@tiens.nguyen/gonext-local-worker 1.0.38 → 1.0.39

package/gonext-local-worker.mjs

@@ -278,6 +278,25 @@ function toOpenAIMessages(messages) {
   });
 }
 
+function parseCompletionTokens(usage) {
+  if (!usage || typeof usage !== "object") {
+    return null;
+  }
+  if (typeof usage.completion_tokens === "number") {
+    return usage.completion_tokens;
+  }
+  if (typeof usage.output_tokens === "number") {
+    return usage.output_tokens;
+  }
+  return null;
+}
+
+function shouldRetryWithoutUsage(err) {
+  const msg =
+    err instanceof Error ? err.message.toLowerCase() : String(err).toLowerCase();
+  return msg.includes("stream_options") || msg.includes("include_usage");
+}
+
 async function runChatJob(job) {
   const { jobId, payload } = job;
   if (!payload || !Array.isArray(payload.messages)) {
@@ -370,18 +389,34 @@ async function runChatJob(job) {
   };
 
   try {
-    const stream = await client.chat.completions.create({
+    const streamRequest = {
       model: payload.modelId,
       messages: toOpenAIMessages(payload.messages),
       stream: true,
       temperature: 0,
-    });
+    };
+    const stream = await client.chat.completions
+      .create({
+        ...streamRequest,
+        stream_options: { include_usage: true },
+      })
+      .catch(async (e) => {
+        if (!shouldRetryWithoutUsage(e)) {
+          throw e;
+        }
+        return client.chat.completions.create(streamRequest);
+      });
 
     let tokenCount = 0;
+    let completionTokensFromUsage = null;
     let isStartThinking = false;
     let isEndThinking = false;
 
     for await (const chunk of stream) {
+      const usageTokens = parseCompletionTokens(chunk.usage);
+      if (usageTokens !== null) {
+        completionTokensFromUsage = usageTokens;
+      }
       const delta = chunk.choices[0]?.delta;
       const content = delta?.content ?? "";
       const reasoningContent = delta?.reasoning_content;
@@ -419,7 +454,10 @@ async function runChatJob(job) {
       body: JSON.stringify({
         jobStatus: "completed",
         resultText: fullText,
-        tokenCount: Math.max(1, tokenCount),
+        tokenCount:
+          completionTokensFromUsage !== null
+            ? completionTokensFromUsage
+            : Math.max(1, tokenCount),
         totalTimeSeconds,
       }),
     });

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tiens.nguyen/gonext-local-worker",
-  "version": "1.0.38",
+  "version": "1.0.39",
   "description": "Polls GoNext cloud API for async local LLM jobs and runs them against Ollama/OpenAI-compatible servers on this Mac",
   "type": "module",
   "license": "MIT",