Refactor _llm.py to handle UnicodeDecodeError when decoding chunks and add timeout to LLM API request

_llm.py

@@ -37,7 +37,10 @@ class LLM:
         buffer = ""
         for chunk in response.iter_content(chunk_size=64):
             if chunk:
+                try:
                     buffer += chunk.decode('utf-8')
+                except UnicodeDecodeError:
+                    continue
                 while "\n" in buffer:
                     line, buffer = buffer.split("\n", 1)
                     if line:
@@ -74,6 +77,7 @@ class LLM:
             json=data,
             auth=HTTPBasicAuth(os.getenv('LLM_API_USER'), os.getenv('LLM_API_PWD_LASSE')),
             stream=stream,
+            timeout= 3600,
         )
 
         if response.status_code == 404: