fix - use anthropic class for clients

litellm/llms/anthropic.py

@@ -8,7 +8,7 @@ from litellm.utils import ModelResponse, Usage, map_finish_reason, CustomStreamW
 import litellm
 from .prompt_templates.factory import prompt_factory, custom_prompt
 from litellm.llms.custom_httpx.http_handler import AsyncHTTPHandler
-
+from .base import BaseLLM
 import httpx
 
 
@@ -19,9 +19,6 @@ class AnthropicConstants(Enum):
     # constants from https://github.com/anthropics/anthropic-sdk-python/blob/main/src/anthropic/_constants.py
 
 
-async_handler = AsyncHTTPHandler(timeout=httpx.Timeout(timeout=600.0, connect=5.0))
-
-
 class AnthropicError(Exception):
     def __init__(self, status_code, message):
         self.status_code = status_code
@@ -105,7 +102,15 @@ def validate_environment(api_key, user_headers):
     return headers
 
 
-def process_response(
+class AnthropicChatCompletion(BaseLLM):
+    def __init__(self) -> None:
+        super().__init__()
+        self.async_handler = AsyncHTTPHandler(
+            timeout=httpx.Timeout(timeout=600.0, connect=5.0)
+        )
+
+    def process_response(
+        self,
         model,
         response,
         model_response,
@@ -116,7 +121,7 @@ def process_response(
         data,
         messages,
         print_verbose,
-):
+    ):
         ## LOGGING
         logging_obj.post_call(
             input=messages,
@@ -129,7 +134,9 @@ def process_response(
         try:
             completion_response = response.json()
         except:
-        raise AnthropicError(message=response.text, status_code=response.status_code)
+            raise AnthropicError(
+                message=response.text, status_code=response.status_code
+            )
         if "error" in completion_response:
             raise AnthropicError(
                 message=str(completion_response["error"]),
@@ -232,8 +239,8 @@ def process_response(
         model_response.usage = usage
         return model_response
 
-
-async def acompletion_stream_function(
+    async def acompletion_stream_function(
+        self,
         model: str,
         messages: list,
         api_base: str,
@@ -250,13 +257,15 @@ async def acompletion_stream_function(
         litellm_params=None,
         logger_fn=None,
         headers={},
-):
-    response = await async_handler.post(
+    ):
+        response = await self.async_handler.post(
             api_base, headers=headers, data=json.dumps(data)
         )
 
         if response.status_code != 200:
-        raise AnthropicError(status_code=response.status_code, message=response.text)
+            raise AnthropicError(
+                status_code=response.status_code, message=response.text
+            )
 
         completion_stream = response.aiter_lines()
 
@@ -268,8 +277,8 @@ async def acompletion_stream_function(
         )
         return streamwrapper
 
-
-async def acompletion_function(
+    async def acompletion_function(
+        self,
         model: str,
         messages: list,
         api_base: str,
@@ -286,11 +295,11 @@ async def acompletion_function(
         litellm_params=None,
         logger_fn=None,
         headers={},
-):
-    response = await async_handler.post(
+    ):
+        response = await self.async_handler.post(
             api_base, headers=headers, data=json.dumps(data)
         )
-    return process_response(
+        return self.process_response(
             model=model,
             response=response,
             model_response=model_response,
@@ -303,8 +312,8 @@ async def acompletion_function(
             print_verbose=print_verbose,
         )
 
-
-def completion(
+    def completion(
+        self,
         model: str,
         messages: list,
         api_base: str,
@@ -319,7 +328,7 @@ def completion(
         litellm_params=None,
         logger_fn=None,
         headers={},
-):
+    ):
         headers = validate_environment(api_key, headers)
         _is_function_call = False
         messages = copy.deepcopy(messages)
@@ -400,7 +409,7 @@ def completion(
             ):  # if function call - fake the streaming (need complete blocks for output parsing in openai format)
                 print_verbose("makes async anthropic streaming POST request")
                 data["stream"] = stream
-            return acompletion_stream_function(
+                return self.acompletion_stream_function(
                     model=model,
                     messages=messages,
                     data=data,
@@ -419,7 +428,7 @@ def completion(
                     headers=headers,
                 )
             else:
-            return acompletion_function(
+                return self.acompletion_function(
                     model=model,
                     messages=messages,
                     data=data,
@@ -466,12 +475,14 @@ def completion(
                 return streaming_response
 
             else:
-            response = requests.post(api_base, headers=headers, data=json.dumps(data))
+                response = requests.post(
+                    api_base, headers=headers, data=json.dumps(data)
+                )
                 if response.status_code != 200:
                     raise AnthropicError(
                         status_code=response.status_code, message=response.text
                     )
-    return process_response(
+        return self.process_response(
             model=model,
             response=response,
             model_response=model_response,
@@ -484,6 +495,10 @@ def completion(
             print_verbose=print_verbose,
         )
 
+    def embedding(self):
+        # logic for parsing in - calling - parsing out model embedding calls
+        pass
+
 
 class ModelResponseIterator:
     def __init__(self, model_response):
@@ -509,8 +524,3 @@ class ModelResponseIterator:
             raise StopAsyncIteration
         self.is_done = True
         return self.model_response
-
-
-def embedding():
-    # logic for parsing in - calling - parsing out model embedding calls
-    pass

litellm/main.py

@@ -39,7 +39,6 @@ from litellm.utils import (
     get_optional_params_image_gen,
 )
 from .llms import (
-    anthropic,
     anthropic_text,
     together_ai,
     ai21,
@@ -68,6 +67,7 @@ from .llms import (
 from .llms.openai import OpenAIChatCompletion, OpenAITextCompletion
 from .llms.azure import AzureChatCompletion
 from .llms.azure_text import AzureTextCompletion
+from .llms.anthropic import AnthropicChatCompletion
 from .llms.huggingface_restapi import Huggingface
 from .llms.prompt_templates.factory import (
     prompt_factory,
@@ -99,6 +99,7 @@ from litellm.utils import (
 dotenv.load_dotenv()  # Loading env variables using dotenv
 openai_chat_completions = OpenAIChatCompletion()
 openai_text_completions = OpenAITextCompletion()
+anthropic_chat_completions = AnthropicChatCompletion()
 azure_chat_completions = AzureChatCompletion()
 azure_text_completions = AzureTextCompletion()
 huggingface = Huggingface()
@@ -1181,7 +1182,7 @@ def completion(
                     or get_secret("ANTHROPIC_API_BASE")
                     or "https://api.anthropic.com/v1/messages"
                 )
-                response = anthropic.completion(
+                response = anthropic_chat_completions.completion(
                     model=model,
                     messages=messages,
                     api_base=api_base,