsteinPromptTemplateLatencyBreakdownItemRepresentation:
+    description: Latency breakdown item with reference name and latency metrics
+    type: object
+    properties:
+      latency:
+        description: Latency information for this breakdown item
+        type: EinsteinPromptTemplateLatencyRepresentation
+      referenceName:
+        description: Reference name for the breakdown item
+        type: string
+  EinsteinPromptTemplateLatencyExecutionRepresentation:
+    description: Latency execution breakdown with metrics for all execution steps
+    type: object
+    properties:
+      generation:
+        description: Latency metrics for generation step
+        type: EinsteinPromptTemplateLatencyStepRepresentation
+        required: false
+      hydration:
+        description: Latency metrics for hydration step
+        type: EinsteinPromptTemplateLatencyStepRepresentation
+        required: false
+      loadTemplate:
+        description: Latency metrics for loading the template
+        type: EinsteinPromptTemplateLatencyStepRepresentation
+      mask:
+        description: Latency metrics for masking step
+        type: EinsteinPromptTemplateLatencyStepRepresentation
+        required: false
+      resolveData:
+        description: Latency metrics for resolving all data (inputs and data providers)
+          with breakdown
+        type: EinsteinPromptTemplateLatencyStepRepresentation
+        required: false
+      summarize:
+        description: Latency metrics for summarization step
+        type: EinsteinPromptTemplateLatencyStepRepresentation
+        required: false
+  EinsteinPromptTemplateLatencyRepresentation:
+    description: Latency information with duration and start time
+    type: object
+    properties:
+      duration:
+        description: Duration of the operation in milliseconds
+        type: integer
+      startTime:
+        description: Start time of the operation in milliseconds
+        type: integer
+  EinsteinPromptTemplateLatencyStepRepresentation:
+    description: Latency step information with latency metrics and breakdown
+    type: object
+    properties:
+      breakdown:
+        description: Breakdown of latency by reference name
+        type: array
+        required: false
+        items:
+          type: EinsteinPromptTemplateLatencyBreakdownItemRepresentation
+      latency:
+        description: Latency information for this step
+        type: EinsteinPromptTemplateLatencyRepresentation
+  EinsteinPromptTemplatePromptTokenUsageRepresentation:
+    description: Output representation for prompt token usage with breakdown
+    type: object
+    properties:
+      breakdown:
+        description: Breakdown of token usage by source (text expressions, files,
+          etc.)
+        required: false
+        type: array
+        items:
+          type: EinsteinPromptTemplateTokenBreakdownRepresentation
+      total:
+        description: Total number of tokens used in the prompt
+        type: integer
+  EinsteinPromptTemplateTokenUsageRepresentation:
+    description: Output representation for token usage and its breakdown
+    type: object
+    properties:
+      completion:
+        description: Number of tokens used for completion
+        type: integer
+      prompt:
+        description: Prompt token usage details including breakdown by source
+        type: EinsteinPromptTemplatePromptTokenUsageRepresentation
+      total:
+        description: Total number of tokens used in the request
+        type: integer
+  EinsteinPromptTemplateTokenBreakdownRepresentation:
+    description: Output representation for a single token breakdown item by source
+    type: object
+    properties:
+      occurrences:
+        description: Number of times this expression appears in the prompt
+        type: integer
+      role:
+        description: Role associated with source of token.
+        type: string
+      source:
+        description: Specific part of prompt that contributed to tokens.
+        type: string
+      tokens:
+        description: Number of tokens contributed by this source
+        type: integer
   WrappedListString:
     description: Wrapped list of String for use in Apex.
     type: object