Merge branch 'main' into ODSC-63451/oci_odsc_embedding

llama-index-integrations/llms/llama-index-llms-huggingface/llama_index/llms/huggingface/base.py

@@ -48,7 +48,7 @@
     to_tgi_messages,
     force_single_tool_call,
     resolve_tgi_function_call,
-    get_max_input_length,
+    get_max_total_tokens,
     resolve_tool_choice,
 )
 from transformers import (
@@ -137,7 +137,7 @@ def completion_to_prompt(completion):
     )
     context_window: int = Field(
         default=DEFAULT_CONTEXT_WINDOW,
-        description="The maximum number of tokens available for input.",
+        description=(LLMMetadata.model_fields["context_window"].description),
         gt=0,
     )
     max_new_tokens: int = Field(
@@ -752,7 +752,10 @@ class TextGenerationInference(FunctionCallingLLM):
 
     context_window: int = Field(
         default=DEFAULT_CONTEXT_WINDOW,
-        description=("Maximum input length in tokens returned from TGI endpoint"),
+        description=(
+            LLMMetadata.model_fields["context_window"].description
+            + " Maximum total tokens returned from TGI endpoint."
+        ),
     )
     is_chat_model: bool = Field(
         default=True,
@@ -819,7 +822,7 @@ def __init__(
             logger.warning(f"TGI client has no function call support: {e}")
             is_function_calling_model = False
 
-        context_window = get_max_input_length(model_url) or DEFAULT_CONTEXT_WINDOW
+        context_window = get_max_total_tokens(model_url) or DEFAULT_CONTEXT_WINDOW
 
         super().__init__(
             context_window=context_window,

llama-index-integrations/llms/llama-index-llms-huggingface/llama_index/llms/huggingface/utils.py

@@ -24,10 +24,20 @@ def resolve_tgi_function_call(url: str) -> bool:
         )
 
 
-def get_max_input_length(url: str) -> Union[int, None]:
+def get_max_input_tokens(url: str) -> Union[int, None]:
     url = f"{url}/info"
     model_info = dict(requests.get(url).json())
-    return model_info.get("max_input_length", None)
+    tgi_version = model_info.get("version", None)
+    if version.parse(tgi_version) >= version.parse("2.1.0"):
+        return model_info.get("max_input_tokens", None)
+    else:
+        return model_info.get("max_input_length", None)
+
+
+def get_max_total_tokens(url: str) -> Union[int, None]:
+    url = f"{url}/info"
+    model_info = dict(requests.get(url).json())
+    return model_info.get("max_total_tokens", None)
 
 
 def to_tgi_messages(messages: Sequence[ChatMessage]) -> Sequence[Message]:

llama-index-integrations/llms/llama-index-llms-huggingface/pyproject.toml

@@ -28,7 +28,7 @@ exclude = ["**/BUILD"]
 license = "MIT"
 name = "llama-index-llms-huggingface"
 readme = "README.md"
-version = "0.4.0"
+version = "0.4.1"
 
 [tool.poetry.dependencies]
 python = ">=3.9,<4.0"

llama-index-integrations/llms/llama-index-llms-nvidia/llama_index/llms/nvidia/utils.py

@@ -14,6 +14,7 @@
     "meta/codellama-70b": 1024,
     "meta/llama3-70b-instruct": 8192,
     "meta/llama3-8b-instruct": 8192,
+    "meta/llama-3.3-70b-instruct": 128000,
     "microsoft/phi-3-medium-4k-instruct": 1024,
     "microsoft/phi-3-mini-128k-instruct": 2048,
     "microsoft/phi-3-mini-4k-instruct": 2048,
@@ -37,6 +38,7 @@
     "meta/llama-3.1-8b-instruct",
     "meta/llama-3.1-70b-instruct",
     "meta/llama-3.1-405b-instruct",
+    "meta/llama-3.3-70b-instruct",
     "mistralai/mistral-large-2-instruct",
 )
 

llama-index-integrations/llms/llama-index-llms-nvidia/pyproject.toml

@@ -30,7 +30,7 @@ license = "MIT"
 name = "llama-index-llms-nvidia"
 packages = [{include = "llama_index/"}]
 readme = "README.md"
-version = "0.3.0"
+version = "0.3.1"
 
 [tool.poetry.dependencies]
 python = ">=3.9,<4.0"

llama-index-integrations/llms/llama-index-llms-openai/llama_index/llms/openai/utils.py

@@ -285,6 +285,10 @@ def to_openai_message_dict(
             msg = f"Unsupported content block type: {type(block).__name__}"
             raise ValueError(msg)
 
+    # NOTE: Sending a blank string to openai will cause an error.
+    # This will commonly happen with tool calls.
+    content_txt = None if content_txt == "" else content_txt
+
     # NOTE: Despite what the openai docs say, if the role is ASSISTANT, SYSTEM
     # or TOOL, 'content' cannot be a list and must be string instead.
     # Furthermore, if all blocks are text blocks, we can use the content_txt

llama-index-integrations/llms/llama-index-llms-openai/pyproject.toml

@@ -29,7 +29,7 @@ exclude = ["**/BUILD"]
 license = "MIT"
 name = "llama-index-llms-openai"
 readme = "README.md"
-version = "0.3.9"
+version = "0.3.10"
 
 [tool.poetry.dependencies]
 python = ">=3.9,<4.0"

llama-index-integrations/llms/llama-index-llms-openai/tests/test_openai_utils.py

@@ -67,7 +67,7 @@ def openi_message_dicts_with_function_calling() -> List[ChatCompletionMessagePar
         },
         {
             "role": "assistant",
-            "content": "",
+            "content": None,
             "function_call": {
                 "name": "get_current_weather",
                 "arguments": '{ "location": "Boston, MA"}',

llama-index-integrations/llms/llama-index-llms-text-generation-inference/llama_index/llms/text_generation_inference/base.py

@@ -38,7 +38,7 @@
     to_tgi_messages,
     force_single_tool_call,
     resolve_tgi_function_call,
-    get_max_input_length,
+    get_max_total_tokens,
     resolve_tool_choice,
 )
 from text_generation import (
@@ -101,7 +101,10 @@ class TextGenerationInference(FunctionCallingLLM):
 
     context_window: int = Field(
         default=DEFAULT_CONTEXT_WINDOW,
-        description=("Maximum input length in tokens returned from TGI endpoint"),
+        description=(
+            LLMMetadata.model_fields["context_window"].description
+            + " Maximum total tokens returned from TGI endpoint."
+        ),
     )
     is_chat_model: bool = Field(
         default=True,
@@ -155,7 +158,7 @@ def __init__(
             logger.warning(f"TGI client has no function call support: {e}")
             is_function_calling_model = False
 
-        context_window = get_max_input_length(model_url) or DEFAULT_CONTEXT_WINDOW
+        context_window = get_max_total_tokens(model_url) or DEFAULT_CONTEXT_WINDOW
 
         super().__init__(
             context_window=context_window,

llama-index-integrations/llms/llama-index-llms-text-generation-inference/llama_index/llms/text_generation_inference/utils.py

@@ -24,10 +24,20 @@ def resolve_tgi_function_call(url: str) -> bool:
         )
 
 
-def get_max_input_length(url: str) -> Union[int, None]:
+def get_max_input_tokens(url: str) -> Union[int, None]:
     url = f"{url}/info"
     model_info = dict(requests.get(url).json())
-    return model_info.get("max_input_length", None)
+    tgi_version = model_info.get("version", None)
+    if version.parse(tgi_version) >= version.parse("2.1.0"):
+        return model_info.get("max_input_tokens", None)
+    else:
+        return model_info.get("max_input_length", None)
+
+
+def get_max_total_tokens(url: str) -> Union[int, None]:
+    url = f"{url}/info"
+    model_info = dict(requests.get(url).json())
+    return model_info.get("max_total_tokens", None)
 
 
 def to_tgi_messages(messages: Sequence[ChatMessage]) -> Sequence[Message]:

llama-index-integrations/llms/llama-index-llms-text-generation-inference/pyproject.toml

@@ -27,7 +27,7 @@ exclude = ["**/BUILD"]
 license = "MIT"
 name = "llama-index-llms-text-generation-inference"
 readme = "README.md"
-version = "0.3.0"
+version = "0.3.1"
 
 [tool.poetry.dependencies]
 python = ">=3.9,<4.0"