nvidia-nat-llama-index 1.4.0a20251015__py3-none-any.whl → 1.4.0a20251021__py3-none-any.whl

nat/plugins/llama_index/llm.py

@@ -19,6 +19,7 @@ from typing import TypeVar
 from nat.builder.builder import Builder
 from nat.builder.framework_enum import LLMFrameworkEnum
 from nat.cli.register_workflow import register_llm_client
+from nat.data_models.llm import APITypeEnum
 from nat.data_models.llm import LLMBaseConfig
 from nat.data_models.retry_mixin import RetryMixin
 from nat.data_models.thinking_mixin import ThinkingMixin
@@ -31,6 +32,7 @@ from nat.llm.utils.thinking import BaseThinkingInjector
 from nat.llm.utils.thinking import FunctionArgumentWrapper
 from nat.llm.utils.thinking import patch_with_thinking
 from nat.utils.exception_handlers.automatic_retries import patch_with_retry
+from nat.utils.responses_api import validate_no_responses_api
 from nat.utils.type_utils import override
 
 ModelType = TypeVar("ModelType")
@@ -44,8 +46,16 @@ def _patch_llm_based_on_config(client: ModelType, llm_config: LLMBaseConfig) ->
 
         @override
         def inject(self, messages: Sequence[ChatMessage], *args, **kwargs) -> FunctionArgumentWrapper:
-            new_messages = [ChatMessage(role="system", content=self.system_prompt)] + list(messages)
-            return FunctionArgumentWrapper(new_messages, *args, **kwargs)
+            for i, message in enumerate(messages):
+                if message.role == "system":
+                    if self.system_prompt not in str(message.content):
+                        messages = list(messages)
+                        messages[i] = ChatMessage(role="system", content=f"{message.content}\n{self.system_prompt}")
+                    break
+            else:
+                messages = list(messages)
+                messages.insert(0, ChatMessage(role="system", content=self.system_prompt))
+            return FunctionArgumentWrapper(messages, *args, **kwargs)
 
     if isinstance(llm_config, RetryMixin):
         client = patch_with_retry(client,
@@ -74,8 +84,10 @@ async def aws_bedrock_llama_index(llm_config: AWSBedrockModelConfig, _builder: B
 
     from llama_index.llms.bedrock import Bedrock
 
+    validate_no_responses_api(llm_config, LLMFrameworkEnum.LLAMA_INDEX)
+
     # LlamaIndex uses context_size instead of max_tokens
-    llm = Bedrock(**llm_config.model_dump(exclude={"type", "top_p", "thinking"}, by_alias=True))
+    llm = Bedrock(**llm_config.model_dump(exclude={"type", "top_p", "thinking", "api_type"}, by_alias=True))
 
     yield _patch_llm_based_on_config(llm, llm_config)
 
@@ -85,7 +97,9 @@ async def azure_openai_llama_index(llm_config: AzureOpenAIModelConfig, _builder:
 
     from llama_index.llms.azure_openai import AzureOpenAI
 
-    llm = AzureOpenAI(**llm_config.model_dump(exclude={"type", "thinking"}, by_alias=True))
+    validate_no_responses_api(llm_config, LLMFrameworkEnum.LLAMA_INDEX)
+
+    llm = AzureOpenAI(**llm_config.model_dump(exclude={"type", "thinking", "api_type"}, by_alias=True))
 
     yield _patch_llm_based_on_config(llm, llm_config)
 
@@ -95,7 +109,9 @@ async def nim_llama_index(llm_config: NIMModelConfig, _builder: Builder):
 
     from llama_index.llms.nvidia import NVIDIA
 
-    llm = NVIDIA(**llm_config.model_dump(exclude={"type", "thinking"}, by_alias=True, exclude_none=True))
+    validate_no_responses_api(llm_config, LLMFrameworkEnum.LLAMA_INDEX)
+
+    llm = NVIDIA(**llm_config.model_dump(exclude={"type", "thinking", "api_type"}, by_alias=True, exclude_none=True))
 
     yield _patch_llm_based_on_config(llm, llm_config)
 
@@ -104,8 +120,14 @@ async def nim_llama_index(llm_config: NIMModelConfig, _builder: Builder):
 async def openai_llama_index(llm_config: OpenAIModelConfig, _builder: Builder):
 
     from llama_index.llms.openai import OpenAI
+    from llama_index.llms.openai import OpenAIResponses
 
-    llm = OpenAI(**llm_config.model_dump(exclude={"type", "thinking"}, by_alias=True, exclude_none=True))
+    if llm_config.api_type == APITypeEnum.RESPONSES:
+        llm = OpenAIResponses(
+            **llm_config.model_dump(exclude={"type", "thinking", "api_type"}, by_alias=True, exclude_none=True))
+    else:
+        llm = OpenAI(
+            **llm_config.model_dump(exclude={"type", "thinking", "api_type"}, by_alias=True, exclude_none=True))
 
     yield _patch_llm_based_on_config(llm, llm_config)
 
@@ -115,6 +137,8 @@ async def litellm_llama_index(llm_config: LiteLlmModelConfig, _builder: Builder)
 
     from llama_index.llms.litellm import LiteLLM
 
-    llm = LiteLLM(**llm_config.model_dump(exclude={"type", "thinking"}, by_alias=True, exclude_none=True))
+    validate_no_responses_api(llm_config, LLMFrameworkEnum.LLAMA_INDEX)
+
+    llm = LiteLLM(**llm_config.model_dump(exclude={"type", "thinking", "api_type"}, by_alias=True, exclude_none=True))
 
     yield _patch_llm_based_on_config(llm, llm_config)

{nvidia_nat_llama_index-1.4.0a20251015.dist-info → nvidia_nat_llama_index-1.4.0a20251021.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nvidia-nat-llama-index
-Version: 1.4.0a20251015
+Version: 1.4.0a20251021
 Summary: Subpackage for Llama-Index integration in NeMo Agent toolkit
 Author: NVIDIA Corporation
 Maintainer: NVIDIA Corporation
@@ -16,18 +16,18 @@ Requires-Python: <3.14,>=3.11
 Description-Content-Type: text/markdown
 License-File: LICENSE-3rd-party.txt
 License-File: LICENSE.md
-Requires-Dist: nvidia-nat==v1.4.0a20251015
-Requires-Dist: llama-index-core~=0.12.21
+Requires-Dist: nvidia-nat==v1.4.0a20251021
+Requires-Dist: llama-index-core~=0.12.40
 Requires-Dist: llama-index-embeddings-azure-openai~=0.3.9
 Requires-Dist: llama-index-embeddings-nvidia~=0.3.1
 Requires-Dist: llama-index-embeddings-openai~=0.3.1
 Requires-Dist: llama-index-llms-azure-openai~=0.3.2
 Requires-Dist: llama-index-llms-bedrock~=0.3.8
 Requires-Dist: llama-index-llms-litellm~=0.5.1
-Requires-Dist: llama-index-llms-nvidia~=0.3.1
-Requires-Dist: llama-index-llms-openai~=0.3.42
+Requires-Dist: llama-index-llms-nvidia~=0.3.4
+Requires-Dist: llama-index-llms-openai<1.0.0,>=0.4.2
 Requires-Dist: llama-index-readers-file~=0.4.4
-Requires-Dist: llama-index~=0.12.21
+Requires-Dist: llama-index~=0.12.40
 Dynamic: license-file
 
 <!--

{nvidia_nat_llama_index-1.4.0a20251015.dist-info → nvidia_nat_llama_index-1.4.0a20251021.dist-info}/RECORD

@@ -1,13 +1,13 @@
 nat/meta/pypi.md,sha256=s9C3pgWB0HLIXTx5QPryNOWN0O2fIRIap0p9_zCHlTs,1112
 nat/plugins/llama_index/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nat/plugins/llama_index/embedder.py,sha256=4c4uAa19UArqA_Npe5mwEnUhCaguTbyWcAntBzcPEeQ,3493
-nat/plugins/llama_index/llm.py,sha256=ms6qwcBISWhiiOhIUXFNuhpDenTIS5s7U47wB3fEgg0,4944
+nat/plugins/llama_index/llm.py,sha256=Vd331oSjTH4a9IAh3mgULXd3Ao9pBJ54WPQsXIz8drI,6088
 nat/plugins/llama_index/register.py,sha256=1x_b8u6cuQwh4Iz_7TcIFWXvLIL9IIKUPE-zR9d6ug8,859
 nat/plugins/llama_index/tool_wrapper.py,sha256=VFKMIIeLdWqHwW2Ax11E2w-_9w3ow6Iuhra1Hk78RYM,1387
-nvidia_nat_llama_index-1.4.0a20251015.dist-info/licenses/LICENSE-3rd-party.txt,sha256=fOk5jMmCX9YoKWyYzTtfgl-SUy477audFC5hNY4oP7Q,284609
-nvidia_nat_llama_index-1.4.0a20251015.dist-info/licenses/LICENSE.md,sha256=QwcOLU5TJoTeUhuIXzhdCEEDDvorGiC6-3YTOl4TecE,11356
-nvidia_nat_llama_index-1.4.0a20251015.dist-info/METADATA,sha256=yahkKb4XlatyqkaBJQSTEr_hmEoKBMBGrxvKqCJlK98,2422
-nvidia_nat_llama_index-1.4.0a20251015.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-nvidia_nat_llama_index-1.4.0a20251015.dist-info/entry_points.txt,sha256=2LqRRju5448P2v8B3y6TSPnk-nOd5T3AmV5JibCnoQc,68
-nvidia_nat_llama_index-1.4.0a20251015.dist-info/top_level.txt,sha256=8-CJ2cP6-f0ZReXe5Hzqp-5pvzzHz-5Ds5H2bGqh1-U,4
-nvidia_nat_llama_index-1.4.0a20251015.dist-info/RECORD,,
+nvidia_nat_llama_index-1.4.0a20251021.dist-info/licenses/LICENSE-3rd-party.txt,sha256=fOk5jMmCX9YoKWyYzTtfgl-SUy477audFC5hNY4oP7Q,284609
+nvidia_nat_llama_index-1.4.0a20251021.dist-info/licenses/LICENSE.md,sha256=QwcOLU5TJoTeUhuIXzhdCEEDDvorGiC6-3YTOl4TecE,11356
+nvidia_nat_llama_index-1.4.0a20251021.dist-info/METADATA,sha256=M8IG_Is0JJHy65XlNfqBNmq1tAY2ds8JPLLoeGiLtMc,2428
+nvidia_nat_llama_index-1.4.0a20251021.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nvidia_nat_llama_index-1.4.0a20251021.dist-info/entry_points.txt,sha256=2LqRRju5448P2v8B3y6TSPnk-nOd5T3AmV5JibCnoQc,68
+nvidia_nat_llama_index-1.4.0a20251021.dist-info/top_level.txt,sha256=8-CJ2cP6-f0ZReXe5Hzqp-5pvzzHz-5Ds5H2bGqh1-U,4
+nvidia_nat_llama_index-1.4.0a20251021.dist-info/RECORD,,