langchain-ollama 0.1.0rc0__tar.gz → 0.1.1__tar.gz

{langchain_ollama-0.1.0rc0 → langchain_ollama-0.1.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: langchain-ollama
-Version: 0.1.0rc0
+Version: 0.1.1
 Summary: An integration package connecting Ollama and LangChain
 Home-page: https://github.com/langchain-ai/langchain
 License: MIT
@@ -12,8 +12,9 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Requires-Dist: langchain-core (>=0.2.20,<0.3.0)
-Requires-Dist: ollama (>=0.2.1,<1)
+Requires-Dist: ollama (>=0.3.0,<1)
 Project-URL: Repository, https://github.com/langchain-ai/langchain
+Project-URL: Release Notes, https://github.com/langchain-ai/langchain/releases?q=tag%3A%22langchain-ollama%3D%3D0%22&expanded=true
 Project-URL: Source Code, https://github.com/langchain-ai/langchain/tree/master/libs/partners/ollama
 Description-Content-Type: text/markdown
 
@@ -37,7 +38,7 @@ You can download it [here](https://ollama.com/download).
 ```python
 from langchain_ollama import ChatOllama
 
-llm = ChatOllama(model="llama3")
+llm = ChatOllama(model="llama3-groq-tool-use")
 llm.invoke("Sing a ballad of LangChain.")
 ```
 
@@ -58,7 +59,7 @@ embeddings.embed_query("What is the meaning of life?")
 ```python
 from langchain_ollama import OllamaLLM
 
-llm = OllamaLLM()
+llm = OllamaLLM(model="llama3")
 llm.invoke("The meaning of life is")
 ```
 

{langchain_ollama-0.1.0rc0 → langchain_ollama-0.1.1}/README.md

@@ -18,7 +18,7 @@ You can download it [here](https://ollama.com/download).
 ```python
 from langchain_ollama import ChatOllama
 
-llm = ChatOllama(model="llama3")
+llm = ChatOllama(model="llama3-groq-tool-use")
 llm.invoke("Sing a ballad of LangChain.")
 ```
 
@@ -39,6 +39,6 @@ embeddings.embed_query("What is the meaning of life?")
 ```python
 from langchain_ollama import OllamaLLM
 
-llm = OllamaLLM()
+llm = OllamaLLM(model="llama3")
 llm.invoke("The meaning of life is")
 ```

{langchain_ollama-0.1.0rc0 → langchain_ollama-0.1.1}/langchain_ollama/chat_models.py

@@ -17,7 +17,6 @@ from typing import (
 )
 from uuid import uuid4
 
-import ollama
 from langchain_core.callbacks import (
     CallbackManagerForLLMRun,
 )
@@ -36,11 +35,10 @@ from langchain_core.messages import (
 from langchain_core.messages.ai import UsageMetadata
 from langchain_core.messages.tool import tool_call
 from langchain_core.outputs import ChatGeneration, ChatGenerationChunk, ChatResult
-from langchain_core.pydantic_v1 import BaseModel
 from langchain_core.runnables import Runnable
 from langchain_core.tools import BaseTool
 from langchain_core.utils.function_calling import convert_to_openai_tool
-from ollama import AsyncClient, Message, Options
+from ollama import AsyncClient, Client, Message, Options
 
 
 def _get_usage_metadata_from_generation_info(
@@ -218,9 +216,32 @@ class ChatOllama(BaseChatModel):
         .. code-block:: python
 
             '{"location": "Pune, India", "time_of_day": "morning"}'
+
+    Tool Calling:
+        .. warning::
+            Ollama currently does not support streaming for tools
+
+        .. code-block:: python
+
+            from langchain_ollama import ChatOllama
+            from langchain_core.pydantic_v1 import BaseModel, Field
+
+            class Multiply(BaseModel):
+                a: int = Field(..., description="First integer")
+                b: int = Field(..., description="Second integer")
+
+            ans = await chat.invoke("What is 45*67")
+            ans.tool_calls
+
+        .. code-block:: python
+
+            [{'name': 'Multiply',
+            'args': {'a': 45, 'b': 67},
+            'id': '420c3f3b-df10-4188-945f-eb3abdb40622',
+            'type': 'tool_call'}]
     """  # noqa: E501
 
-    model: str = "llama2"
+    model: str
     """Model name to use."""
 
     mirostat: Optional[int] = None
@@ -269,6 +290,11 @@ class ChatOllama(BaseChatModel):
     """The temperature of the model. Increasing the temperature will
     make the model answer more creatively. (Default: 0.8)"""
 
+    seed: Optional[int] = None
+    """Sets the random number seed to use for generation. Setting this
+    to a specific number will make the model generate the same text for
+    the same prompt."""
+
     stop: Optional[List[str]] = None
     """Sets the stop tokens to use."""
 
@@ -293,6 +319,9 @@ class ChatOllama(BaseChatModel):
     keep_alive: Optional[Union[int, str]] = None
     """How long the model will stay loaded into memory."""
 
+    base_url: Optional[str] = None
+    """Base url the model is hosted under."""
+
     @property
     def _default_params(self) -> Dict[str, Any]:
         """Get the default parameters for calling Ollama."""
@@ -310,6 +339,7 @@ class ChatOllama(BaseChatModel):
                 "repeat_last_n": self.repeat_last_n,
                 "repeat_penalty": self.repeat_penalty,
                 "temperature": self.temperature,
+                "seed": self.seed,
                 "stop": self.stop,
                 "tfs_z": self.tfs_z,
                 "top_k": self.top_k,
@@ -323,7 +353,7 @@ class ChatOllama(BaseChatModel):
     ) -> Sequence[Message]:
         ollama_messages: List = []
         for message in messages:
-            role = ""
+            role: Literal["user", "assistant", "system", "tool"]
             tool_call_id: Optional[str] = None
             tool_calls: Optional[List[Dict[str, Any]]] = None
             if isinstance(message, HumanMessage):
@@ -360,11 +390,13 @@ class ChatOllama(BaseChatModel):
                         image_url = None
                         temp_image_url = content_part.get("image_url")
                         if isinstance(temp_image_url, str):
-                            image_url = content_part["image_url"]
+                            image_url = temp_image_url
                         elif (
-                            isinstance(temp_image_url, dict) and "url" in temp_image_url
+                            isinstance(temp_image_url, dict)
+                            and "url" in temp_image_url
+                            and isinstance(temp_image_url["url"], str)
                         ):
-                            image_url = temp_image_url
+                            image_url = temp_image_url["url"]
                         else:
                             raise ValueError(
                                 "Only string image_url or dict with string 'url' "
@@ -385,15 +417,16 @@ class ChatOllama(BaseChatModel):
                             "Must either have type 'text' or type 'image_url' "
                             "with a string 'image_url' field."
                         )
-            msg = {
+            # Should convert to ollama.Message once role includes tool, and tool_call_id is in Message # noqa: E501
+            msg: dict = {
                 "role": role,
                 "content": content,
                 "images": images,
             }
+            if tool_calls:
+                msg["tool_calls"] = tool_calls  # type: ignore
             if tool_call_id:
                 msg["tool_call_id"] = tool_call_id
-            if tool_calls:
-                msg["tool_calls"] = tool_calls
             ollama_messages.append(msg)
 
         return ollama_messages
@@ -415,15 +448,26 @@ class ChatOllama(BaseChatModel):
                 params[key] = kwargs[key]
 
         params["options"]["stop"] = stop
-        async for part in await AsyncClient().chat(
-            model=params["model"],
-            messages=ollama_messages,
-            stream=True,
-            options=Options(**params["options"]),
-            keep_alive=params["keep_alive"],
-            format=params["format"],
-        ):  # type:ignore
-            yield part
+        if "tools" in kwargs:
+            yield await AsyncClient(host=self.base_url).chat(
+                model=params["model"],
+                messages=ollama_messages,
+                stream=False,
+                options=Options(**params["options"]),
+                keep_alive=params["keep_alive"],
+                format=params["format"],
+                tools=kwargs["tools"],
+            )  # type:ignore
+        else:
+            async for part in await AsyncClient(host=self.base_url).chat(
+                model=params["model"],
+                messages=ollama_messages,
+                stream=True,
+                options=Options(**params["options"]),
+                keep_alive=params["keep_alive"],
+                format=params["format"],
+            ):  # type:ignore
+                yield part
 
     def _create_chat_stream(
         self,
@@ -443,25 +487,17 @@ class ChatOllama(BaseChatModel):
 
         params["options"]["stop"] = stop
         if "tools" in kwargs:
-            # tools not supported by sdk yet.
-            req = {
-                "model": params["model"],
-                "messages": ollama_messages,
-                "stream": False,
-                "format": params["format"],
-                "options": Options(**params["options"]),
-                "keep_alive": params["keep_alive"],
-                "tools": kwargs["tools"],
-            }
-            it = ollama._client._request_stream(
-                "POST",
-                "/api/chat",
-                json=req,
+            yield Client(host=self.base_url).chat(
+                model=params["model"],
+                messages=ollama_messages,
                 stream=False,
+                options=Options(**params["options"]),
+                keep_alive=params["keep_alive"],
+                format=params["format"],
+                tools=kwargs["tools"],
             )
-            yield cast(Mapping[str, Any], it)
         else:
-            yield from ollama.chat(
+            yield from Client(host=self.base_url).chat(
                 model=params["model"],
                 messages=ollama_messages,
                 stream=True,
@@ -686,8 +722,19 @@ class ChatOllama(BaseChatModel):
 
     def bind_tools(
         self,
-        tools: Sequence[Union[Dict[str, Any], Type[BaseModel], Callable, BaseTool]],
+        tools: Sequence[Union[Dict[str, Any], Type, Callable, BaseTool]],
         **kwargs: Any,
     ) -> Runnable[LanguageModelInput, BaseMessage]:
+        """Bind tool-like objects to this chat model.
+
+        Assumes model is compatible with OpenAI tool-calling API.
+
+        Args:
+            tools: A list of tool definitions to bind to this chat model.
+                Supports any tool definition handled by
+                :meth:`langchain_core.utils.function_calling.convert_to_openai_tool`.
+            kwargs: Any additional parameters are passed directly to
+                ``self.bind(**kwargs)``.
+        """  # noqa: E501
         formatted_tools = [convert_to_openai_tool(tool) for tool in tools]
         return super().bind(tools=formatted_tools, **kwargs)

{langchain_ollama-0.1.0rc0 → langchain_ollama-0.1.1}/langchain_ollama/embeddings.py

@@ -14,11 +14,11 @@ class OllamaEmbeddings(BaseModel, Embeddings):
 
             from langchain_ollama import OllamaEmbeddings
 
-            model = OllamaEmbeddings(model="llama3")
+            embedder = OllamaEmbeddings(model="llama3")
             embedder.embed_query("what is the place that jonathan worked at?")
     """
 
-    model: str = "llama2"
+    model: str
     """Model name to use."""
 
     class Config:
@@ -28,9 +28,7 @@ class OllamaEmbeddings(BaseModel, Embeddings):
 
     def embed_documents(self, texts: List[str]) -> List[List[float]]:
         """Embed search docs."""
-        embedded_docs = []
-        for doc in texts:
-            embedded_docs.append(list(ollama.embeddings(self.model, doc)["embedding"]))
+        embedded_docs = ollama.embed(self.model, texts)["embeddings"]
         return embedded_docs
 
     def embed_query(self, text: str) -> List[float]:
@@ -39,11 +37,7 @@ class OllamaEmbeddings(BaseModel, Embeddings):
 
     async def aembed_documents(self, texts: List[str]) -> List[List[float]]:
         """Embed search docs."""
-        embedded_docs = []
-        for doc in texts:
-            embedded_docs.append(
-                list((await AsyncClient().embeddings(self.model, doc))["embedding"])
-            )
+        embedded_docs = (await AsyncClient().embed(self.model, texts))["embeddings"]
         return embedded_docs
 
     async def aembed_query(self, text: str) -> List[float]:

{langchain_ollama-0.1.0rc0 → langchain_ollama-0.1.1}/langchain_ollama/llms.py

@@ -34,7 +34,7 @@ class OllamaLLM(BaseLLM):
             model.invoke("Come up with 10 names for a song about parrots")
     """
 
-    model: str = "llama2"
+    model: str
     """Model name to use."""
 
     mirostat: Optional[int] = None
@@ -205,9 +205,9 @@ class OllamaLLM(BaseLLM):
             if not isinstance(stream_resp, str):
                 chunk = GenerationChunk(
                     text=stream_resp["response"] if "response" in stream_resp else "",
-                    generation_info=dict(stream_resp)
-                    if stream_resp.get("done") is True
-                    else None,
+                    generation_info=(
+                        dict(stream_resp) if stream_resp.get("done") is True else None
+                    ),
                 )
                 if final_chunk is None:
                     final_chunk = chunk
@@ -237,9 +237,9 @@ class OllamaLLM(BaseLLM):
             if not isinstance(stream_resp, str):
                 chunk = GenerationChunk(
                     text=stream_resp["response"] if "response" in stream_resp else "",
-                    generation_info=dict(stream_resp)
-                    if stream_resp.get("done") is True
-                    else None,
+                    generation_info=(
+                        dict(stream_resp) if stream_resp.get("done") is True else None
+                    ),
                 )
                 if final_chunk is None:
                     final_chunk = chunk
@@ -304,12 +304,10 @@ class OllamaLLM(BaseLLM):
         for stream_resp in self._create_generate_stream(prompt, stop, **kwargs):
             if not isinstance(stream_resp, str):
                 chunk = GenerationChunk(
-                    text=stream_resp["message"]["content"]
-                    if "message" in stream_resp
-                    else "",
-                    generation_info=dict(stream_resp)
-                    if stream_resp.get("done") is True
-                    else None,
+                    text=(stream_resp.get("response", "")),
+                    generation_info=(
+                        dict(stream_resp) if stream_resp.get("done") is True else None
+                    ),
                 )
                 if run_manager:
                     run_manager.on_llm_new_token(
@@ -328,12 +326,10 @@ class OllamaLLM(BaseLLM):
         async for stream_resp in self._acreate_generate_stream(prompt, stop, **kwargs):
             if not isinstance(stream_resp, str):
                 chunk = GenerationChunk(
-                    text=stream_resp["message"]["content"]
-                    if "message" in stream_resp
-                    else "",
-                    generation_info=dict(stream_resp)
-                    if stream_resp.get("done") is True
-                    else None,
+                    text=(stream_resp.get("response", "")),
+                    generation_info=(
+                        dict(stream_resp) if stream_resp.get("done") is True else None
+                    ),
                 )
                 if run_manager:
                     await run_manager.on_llm_new_token(

{langchain_ollama-0.1.0rc0 → langchain_ollama-0.1.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "langchain-ollama"
-version = "0.1.0rc0"
+version = "0.1.1"
 description = "An integration package connecting Ollama and LangChain"
 authors = []
 readme = "README.md"
@@ -9,10 +9,11 @@ license = "MIT"
 
 [tool.poetry.urls]
 "Source Code" = "https://github.com/langchain-ai/langchain/tree/master/libs/partners/ollama"
+"Release Notes" = "https://github.com/langchain-ai/langchain/releases?q=tag%3A%22langchain-ollama%3D%3D0%22&expanded=true"
 
 [tool.poetry.dependencies]
 python = ">=3.8.1,<4.0"
-ollama = ">=0.2.1,<1"
+ollama = ">=0.3.0,<1"
 langchain-core = "^0.2.20"
 
 [tool.poetry.group.test]