@f5-sales-demo/pi-ai 20.13.2 → 20.15.0

package/package.json

@@ -1,7 +1,7 @@
 {
 	"type": "module",
 	"name": "@f5-sales-demo/pi-ai",
-	"version": "20.13.2",
+	"version": "20.15.0",
 	"description": "Unified LLM API with automatic model discovery and provider configuration",
 	"homepage": "https://github.com/f5-sales-demo/xcsh",
 	"author": "Can Boluk",
@@ -46,7 +46,7 @@
 		"@anthropic-ai/sdk": "^0.115",
 		"@aws-sdk/client-bedrock-runtime": "^3",
 		"@bufbuild/protobuf": "^2.11",
-		"@f5-sales-demo/pi-utils": "20.13.2",
+		"@f5-sales-demo/pi-utils": "20.15.0",
 		"@google/genai": "^2.15",
 		"@sinclair/typebox": "^0.34",
 		"@smithy/node-http-handler": "^4.4",

package/src/providers/google-gemini-cli.ts

@@ -69,6 +69,10 @@ const ANTIGRAVITY_VERTEX_MODELS: Readonly<Record<string, string>> = Object.freez
 	"gemini-3.1-pro-high-vertex": "gemini-3.1-pro-preview",
 });
 
+export function resolveAntigravityServingModelId(requestedModelId: string): string {
+	return ANTIGRAVITY_VERTEX_MODELS[requestedModelId] ?? requestedModelId;
+}
+
 /**
  * Build a User-Agent string that identifies as Gemini CLI to unlock higher rate limits.
  * Uses the same format as the official Gemini CLI (v0.35+):
@@ -477,6 +481,7 @@ export const streamGoogleGeminiCli: StreamFunction<"google-gemini-cli"> = (
 			api: "google-gemini-cli" as Api,
 			provider: model.provider,
 			model: model.id,
+			responseAttribution: { requestedModel: model.id },
 			usage: {
 				input: 0,
 				output: 0,
@@ -687,6 +692,14 @@ export const streamGoogleGeminiCli: StreamFunction<"google-gemini-cli"> = (
 					options?.signal,
 				)) {
 					const responseData = chunk.response ?? chunk;
+					if (responseData.modelVersion) {
+						output.responseAttribution = {
+							...(output.responseAttribution ?? { requestedModel: model.id }),
+							responseModel: responseData.modelVersion,
+							responseModelSource: "response-body",
+						};
+					}
+					if (responseData.responseId) output.responseId = responseData.responseId;
 
 					const candidate = responseData.candidates?.[0];
 					if (candidate?.content?.parts) {

package/src/providers/openai-codex-responses.ts

@@ -385,6 +385,7 @@ function createAssistantOutput(model: Model<"openai-codex-responses">): Assistan
 		api: "openai-codex-responses" as Api,
 		provider: model.provider,
 		model: model.id,
+		responseAttribution: { requestedModel: model.id },
 		usage: createEmptyUsage(),
 		stopReason: "stop",
 		timestamp: Date.now(),
@@ -816,7 +817,8 @@ function handleCodexStreamEvent(args: {
 	}
 
 	if (eventType === "response.created") {
-		return handleResponseCreated(runtime, rawEvent);
+		handleResponseCreated(model, output, runtime, rawEvent);
+		return firstTokenTime;
 	}
 
 	if (eventType === "response.completed" || eventType === "response.done" || eventType === "response.incomplete") {
@@ -1011,13 +1013,27 @@ function handleOutputItemDone(
 	void model;
 }
 
-function handleResponseCreated(runtime: CodexStreamRuntime, rawEvent: Record<string, unknown>): number | undefined {
-	const response = (rawEvent as { response?: { id?: string } }).response;
+function handleResponseCreated(
+	model: Model<"openai-codex-responses">,
+	output: AssistantMessage,
+	runtime: CodexStreamRuntime,
+	rawEvent: Record<string, unknown>,
+): void {
+	const response = (rawEvent as { response?: { id?: string; model?: string } }).response;
+	if (typeof response?.id === "string" && response.id.length > 0) {
+		output.responseId = response.id;
+	}
+	if (typeof response?.model === "string" && response.model.length > 0) {
+		output.responseAttribution = {
+			...(output.responseAttribution ?? { requestedModel: model.id }),
+			responseModel: response.model,
+			responseModelSource: "response-body",
+		};
+	}
 	const state = runtime.websocketState;
 	if (runtime.transport === "websocket" && state && typeof response?.id === "string" && response.id.length > 0) {
 		state.lastResponseId = response.id;
 	}
-	return undefined;
 }
 
 function handleResponseCompleted(
@@ -1038,6 +1054,7 @@ function handleResponseCompleted(
 					input_tokens_details?: { cached_tokens?: number };
 				};
 				status?: string;
+				model?: string;
 			};
 		}
 	).response;
@@ -1056,6 +1073,13 @@ function handleResponseCompleted(
 	if (typeof response?.id === "string" && response.id.length > 0) {
 		output.responseId = response.id;
 	}
+	if (typeof response?.model === "string" && response.model.length > 0) {
+		output.responseAttribution = {
+			...(output.responseAttribution ?? { requestedModel: model.id }),
+			responseModel: response.model,
+			responseModelSource: "response-body",
+		};
+	}
 
 	const state = runtime.websocketState;
 	if (runtime.transport === "websocket" && state) {