@llmops/app 0.6.2 → 0.6.3

package/dist/index.cjs

@@ -18384,7 +18384,7 @@ function getGlobalTraceBatchWriter(deps, config$1) {
 
 //#endregion
 //#region src/server/middlewares/costTracking.ts
-const pricingProvider = (0, __llmops_core.getDefaultPricingProvider)();
+const pricingProvider$1 = (0, __llmops_core.getDefaultPricingProvider)();
 /**
 * Transform gateway hook results to our schema format for telemetry
 */
@@ -18654,7 +18654,7 @@ async function processUsageAndLog(params) {
 	let outputCost = 0;
 	let cacheSavings = 0;
 	if (usage && usage.promptTokens + usage.completionTokens > 0) try {
-		const pricing = await pricingProvider.getModelPricing(provider, model);
+		const pricing = await pricingProvider$1.getModelPricing(provider, model);
 		if (pricing) {
 			const costResult = (0, __llmops_core.calculateCacheAwareCost)({
 				promptTokens: usage.promptTokens,
@@ -18873,6 +18873,7 @@ function extractTypedFields(attrs) {
 		output
 	};
 }
+const pricingProvider = (0, __llmops_core.getDefaultPricingProvider)();
 /**
 * OTLP ingestion endpoint
 * Accepts OTLP JSON (ExportTraceServiceRequest) format
@@ -18913,6 +18914,16 @@ const app$3 = new hono.Hono().post("/v1/traces", async (c) => {
 				const durationMs = endTime ? endTime.getTime() - startTime.getTime() : null;
 				const spanStatus = otlpSpan.status?.code ?? 0;
 				const traceStatus = spanStatus === 2 ? "error" : spanStatus === 1 ? "ok" : "unset";
+				let cost = 0;
+				if (typed.provider && typed.model && (typed.promptTokens > 0 || typed.completionTokens > 0)) try {
+					const pricing = await pricingProvider.getModelPricing(typed.provider, typed.model);
+					if (pricing) cost = (0, __llmops_core.calculateCacheAwareCost)({
+						promptTokens: typed.promptTokens,
+						completionTokens: typed.completionTokens
+					}, pricing, typed.provider).totalCost;
+				} catch (e) {
+					__llmops_core.logger.debug(`[OTLP] Failed to calculate cost for ${typed.provider}/${typed.model}: ${e instanceof Error ? e.message : String(e)}`);
+				}
 				const spanData = {
 					traceId: otlpSpan.traceId,
 					spanId: otlpSpan.spanId,
@@ -18929,7 +18940,7 @@ const app$3 = new hono.Hono().post("/v1/traces", async (c) => {
 					promptTokens: typed.promptTokens,
 					completionTokens: typed.completionTokens,
 					totalTokens: typed.totalTokens,
-					cost: 0,
+					cost,
 					source: "otlp",
 					input: typed.input,
 					output: typed.output,
@@ -18955,7 +18966,7 @@ const app$3 = new hono.Hono().post("/v1/traces", async (c) => {
 					totalInputTokens: typed.promptTokens,
 					totalOutputTokens: typed.completionTokens,
 					totalTokens: typed.totalTokens,
-					totalCost: 0,
+					totalCost: cost,
 					tags: {},
 					metadata: {}
 				};

package/dist/index.mjs

@@ -18356,7 +18356,7 @@ function getGlobalTraceBatchWriter(deps, config$1) {
 
 //#endregion
 //#region src/server/middlewares/costTracking.ts
-const pricingProvider = getDefaultPricingProvider();
+const pricingProvider$1 = getDefaultPricingProvider();
 /**
 * Transform gateway hook results to our schema format for telemetry
 */
@@ -18626,7 +18626,7 @@ async function processUsageAndLog(params) {
 	let outputCost = 0;
 	let cacheSavings = 0;
 	if (usage && usage.promptTokens + usage.completionTokens > 0) try {
-		const pricing = await pricingProvider.getModelPricing(provider, model);
+		const pricing = await pricingProvider$1.getModelPricing(provider, model);
 		if (pricing) {
 			const costResult = calculateCacheAwareCost({
 				promptTokens: usage.promptTokens,
@@ -18845,6 +18845,7 @@ function extractTypedFields(attrs) {
 		output
 	};
 }
+const pricingProvider = getDefaultPricingProvider();
 /**
 * OTLP ingestion endpoint
 * Accepts OTLP JSON (ExportTraceServiceRequest) format
@@ -18885,6 +18886,16 @@ const app$3 = new Hono().post("/v1/traces", async (c) => {
 				const durationMs = endTime ? endTime.getTime() - startTime.getTime() : null;
 				const spanStatus = otlpSpan.status?.code ?? 0;
 				const traceStatus = spanStatus === 2 ? "error" : spanStatus === 1 ? "ok" : "unset";
+				let cost = 0;
+				if (typed.provider && typed.model && (typed.promptTokens > 0 || typed.completionTokens > 0)) try {
+					const pricing = await pricingProvider.getModelPricing(typed.provider, typed.model);
+					if (pricing) cost = calculateCacheAwareCost({
+						promptTokens: typed.promptTokens,
+						completionTokens: typed.completionTokens
+					}, pricing, typed.provider).totalCost;
+				} catch (e) {
+					logger.debug(`[OTLP] Failed to calculate cost for ${typed.provider}/${typed.model}: ${e instanceof Error ? e.message : String(e)}`);
+				}
 				const spanData = {
 					traceId: otlpSpan.traceId,
 					spanId: otlpSpan.spanId,
@@ -18901,7 +18912,7 @@ const app$3 = new Hono().post("/v1/traces", async (c) => {
 					promptTokens: typed.promptTokens,
 					completionTokens: typed.completionTokens,
 					totalTokens: typed.totalTokens,
-					cost: 0,
+					cost,
 					source: "otlp",
 					input: typed.input,
 					output: typed.output,
@@ -18927,7 +18938,7 @@ const app$3 = new Hono().post("/v1/traces", async (c) => {
 					totalInputTokens: typed.promptTokens,
 					totalOutputTokens: typed.completionTokens,
 					totalTokens: typed.totalTokens,
-					totalCost: 0,
+					totalCost: cost,
 					tags: {},
 					metadata: {}
 				};

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@llmops/app",
-  "version": "0.6.2",
+  "version": "0.6.3",
   "description": "LLMOps application with server and client",
   "type": "module",
   "license": "Apache-2.0",
@@ -76,8 +76,8 @@
     "react-hook-form": "^7.68.0",
     "recharts": "^3.6.0",
     "uuid": "^13.0.0",
-    "@llmops/core": "^0.6.2",
-    "@llmops/gateway": "^0.6.2"
+    "@llmops/core": "^0.6.3",
+    "@llmops/gateway": "^0.6.3"
   },
   "peerDependencies": {
     "react": "^19.2.1",