Address PR review feedback for LiteLLM provider

- Add "strict": True to json_schema response_format in _invoke_v2 and
  _ainvoke_v2, matching the OpenAI provider pattern
- Set supports_structured_output = False since not all LiteLLM-supported
  providers handle structured output
- Add @rate_limit_handler_decorator to invoke_with_tools and
  @async_rate_limit_handler_decorator to ainvoke_with_tools for
  consistent rate limit handling across all methods
- Add LiteLLM provider documentation to user guide and API reference
- Remove "Generated by Claude Code" comments from test files

Author: Aarish Irani

docs/source/api.rst

@@ -318,6 +318,12 @@ CohereEmbeddings
 .. autoclass:: neo4j_graphrag.embeddings.cohere.CohereEmbeddings
     :members:
 
+LiteLLMEmbeddings
+=================
+
+.. autoclass:: neo4j_graphrag.embeddings.litellm.LiteLLMEmbeddings
+    :members:
+
 **********
 Generation
 **********
@@ -402,6 +408,13 @@ MistralAILLM
     :members:
 
 
+LiteLLMChat
+-----------
+
+.. autoclass:: neo4j_graphrag.llm.litellm_llm.LiteLLMChat
+    :members:
+
+
 Rate Limiting
 =============
 

docs/source/user_guide_rag.rst

@@ -79,6 +79,7 @@ If OpenAI cannot be used directly, there are a few available alternatives:
 - Use Anthropic LLM (Claude...).
 - Use Mistral LLM
 - Use Cohere.
+- Use LiteLLM (100+ providers via a unified interface).
 - Use a local Ollama model.
 - Implement a custom interface.
 - Utilize any LangChain chat model.
@@ -214,6 +215,47 @@ To use Cohere, instantiate the `CohereLLM` class:
 See :ref:`coherellm`.
 
 
+Using LiteLLM (100+ Providers)
+-------------------------------
+
+LiteLLM provides a unified interface to 100+ LLM providers (OpenAI, Anthropic,
+Google, Azure, AWS Bedrock, Ollama, and more) using the OpenAI completion format.
+To use LiteLLM, instantiate the `LiteLLMChat` class:
+
+.. code:: python
+
+    from neo4j_graphrag.llm import LiteLLMChat
+
+    # Use any LiteLLM-supported provider
+    llm = LiteLLMChat(model_name="gpt-4o", api_key="...")
+    llm.invoke("say something")
+
+    # Anthropic
+    llm = LiteLLMChat(model_name="anthropic/claude-sonnet-4-20250514", api_key="...")
+
+    # Ollama (local)
+    llm = LiteLLMChat(model_name="ollama/llama3")
+
+    # Azure OpenAI
+    llm = LiteLLMChat(
+        model_name="azure/my-deployment",
+        api_key="...",
+        api_base="https://my-endpoint.openai.azure.com/",
+    )
+
+    # AWS Bedrock
+    llm = LiteLLMChat(model_name="bedrock/anthropic.claude-v2")
+
+
+.. note::
+
+    In order to run this code, the `litellm` Python package needs to be installed:
+    `pip install "neo4j_graphrag[litellm]"`
+
+
+See :ref:`litellmchat`.
+
+
 Using a Local Model via Ollama
 -------------------------------
 
@@ -620,6 +662,7 @@ Currently, this package supports the following embedders:
 - :ref:`cohereembeddings`
 - :ref:`azureopenaiembeddings`
 - :ref:`ollamaembeddings`
+- :ref:`litellmembeddings`
 
 The `OpenAIEmbeddings` was illustrated previously. Here is how to use the `SentenceTransformerEmbeddings`:
 

src/neo4j_graphrag/llm/litellm_llm.py

@@ -88,7 +88,7 @@ class LiteLLMChat(LLMBase):
         llm = LiteLLMChat(model_name="ollama/llama3")
     """
 
-    supports_structured_output = True
+    supports_structured_output = False
 
     def __init__(
         self,
@@ -145,6 +145,7 @@ async def ainvoke(
         else:
             raise ValueError(f"Invalid input type for ainvoke method - {type(input)}")
 
+    @rate_limit_handler_decorator
     def invoke_with_tools(
         self,
         input: str,
@@ -193,6 +194,7 @@ def invoke_with_tools(
             content=res.choices[0].message.content,
         )
 
+    @async_rate_limit_handler_decorator
     async def ainvoke_with_tools(
         self,
         input: str,
@@ -284,6 +286,7 @@ def _invoke_v2(
                         "type": "json_schema",
                         "json_schema": {
                             "name": response_format.__name__,
+                            "strict": True,
                             "schema": response_format.model_json_schema(),
                         },
                     }
@@ -340,6 +343,7 @@ async def _ainvoke_v2(
                         "type": "json_schema",
                         "json_schema": {
                             "name": response_format.__name__,
+                            "strict": True,
                             "schema": response_format.model_json_schema(),
                         },
                     }

tests/unit/embeddings/test_litellm_embedder.py

@@ -13,8 +13,6 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
-# Generated by Claude Code
-
 import sys
 from typing import Generator
 from unittest.mock import AsyncMock, MagicMock, Mock, patch

tests/unit/llm/test_litellm_llm.py

@@ -13,8 +13,6 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
-# Generated by Claude Code
-
 import sys
 import warnings
 from typing import Generator, List