From f263f9e48a38d6b6f0132d8c6ea2a38a6d19d1b6 Mon Sep 17 00:00:00 2001
From: MohamedBassem <me@mbassem.com>
Date: Sun, 15 Sep 2024 00:03:39 +0000
Subject: feature(worker): Allow configuring inference job timeout and ollama
 keep alive. Fixes #389 #224

---
 apps/workers/inference.ts    | 1 +
 apps/workers/openaiWorker.ts | 2 +-
 2 files changed, 2 insertions(+), 1 deletion(-)

(limited to 'apps/workers')

diff --git a/apps/workers/inference.ts b/apps/workers/inference.ts
index fa83140f..071f4742 100644
--- a/apps/workers/inference.ts
+++ b/apps/workers/inference.ts
@@ -103,6 +103,7 @@ class OllamaInferenceClient implements InferenceClient {
       model: model,
       format: "json",
       stream: true,
+      keep_alive: serverConfig.inference.ollamaKeepAlive,
       messages: [
         { role: "user", content: prompt, images: image ? [image] : undefined },
       ],
diff --git a/apps/workers/openaiWorker.ts b/apps/workers/openaiWorker.ts
index 55695938..8bd2cf4a 100644
--- a/apps/workers/openaiWorker.ts
+++ b/apps/workers/openaiWorker.ts
@@ -81,7 +81,7 @@ export class OpenAiWorker {
       {
         concurrency: 1,
         pollIntervalMs: 1000,
-        timeoutSecs: 30,
+        timeoutSecs: serverConfig.inference.jobTimeoutSec,
       },
     );
 
-- 
cgit v1.2.3-70-g09d2