Merge branch 'main' into litellm_bedrock_anthropic_fix

docs/my-website/docs/observability/langsmith_integration.md

@@ -71,6 +71,23 @@ response = litellm.completion(
 )
 print(response)
 ```
+
+### Make LiteLLM Proxy use Custom `LANGSMITH_BASE_URL`
+
+If you're using a custom LangSmith instance, you can set the
+`LANGSMITH_BASE_URL` environment variable to point to your instance.
+For example, you can make LiteLLM Proxy log to a local LangSmith instance with
+this config:
+
+```yaml
+litellm_settings:
+  success_callback: ["langsmith"]
+
+environment_variables:
+  LANGSMITH_BASE_URL: "http://localhost:1984"
+  LANGSMITH_PROJECT: "litellm-proxy"
+```
+
 ## Support & Talk to Founders
 
 - [Schedule Demo 👋](https://calendly.com/d/4mp-gd3-k5k/berriai-1-1-onboarding-litellm-hosted-version)

docs/my-website/docs/providers/anthropic.md

@@ -232,7 +232,6 @@ response = completion(
     model="anthropic/claude-3-opus-20240229",
     messages=messages,
     tools=tools,
-    extra_headers={"anthropic-beta": "tools-2024-05-16"},
 )
 ```
 
@@ -247,7 +246,6 @@ response = completion(
     messages=messages,
     tools=tools,
     tool_choice={"type": "tool", "name": "get_weather"},
-    extra_headers={"anthropic-beta": "tools-2024-05-16"},
 )
 ```
 

docs/my-website/docs/proxy/enterprise.md

@@ -1,7 +1,8 @@
+import Image from '@theme/IdealImage';
 import Tabs from '@theme/Tabs';
 import TabItem from '@theme/TabItem';
 
-# ✨ Enterprise Features - Content Mod, SSO
+# ✨ Enterprise Features - Content Mod, SSO, Custom Swagger
 
 Features here are behind a commercial license in our `/enterprise` folder. [**See Code**](https://github.com/BerriAI/litellm/tree/main/enterprise)
 
@@ -20,6 +21,7 @@ Features:
 - ✅ Reject calls (incoming / outgoing) with Banned Keywords (e.g. competitors)
 - ✅ Don't log/store specific requests to Langfuse, Sentry, etc. (eg confidential LLM requests)
 - ✅ Tracking Spend for Custom Tags
+- ✅ Custom Branding + Routes on Swagger Docs
 
 
 
@@ -526,4 +528,39 @@ curl -X GET "http://0.0.0.0:4000/spend/tags" \
 
 <!-- ## Tracking Spend per Key
 
-## Tracking Spend per User -->
+## Tracking Spend per User -->
+
+## Swagger Docs - Custom Routes + Branding 
+
+:::info 
+
+Requires a LiteLLM Enterprise key to use. Request one [here](https://calendly.com/d/4mp-gd3-k5k/litellm-1-1-onboarding-chat)
+
+:::
+
+Set LiteLLM Key in your environment
+
+```bash
+LITELLM_LICENSE=""
+```
+
+### Customize Title + Description
+
+In your environment, set: 
+
+```bash
+DOCS_TITLE="TotalGPT"
+DOCS_DESCRIPTION="Sample Company Description"
+```
+
+### Customize Routes
+
+Hide admin routes from users. 
+
+In your environment, set: 
+
+```bash
+DOCS_FILTERED="True" # only shows openai routes to user
+```
+
+<Image img={require('../../img/custom_swagger.png')}  style={{ width: '900px', height: 'auto' }} />

litellm/integrations/langsmith.py

@@ -44,6 +44,8 @@ class LangsmithLogger:
         print_verbose(
             f"Langsmith Logging - project_name: {project_name}, run_name {run_name}"
         )
+        langsmith_base_url = os.getenv("LANGSMITH_BASE_URL", "https://api.smith.langchain.com")
+
         try:
             print_verbose(
                 f"Langsmith Logging - Enters logging function for model {kwargs}"
@@ -86,8 +88,12 @@ class LangsmithLogger:
                 "end_time": end_time,
             }
 
+            url = f"{langsmith_base_url}/runs"
+            print_verbose(
+                f"Langsmith Logging - About to send data to {url} ..."
+            )
             response = requests.post(
-                "https://api.smith.langchain.com/runs",
+                url=url,
                 json=data,
                 headers={"x-api-key": self.langsmith_api_key},
             )

litellm/llms/anthropic.py

@@ -507,7 +507,7 @@ class AnthropicChatCompletion(BaseLLM):
             _is_function_call = True
             if "anthropic-beta" not in headers:
                 # default to v1 of "anthropic-beta"
-                headers["anthropic-beta"] = "tools-2024-04-04"
+                headers["anthropic-beta"] = "tools-2024-05-16"
 
             anthropic_tools = []
             for tool in optional_params["tools"]:

litellm/proxy/_types.py

@@ -87,10 +87,10 @@ class LiteLLMRoutes(enum.Enum):
         # models
         "/models",
         "/v1/models",
+        # token counter
+        "/utils/token_counter",
     ]
 
-    llm_utils_routes: List = ["utils/token_counter"]
-
     info_routes: List = [
         "/key/info",
         "/team/info",

litellm/proxy/auth/litellm_license.py

@@ -0,0 +1,42 @@
+# What is this?
+## If litellm license in env, checks if it's valid
+import os
+from litellm.llms.custom_httpx.http_handler import HTTPHandler
+
+
+class LicenseCheck:
+    """
+    - Check if license in env
+    - Returns if license is valid
+    """
+
+    base_url = "https://license.litellm.ai"
+
+    def __init__(self) -> None:
+        self.license_str = os.getenv("LITELLM_LICENSE", None)
+        self.http_handler = HTTPHandler()
+
+    def _verify(self, license_str: str) -> bool:
+        url = "{}/verify_license/{}".format(self.base_url, license_str)
+
+        try:  # don't impact user, if call fails
+            response = self.http_handler.get(url=url)
+
+            response.raise_for_status()
+
+            response_json = response.json()
+
+            premium = response_json["verify"]
+
+            assert isinstance(premium, bool)
+
+            return premium
+        except Exception as e:
+            return False
+
+    def is_premium(self) -> bool:
+        if self.license_str is None:
+            return False
+        elif self._verify(license_str=self.license_str):
+            return True
+        return False

litellm/proxy/proxy_server.py

@@ -110,6 +110,7 @@ from litellm.router import LiteLLM_Params, Deployment, updateDeployment
 from litellm.router import ModelInfo as RouterModelInfo
 from litellm._logging import verbose_router_logger, verbose_proxy_logger
 from litellm.proxy.auth.handle_jwt import JWTHandler
+from litellm.proxy.auth.litellm_license import LicenseCheck
 from litellm.proxy.hooks.prompt_injection_detection import (
     _OPTIONAL_PromptInjectionDetection,
 )
@@ -150,6 +151,7 @@ from fastapi.responses import (
     ORJSONResponse,
     JSONResponse,
 )
+from fastapi.openapi.utils import get_openapi
 from fastapi.responses import RedirectResponse
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.staticfiles import StaticFiles
@@ -169,17 +171,30 @@ except Exception as e:
     except Exception as e:
         pass
 
+_license_check = LicenseCheck()
+premium_user: bool = _license_check.is_premium()
+
 ui_link = f"/ui/"
 ui_message = (
     f"👉 [```LiteLLM Admin Panel on /ui```]({ui_link}). Create, Edit Keys with SSO"
 )
 
+### CUSTOM BRANDING [ENTERPRISE FEATURE] ###
 _docs_url = None if os.getenv("NO_DOCS", "False") == "True" else "/"
+_title = os.getenv("DOCS_TITLE", "LiteLLM API") if premium_user else "LiteLLM API"
+_description = (
+    os.getenv(
+        "DOCS_DESCRIPTION",
+        f"Proxy Server to call 100+ LLMs in the OpenAI format\n\n{ui_message}",
+    )
+    if premium_user
+    else f"Proxy Server to call 100+ LLMs in the OpenAI format\n\n{ui_message}"
+)
 
 app = FastAPI(
     docs_url=_docs_url,
-    title="LiteLLM API",
-    description=f"Proxy Server to call 100+ LLMs in the OpenAI format\n\n{ui_message}",
+    title=_title,
+    description=_description,
     version=version,
     root_path=os.environ.get(
         "SERVER_ROOT_PATH", ""
@@ -187,6 +202,31 @@ app = FastAPI(
 )
 
 
+### CUSTOM API DOCS [ENTERPRISE FEATURE] ###
+# Custom OpenAPI schema generator to include only selected routes
+def custom_openapi():
+    if app.openapi_schema:
+        return app.openapi_schema
+    openapi_schema = get_openapi(
+        title=app.title,
+        version=app.version,
+        description=app.description,
+        routes=app.routes,
+    )
+    # Filter routes to include only specific ones
+    openai_routes = LiteLLMRoutes.openai_routes.value
+    paths_to_include: dict = {}
+    for route in openai_routes:
+        paths_to_include[route] = openapi_schema["paths"][route]
+    openapi_schema["paths"] = paths_to_include
+    app.openapi_schema = openapi_schema
+    return app.openapi_schema
+
+
+if os.getenv("DOCS_FILTERED", "False") == "True" and premium_user:
+    app.openapi = custom_openapi  # type: ignore
+
+
 class ProxyException(Exception):
     # NOTE: DO NOT MODIFY THIS
     # This is used to map exactly to OPENAI Exceptions

litellm/tests/test_completion.py

@@ -3032,7 +3032,6 @@ def test_mistral_anyscale_stream():
         print(chunk["choices"][0]["delta"].get("content", ""), end="")
 
 
-# test_mistral_anyscale_stream()
 # test_completion_anyscale_2()
 # def test_completion_with_fallbacks_multiple_keys():
 #     print(f"backup key 1: {os.getenv('BACKUP_OPENAI_API_KEY_1')}")

litellm/utils.py

@@ -5122,7 +5122,7 @@ def get_optional_params(
         or "tools" in non_default_params
     ):
         if (
-            custom_llm_provider != "openai"
+            custom_llm_provider == "ollama"
             and custom_llm_provider != "text-completion-openai"
             and custom_llm_provider != "azure"
             and custom_llm_provider != "vertex_ai"
@@ -5136,6 +5136,8 @@ def get_optional_params(
             and custom_llm_provider != "cohere"
             and custom_llm_provider != "bedrock"
             and custom_llm_provider != "ollama_chat"
+            and custom_llm_provider != "openrouter"
+            and custom_llm_provider not in litellm.openai_compatible_providers
         ):
             if custom_llm_provider == "ollama":
                 # ollama actually supports json output

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "litellm"
-version = "1.37.13"
+version = "1.37.15"
 description = "Library to easily interface with LLM API providers"
 authors = ["BerriAI"]
 license = "MIT"
@@ -79,7 +79,7 @@ requires = ["poetry-core", "wheel"]
 build-backend = "poetry.core.masonry.api"
 
 [tool.commitizen]
-version = "1.37.13"
+version = "1.37.15"
 version_files = [
     "pyproject.toml:^version"
 ]