From d46df34ff5fe72ffb88fdd5599e6b5b17d13ff13 Mon Sep 17 00:00:00 2001
From: Krrish Dholakia <krrishdholakia@gmail.com>
Date: Thu, 1 Feb 2024 09:53:33 -0800
Subject: [PATCH] fix(utils.py): fix streaming chunks to not return role,
 unless set

---
 litellm/llms/ollama.py | 10 +++++-----
 litellm/utils.py       |  3 ++-
 2 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/litellm/llms/ollama.py b/litellm/llms/ollama.py
index 0f99622b32..c8847117e3 100644
--- a/litellm/llms/ollama.py
+++ b/litellm/llms/ollama.py
@@ -145,8 +145,8 @@ def get_ollama_response(
         ):  # completion(top_k=3) > cohere_config(top_k=3) <- allows for dynamic variables to be passed in
             optional_params[k] = v
 
-    optional_params["stream"] = optional_params.get("stream", False)
-    data = {"model": model, "prompt": prompt, **optional_params}
+    stream = optional_params.pop("stream", False)
+    data = {"model": model, "prompt": prompt, "options": optional_params}
     ## LOGGING
     logging_obj.pre_call(
         input=None,
@@ -159,7 +159,7 @@ def get_ollama_response(
         },
     )
     if acompletion is True:
-        if optional_params.get("stream", False) == True:
+        if stream == True:
             response = ollama_async_streaming(
                 url=url,
                 data=data,
@@ -176,7 +176,7 @@ def get_ollama_response(
                 logging_obj=logging_obj,
             )
         return response
-    elif optional_params.get("stream", False) == True:
+    elif stream == True:
         return ollama_completion_stream(url=url, data=data, logging_obj=logging_obj)
 
     response = requests.post(url=f"{url}", json=data, timeout=litellm.request_timeout)
@@ -254,7 +254,7 @@ async def ollama_async_streaming(url, data, model_response, encoding, logging_ob
         ) as response:
             if response.status_code != 200:
                 raise OllamaError(
-                    status_code=response.status_code, message=response.text
+                    status_code=response.status_code, message=await response.aread()
                 )
 
             streamwrapper = litellm.CustomStreamWrapper(
diff --git a/litellm/utils.py b/litellm/utils.py
index a954353c60..d3107d758a 100644
--- a/litellm/utils.py
+++ b/litellm/utils.py
@@ -233,7 +233,8 @@ class Delta(OpenAIObject):
     def __init__(self, content=None, role=None, **params):
         super(Delta, self).__init__(**params)
         self.content = content
-        self.role = role
+        if role is not None:
+            self.role = role
 
     def __contains__(self, key):
         # Define custom behavior for the 'in' operator