Support Ollama models (#1596)

mem0/configs/llms/base.py

@@ -11,7 +11,8 @@ class BaseLlmConfig(ABC):
         model: Optional[str] = None,
         temperature: float = 0,
         max_tokens: int = 3000,
-        top_p: float = 1
+        top_p: float = 1,
+        base_url: Optional[str] = None
     ):
         """
         Initializes a configuration class instance for the LLM.
@@ -26,9 +27,12 @@ class BaseLlmConfig(ABC):
         :param top_p: Controls the diversity of words. Higher values (closer to 1) make word selection more diverse,
         defaults to 1
         :type top_p: float, optional
+        :param base_url: The base URL of the LLM, defaults to None
+        :type base_url: Optional[str], optional
         """
         
         self.model = model
         self.temperature = temperature
         self.max_tokens = max_tokens
-        self.top_p = top_p
+        self.top_p = top_p
+        self.base_url = base_url

mem0/llms/ollama.py

@@ -1,29 +1,90 @@
-import ollama
-from mem0.llms.base import LLMBase
+from typing import Dict, List, Optional
 
+try:
+    from ollama import Client
+except ImportError:
+    raise ImportError("Ollama requires extra dependencies. Install with `pip install ollama`") from None
+
+from mem0.llms.base import LLMBase
+from mem0.configs.llms.base import BaseLlmConfig
 
 class OllamaLLM(LLMBase):
-    def __init__(self, model="llama3"):
-        self.model = model
+    def __init__(self, config: Optional[BaseLlmConfig] = None):
+        super().__init__(config)
+
+        if not self.config.model:
+            self.config.model="llama3.1:70b"
+        self.client = Client(host=self.config.base_url)
         self._ensure_model_exists()
 
     def _ensure_model_exists(self):
         """
         Ensure the specified model exists locally. If not, pull it from Ollama.
+        """        
+        local_models = self.client.list()["models"]
+        if not any(model.get("name") == self.config.model for model in local_models):
+            self.client.pull(self.config.model)
+    
+    def _parse_response(self, response, tools):
         """
-        model_list = [m["name"] for m in ollama.list()["models"]]
-        if not any(m.startswith(self.model) for m in model_list):
-            ollama.pull(self.model)
+        Process the response based on whether tools are used or not.
 
-    def generate_response(self, messages):
+        Args:
+            response: The raw response from API.
+            tools: The list of tools provided in the request.
+
+        Returns:
+            str or dict: The processed response.
         """
-        Generate a response based on the given messages using Ollama.
+        if tools:
+            processed_response = {
+                "content": response['message']['content'],
+                "tool_calls": []
+            }
+            
+            if response['message'].get('tool_calls'):
+                for tool_call in response['message']['tool_calls']:
+                    processed_response["tool_calls"].append({
+                        "name": tool_call["function"]["name"],
+                        "arguments": tool_call["function"]["arguments"]
+                    })
+            
+            return processed_response
+        else:
+            return response['message']['content']
+
+    def generate_response(
+        self,
+        messages: List[Dict[str, str]],
+        response_format=None,
+        tools: Optional[List[Dict]] = None,
+        tool_choice: str = "auto",
+    ):
+        """
+        Generate a response based on the given messages using OpenAI.
 
         Args:
             messages (list): List of message dicts containing 'role' and 'content'.
+            response_format (str or object, optional): Format of the response. Defaults to "text".
+            tools (list, optional): List of tools that the model can call. Defaults to None.
+            tool_choice (str, optional): Tool choice method. Defaults to "auto".
 
         Returns:
             str: The generated response.
         """
-        response = ollama.chat(model=self.model, messages=messages)
-        return response["message"]["content"]
+        params = {
+            "model": self.config.model, 
+            "messages": messages, 
+            "options": {
+                "temperature": self.config.temperature, 
+                "num_predict": self.config.max_tokens, 
+                "top_p": self.config.top_p
+            }  
+        }
+        if response_format:
+            params["format"] = response_format
+        if tools:
+            params["tools"] = tools
+
+        response = self.client.chat(**params)
+        return self._parse_response(response, tools)

mem0/utils/factory.py

@@ -17,6 +17,7 @@ class LlmFactory:
         "together": "mem0.llms.together.TogetherLLM",
         "aws_bedrock": "mem0.llms.aws_bedrock.AWSBedrockLLM",
         "litellm": "mem0.llms.litellm.LiteLLM",
+        "ollama": "mem0.llms.ollama.OllamaLLM",
     }
 
     @classmethod