Added factory for creating LLM providers, added Deepseek provider

examples/try.py

@@ -1,74 +1,50 @@
 import os
 import sys
 
-from langchain_anthropic import ChatAnthropic
-from langchain_openai import ChatOpenAI
-from langchain_google_genai import ChatGoogleGenerativeAI
-
+from mlx_use.llm.factory import LLMFactory
+from mlx_use.llm.providers import LLMConfig, LLM
 
 sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-import argparse
 import asyncio
 
 from mlx_use import Agent
-from pydantic import SecretStr
 from mlx_use.controller.service import Controller
 
+config = LLMConfig(
+    provider_name=LLM.OPENAI,
+    env_var="OPENAI_API_KEY",
+    params={"model": "gpt-4o"}
+)
 
-def set_llm(llm_provider:str = None):
-	if not llm_provider:
-		raise ValueError("No llm provider was set")
-
-	if llm_provider == "OAI" and os.getenv('OPENAI_API_KEY'):
-		return ChatOpenAI(model='gpt-4', api_key=SecretStr(os.getenv('OPENAI_API_KEY')))
-
-	if llm_provider == "google" and os.getenv('GEMINI_API_KEY'):
-		return ChatGoogleGenerativeAI(model='gemini-2.0-flash-exp', api_key=SecretStr(os.getenv('GEMINI_API_KEY')))
-
-	if llm_provider == "anthropic" and os.getenv('ANTHROPIC_API_KEY'):
-		return ChatAnthropic(model='claude-3-sonnet-20240229', api_key=SecretStr(os.getenv('ANTHROPIC_API_KEY')))
-
-	return None
-
-# Try to set LLM based on available API keys
-llm = None
-if os.getenv('GEMINI_API_KEY'):
-	llm = set_llm('google')
-elif os.getenv('OPENAI_API_KEY'):
-	llm = set_llm('OAI')
-elif os.getenv('ANTHROPIC_API_KEY'):
-	llm = set_llm('anthropic')
-
-if not llm:
-	raise ValueError("No API keys found. Please set at least one of GEMINI_API_KEY, OPENAI_API_KEY, or ANTHROPIC_API_KEY in your .env file")
+provider = LLMFactory.create_provider(config)
+llm = provider.get_llm()
 
 controller = Controller()
 
 
 async def main():
-
-	agent_greeting = Agent(
-		task='Say "Hi there $whoami,  What can I do for you today?"',
-		llm=llm,
-		controller=controller,
-		use_vision=False,
-		max_actions_per_step=1,
-		max_failures=5
-	)
-
-	await agent_greeting.run(max_steps=25)
-	task = input("Enter the task: ")
-
-	agent_task = Agent(
-		task=task,
-		llm=llm,
-		controller=controller,
-		use_vision=False,
-		max_actions_per_step=4,
-		max_failures=5
-	)
-
-	await agent_task.run(max_steps=25)
+    agent_greeting = Agent(
+        task='Say "Hi there $whoami,  What can I do for you today?"',
+        llm=llm,
+        controller=controller,
+        use_vision=False,
+        max_actions_per_step=1,
+        max_failures=5
+    )
+
+    await agent_greeting.run(max_steps=25)
+    task = input("Enter the task: ")
+
+    agent_task = Agent(
+        task=task,
+        llm=llm,
+        controller=controller,
+        use_vision=False,
+        max_actions_per_step=4,
+        max_failures=5
+    )
+
+    await agent_task.run(max_steps=25)
 
 
 asyncio.run(main())

mlx_use/llm/factory.py

@@ -0,0 +1,10 @@
+from mlx_use.llm.providers import LLMProvider, GenericLLMProvider, LLMConfig
+
+
+class LLMFactory:
+    """Factory for creating LLM providers"""
+
+    @staticmethod
+    def create_provider(config: LLMConfig) -> LLMProvider:
+        """Create an LLM provider based on the given name"""
+        return GenericLLMProvider(config)

mlx_use/llm/providers.py

@@ -0,0 +1,66 @@
+import os
+from abc import ABC, abstractmethod
+from enum import Enum
+from typing import Optional, Dict, Any
+
+from langchain_anthropic import ChatAnthropic
+from langchain_core.language_models import BaseChatModel
+from langchain_google_genai import ChatGoogleGenerativeAI
+from langchain_openai import ChatOpenAI
+from pydantic import SecretStr
+
+
+class LLM(Enum):
+    DEEPSEEK = "deepseek"
+    CLAUDE = "claude"
+    OPENAI = "openai"
+    GEMINI = "gemini"
+
+
+class LLMConfig:
+    """Configuration class for LLM providers."""
+
+    PROVIDERS = {
+        LLM.DEEPSEEK: {"class": ChatOpenAI, "default_params": {"base_url": "https://api.deepseek.com/v1", "model": "deepseek-chat"}},
+        LLM.CLAUDE: {"class": ChatAnthropic, "default_params": {"model_name": "claude-3-7-sonnet-20250219"}},
+        LLM.OPENAI: {"class": ChatOpenAI, "default_params": {"model": "gpt-4o"}},
+        LLM.GEMINI: {"class": ChatGoogleGenerativeAI, "default_params": {"model": "gemini-2.0-flash"}},
+    }
+
+    def __init__(self, provider_name: LLM, env_var: str, params: Optional[Dict[str, Any]] = None):
+        self.provider_class = self.PROVIDERS[provider_name]["class"]
+        self.env_var = env_var
+        self.params = {**self.PROVIDERS[provider_name].get("default_params", {}), **(params or {})}
+
+    def get_api_key(self) -> str:
+        api_key = os.getenv(self.env_var)
+        if not api_key:
+            raise ValueError(f"{self.env_var} environment variable not set")
+        return api_key
+
+
+class LLMProvider(ABC):
+    """Strategy interface for LLM providers"""
+
+    @abstractmethod
+    def get_llm(self, config: Optional[LLMConfig] = None) -> BaseChatModel:
+        """Return the LLM instance"""
+        pass
+
+
+class GenericLLMProvider(LLMProvider):
+    """Generic LLM provider implementation using dynamic configuration."""
+
+    def __init__(self, config: Optional[LLMConfig] = None):
+        self.config = config
+
+    def get_llm(self, params: Optional[Dict[str, Any]] = None) -> BaseChatModel:
+        try:
+            merged_params = {**self.config.params, **(params or {})}
+            return self.config.provider_class(
+                api_key=SecretStr(self.config.get_api_key()),
+                **merged_params
+            )
+        except Exception as e:
+            print(f"Error initializing {self.config.provider_class} LLM: {e}")
+            raise