tinyagent-py 0.0.7__py3-none-any.whl → 0.0.9__py3-none-any.whl

tinyagent/code_agent/tiny_code_agent.py

@@ -0,0 +1,573 @@
+import traceback
+from textwrap import dedent
+from typing import Optional, List, Dict, Any
+from pathlib import Path
+from tinyagent import TinyAgent, tool
+from tinyagent.hooks.logging_manager import LoggingManager
+from .providers.base import CodeExecutionProvider
+from .providers.modal_provider import ModalProvider
+from .helper import translate_tool_for_code_agent, load_template, render_system_prompt, prompt_code_example, prompt_qwen_helper
+
+
+class TinyCodeAgent:
+    """
+    A TinyAgent specialized for code execution tasks.
+    
+    This class provides a high-level interface for creating agents that can execute
+    Python code using various providers (Modal, Docker, local execution, etc.).
+    """
+    
+    def __init__(
+        self,
+        model: str = "gpt-4.1-mini",
+        api_key: Optional[str] = None,
+        log_manager: Optional[LoggingManager] = None,
+        provider: str = "modal",
+        tools: Optional[List[Any]] = None,
+        code_tools: Optional[List[Any]] = None,
+        authorized_imports: Optional[List[str]] = None,
+        system_prompt_template: Optional[str] = None,
+        provider_config: Optional[Dict[str, Any]] = None,
+        user_variables: Optional[Dict[str, Any]] = None,
+        pip_packages: Optional[List[str]] = None,
+        local_execution: bool = False,
+        **agent_kwargs
+    ):
+        """
+        Initialize TinyCodeAgent.
+        
+        Args:
+            model: The language model to use
+            api_key: API key for the model
+            log_manager: Optional logging manager
+            provider: Code execution provider ("modal", "local", etc.)
+            tools: List of tools available to the LLM (regular tools)
+            code_tools: List of tools available in the Python execution environment
+            authorized_imports: List of authorized Python imports
+            system_prompt_template: Path to custom system prompt template
+            provider_config: Configuration for the code execution provider
+            user_variables: Dictionary of variables to make available in Python environment
+            pip_packages: List of additional Python packages to install in Modal environment
+            local_execution: If True, uses Modal's .local() method for local execution. 
+                                If False, uses Modal's .remote() method for cloud execution (default: False)
+            **agent_kwargs: Additional arguments passed to TinyAgent
+        """
+        self.model = model
+        self.api_key = api_key
+        self.log_manager = log_manager
+        self.tools = tools or []  # LLM tools
+        self.code_tools = code_tools or []  # Python environment tools
+        self.authorized_imports = authorized_imports or ["tinyagent", "gradio", "requests", "asyncio"]
+        self.provider_config = provider_config or {}
+        self.user_variables = user_variables or {}
+        self.pip_packages = pip_packages or []
+        self.local_execution = local_execution
+        
+        # Create the code execution provider
+        self.code_provider = self._create_provider(provider, self.provider_config)
+        
+        # Set user variables in the provider
+        if self.user_variables:
+            self.code_provider.set_user_variables(self.user_variables)
+        
+        # Build system prompt
+        self.system_prompt = self._build_system_prompt(system_prompt_template)
+        
+        # Create the underlying TinyAgent
+        self.agent = TinyAgent(
+            model=model,
+            api_key=api_key,
+            system_prompt=self.system_prompt,
+            logger=log_manager.get_logger('tinyagent.tiny_agent') if log_manager else None,
+            **agent_kwargs
+        )
+        
+        # Add the code execution tool
+        self._setup_code_execution_tool()
+        
+        # Add LLM tools (not code tools - those go to the provider)
+        if self.tools:
+            self.agent.add_tools(self.tools)
+    
+    def _create_provider(self, provider_type: str, config: Dict[str, Any]) -> CodeExecutionProvider:
+        """Create a code execution provider based on the specified type."""
+        if provider_type.lower() == "modal":
+            # Merge pip_packages from both sources (direct parameter and provider_config)
+            config_pip_packages = config.get("pip_packages", [])
+            final_pip_packages = list(set(self.pip_packages + config_pip_packages))
+            
+            final_config = config.copy()
+            final_config["pip_packages"] = final_pip_packages
+            
+            return ModalProvider(
+                log_manager=self.log_manager,
+                code_tools=self.code_tools,
+                local_execution=self.local_execution,
+                **final_config
+            )
+        else:
+            raise ValueError(f"Unsupported provider type: {provider_type}")
+    
+    def _build_system_prompt(self, template_path: Optional[str] = None) -> str:
+        """Build the system prompt for the code agent."""
+        # Use default template if none provided
+        if template_path is None:
+            template_path = str(Path(__file__).parent.parent / "prompts" / "code_agent.yaml")
+        
+        # Translate code tools to code agent format
+        code_tools_metadata = {}
+        for tool in self.code_tools:
+            if hasattr(tool, '_tool_metadata'):
+                metadata = translate_tool_for_code_agent(tool)
+                code_tools_metadata[metadata["name"]] = metadata
+        
+        # Load and render template
+        try:
+            template_str = load_template(template_path)
+            system_prompt = render_system_prompt(
+                template_str, 
+                code_tools_metadata, 
+                {}, 
+                self.authorized_imports
+            )
+            base_prompt = system_prompt + prompt_code_example + prompt_qwen_helper
+        except Exception as e:
+            # Fallback to a basic prompt if template loading fails
+            traceback.print_exc()
+            print(f"Failed to load template from {template_path}: {e}")
+            base_prompt = self._get_fallback_prompt()
+        
+        # Add user variables information to the prompt
+        if self.user_variables:
+            variables_info = self._build_variables_prompt()
+            base_prompt += "\n\n" + variables_info
+        
+        return base_prompt
+    
+    def _get_fallback_prompt(self) -> str:
+        """Get a fallback system prompt if template loading fails."""
+        return dedent("""
+        You are a helpful AI assistant that can execute Python code to solve problems.
+        
+        You have access to a run_python tool that can execute Python code in a sandboxed environment.
+        Use this tool to solve computational problems, analyze data, or perform any task that requires code execution.
+        
+        When writing code:
+        - Always think step by step about the task
+        - Use print() statements to show intermediate results
+        - Handle errors gracefully
+        - Provide clear explanations of your approach
+        
+        The user cannot see the direct output of run_python, so use final_answer to show results.
+        """)
+    
+    def _build_variables_prompt(self) -> str:
+        """Build the variables section for the system prompt."""
+        if not self.user_variables:
+            return ""
+        
+        variables_lines = ["## Available Variables", ""]
+        variables_lines.append("The following variables are pre-loaded and available in your Python environment:")
+        variables_lines.append("")
+        
+        for var_name, var_value in self.user_variables.items():
+            var_type = type(var_value).__name__
+            
+            # Try to get a brief description of the variable
+            if hasattr(var_value, 'shape') and hasattr(var_value, 'dtype'):
+                # Likely numpy array or pandas DataFrame
+                if hasattr(var_value, 'columns'):
+                    # DataFrame
+                    desc = f"DataFrame with shape {var_value.shape} and columns: {list(var_value.columns)}"
+                else:
+                    # Array
+                    desc = f"Array with shape {var_value.shape} and dtype {var_value.dtype}"
+            elif isinstance(var_value, (list, tuple)):
+                length = len(var_value)
+                if length > 0:
+                    first_type = type(var_value[0]).__name__
+                    desc = f"{var_type} with {length} items (first item type: {first_type})"
+                else:
+                    desc = f"Empty {var_type}"
+            elif isinstance(var_value, dict):
+                keys_count = len(var_value)
+                if keys_count > 0:
+                    sample_keys = list(var_value.keys())[:3]
+                    desc = f"Dictionary with {keys_count} keys. Sample keys: {sample_keys}"
+                else:
+                    desc = "Empty dictionary"
+            elif isinstance(var_value, str):
+                length = len(var_value)
+                preview = var_value[:50] + "..." if length > 50 else var_value
+                desc = f"String with {length} characters: '{preview}'"
+            else:
+                desc = f"{var_type}: {str(var_value)[:100]}"
+            
+            variables_lines.append(f"- **{var_name}** ({var_type}): {desc}")
+        
+        variables_lines.extend([
+            "",
+            "These variables are already loaded and ready to use in your code. You don't need to import or define them.",
+            "You can directly reference them by name in your Python code."
+        ])
+        
+        return "\n".join(variables_lines)
+    
+    def _build_code_tools_prompt(self) -> str:
+        """Build the code tools section for the system prompt."""
+        if not self.code_tools:
+            return ""
+        
+        code_tools_lines = ["## Available Code Tools", ""]
+        code_tools_lines.append("The following code tools are available in your Python environment:")
+        code_tools_lines.append("")
+        
+        for tool in self.code_tools:
+            if hasattr(tool, '_tool_metadata'):
+                metadata = translate_tool_for_code_agent(tool)
+                desc = f"- **{metadata['name']}** ({metadata['type']}): {metadata['description']}"
+                code_tools_lines.append(desc)
+        
+        code_tools_lines.extend([
+            "",
+            "These tools are already loaded and ready to use in your code. You don't need to import or define them.",
+            "You can directly reference them by name in your Python code."
+        ])
+        
+        return "\n".join(code_tools_lines)
+    
+    def _setup_code_execution_tool(self):
+        """Set up the run_python tool using the code provider."""
+        @tool(name="run_python", description=dedent("""
+        This tool receives Python code and executes it in a sandboxed environment.
+        During each intermediate step, you can use 'print()' to save important information.
+        These print outputs will appear in the 'Observation:' field for the next step.
+
+        Args:
+            code_lines: list[str]: The Python code to execute as a list of strings.
+                Your code should include all necessary steps for successful execution,
+                cover edge cases, and include error handling.
+                Each line should be an independent line of code.
+
+        Returns:
+            Status of code execution or error message.
+        """))
+        async def run_python(code_lines: List[str], timeout: int = 120) -> str:
+            """Execute Python code using the configured provider."""
+            try:
+                result = await self.code_provider.execute_python(code_lines, timeout)
+                return str(result)
+            except Exception as e:
+                return f"Error executing code: {str(e)}"
+        
+        self.agent.add_tool(run_python)
+    
+    async def run(self, user_input: str, max_turns: int = 10) -> str:
+        """
+        Run the code agent with the given input.
+        
+        Args:
+            user_input: The user's request or question
+            max_turns: Maximum number of conversation turns
+            
+        Returns:
+            The agent's response
+        """
+        return await self.agent.run(user_input, max_turns)
+    
+    async def connect_to_server(self, command: str, args: List[str], **kwargs):
+        """Connect to an MCP server."""
+        return await self.agent.connect_to_server(command, args, **kwargs)
+    
+    def add_callback(self, callback):
+        """Add a callback to the agent."""
+        self.agent.add_callback(callback)
+    
+    def add_tool(self, tool):
+        """Add a tool to the agent (LLM tool)."""
+        self.agent.add_tool(tool)
+    
+    def add_tools(self, tools: List[Any]):
+        """Add multiple tools to the agent (LLM tools)."""
+        self.agent.add_tools(tools)
+    
+    def add_code_tool(self, tool):
+        """
+        Add a code tool that will be available in the Python execution environment.
+        
+        Args:
+            tool: The tool to add to the code execution environment
+        """
+        self.code_tools.append(tool)
+        # Update the provider with the new code tools
+        self.code_provider.set_code_tools(self.code_tools)
+        # Rebuild system prompt to include new code tools info
+        self.system_prompt = self._build_system_prompt()
+        # Update the agent's system prompt
+        self.agent.system_prompt = self.system_prompt
+    
+    def add_code_tools(self, tools: List[Any]):
+        """
+        Add multiple code tools that will be available in the Python execution environment.
+        
+        Args:
+            tools: List of tools to add to the code execution environment
+        """
+        self.code_tools.extend(tools)
+        # Update the provider with the new code tools
+        self.code_provider.set_code_tools(self.code_tools)
+        # Rebuild system prompt to include new code tools info
+        self.system_prompt = self._build_system_prompt()
+        # Update the agent's system prompt
+        self.agent.system_prompt = self.system_prompt
+    
+    def remove_code_tool(self, tool_name: str):
+        """
+        Remove a code tool by name.
+        
+        Args:
+            tool_name: Name of the tool to remove
+        """
+        self.code_tools = [tool for tool in self.code_tools 
+                          if not (hasattr(tool, '_tool_metadata') and 
+                                tool._tool_metadata.get('name') == tool_name)]
+        # Update the provider
+        self.code_provider.set_code_tools(self.code_tools)
+        # Rebuild system prompt
+        self.system_prompt = self._build_system_prompt()
+        # Update the agent's system prompt
+        self.agent.system_prompt = self.system_prompt
+    
+    def get_code_tools(self) -> List[Any]:
+        """
+        Get a copy of current code tools.
+        
+        Returns:
+            List of current code tools
+        """
+        return self.code_tools.copy()
+    
+    def get_llm_tools(self) -> List[Any]:
+        """
+        Get a copy of current LLM tools.
+        
+        Returns:
+            List of current LLM tools
+        """
+        return self.tools.copy()
+    
+    def set_user_variables(self, variables: Dict[str, Any]):
+        """
+        Set user variables that will be available in the Python environment.
+        
+        Args:
+            variables: Dictionary of variable name -> value pairs
+        """
+        self.user_variables = variables.copy()
+        self.code_provider.set_user_variables(self.user_variables)
+        # Rebuild system prompt to include new variables info
+        self.system_prompt = self._build_system_prompt()
+        # Update the agent's system prompt
+        self.agent.system_prompt = self.system_prompt
+    
+    def add_user_variable(self, name: str, value: Any):
+        """
+        Add a single user variable.
+        
+        Args:
+            name: Variable name
+            value: Variable value
+        """
+        self.user_variables[name] = value
+        self.code_provider.set_user_variables(self.user_variables)
+        # Rebuild system prompt to include new variables info
+        self.system_prompt = self._build_system_prompt()
+        # Update the agent's system prompt
+        self.agent.system_prompt = self.system_prompt
+    
+    def remove_user_variable(self, name: str):
+        """
+        Remove a user variable.
+        
+        Args:
+            name: Variable name to remove
+        """
+        if name in self.user_variables:
+            del self.user_variables[name]
+            self.code_provider.set_user_variables(self.user_variables)
+            # Rebuild system prompt
+            self.system_prompt = self._build_system_prompt()
+            # Update the agent's system prompt
+            self.agent.system_prompt = self.system_prompt
+    
+    def get_user_variables(self) -> Dict[str, Any]:
+        """
+        Get a copy of current user variables.
+        
+        Returns:
+            Dictionary of current user variables
+        """
+        return self.user_variables.copy()
+    
+    def add_pip_packages(self, packages: List[str]):
+        """
+        Add additional pip packages to the Modal environment.
+        Note: This requires recreating the provider, so it's best to set packages during initialization.
+        
+        Args:
+            packages: List of package names to install
+        """
+        self.pip_packages.extend(packages)
+        self.pip_packages = list(set(self.pip_packages))  # Remove duplicates
+        
+        # Note: Adding packages after initialization requires recreating the provider
+        # This is expensive, so it's better to set packages during initialization
+        print("⚠️  Warning: Adding packages after initialization requires recreating the Modal environment.")
+        print("   For better performance, set pip_packages during TinyCodeAgent initialization.")
+        
+        # Recreate the provider with new packages
+        self.code_provider = self._create_provider(self.provider, self.provider_config)
+        
+        # Re-set user variables if they exist
+        if self.user_variables:
+            self.code_provider.set_user_variables(self.user_variables)
+    
+    def get_pip_packages(self) -> List[str]:
+        """
+        Get a copy of current pip packages.
+        
+        Returns:
+            List of pip packages that will be installed in Modal
+        """
+        return self.pip_packages.copy()
+    
+    async def close(self):
+        """Clean up resources."""
+        await self.code_provider.cleanup()
+        await self.agent.close()
+    
+    def clear_conversation(self):
+        """Clear the conversation history."""
+        self.agent.clear_conversation()
+    
+    @property
+    def messages(self):
+        """Get the conversation messages."""
+        return self.agent.messages
+    
+    @property
+    def session_id(self):
+        """Get the session ID."""
+        return self.agent.session_id 
+
+
+# Example usage demonstrating both LLM tools and code tools
+async def run_example():
+    """
+    Example demonstrating TinyCodeAgent with both LLM tools and code tools.
+    Also shows how to use local vs remote execution.
+    
+    LLM tools: Available to the LLM for direct calling
+    Code tools: Available in the Python execution environment
+    """
+    from tinyagent import tool
+    
+    # Example LLM tool - available to the LLM for direct calling
+    @tool(name="search_web", description="Search the web for information")
+    async def search_web(query: str) -> str:
+        """Search the web for information."""
+        return f"Search results for: {query}"
+    
+    # Example code tool - available in Python environment
+    @tool(name="data_processor", description="Process data arrays")
+    def data_processor(data: List[float]) -> Dict[str, Any]:
+        """Process a list of numbers and return statistics."""
+        return {
+            "mean": sum(data) / len(data),
+            "max": max(data),
+            "min": min(data),
+            "count": len(data)
+        }
+    
+    print("🚀 Testing TinyCodeAgent with REMOTE execution (Modal)")
+    # Create TinyCodeAgent with remote execution (default)
+    agent_remote = TinyCodeAgent(
+        model="gpt-4.1-mini",
+        tools=[search_web],  # LLM tools
+        code_tools=[data_processor],  # Code tools
+        user_variables={
+            "sample_data": [1, 2, 3, 4, 5, 10, 15, 20]
+        },
+        local_execution=False  # Remote execution via Modal (default)
+    )
+    
+    # Connect to MCP servers
+    await agent_remote.connect_to_server("npx", ["-y", "@openbnb/mcp-server-airbnb", "--ignore-robots-txt"])
+    await agent_remote.connect_to_server("npx", ["-y", "@modelcontextprotocol/server-sequential-thinking"])
+    
+    # Test the remote agent
+    response_remote = await agent_remote.run("""
+    I have some sample data. Please use the data_processor tool in Python to analyze my sample_data
+    and show me the results.
+    """)
+    
+    print("Remote Agent Response:")
+    print(response_remote)
+    print("\n" + "="*80 + "\n")
+    
+    # Now test with local execution
+    print("🏠 Testing TinyCodeAgent with LOCAL execution")
+    agent_local = TinyCodeAgent(
+        model="gpt-4.1-mini",
+        tools=[search_web],  # LLM tools
+        code_tools=[data_processor],  # Code tools
+        user_variables={
+            "sample_data": [1, 2, 3, 4, 5, 10, 15, 20]
+        },
+        local_execution=True  # Local execution
+    )
+    
+    # Connect to MCP servers
+    await agent_local.connect_to_server("npx", ["-y", "@openbnb/mcp-server-airbnb", "--ignore-robots-txt"])
+    await agent_local.connect_to_server("npx", ["-y", "@modelcontextprotocol/server-sequential-thinking"])
+    
+    # Test the local agent
+    response_local = await agent_local.run("""
+    I have some sample data. Please use the data_processor tool in Python to analyze my sample_data
+    and show me the results.
+    """)
+    
+    print("Local Agent Response:")
+    print(response_local)
+    
+    # Demonstrate adding tools dynamically
+    @tool(name="validator", description="Validate processed results")
+    def validator(results: Dict[str, Any]) -> bool:
+        """Validate that results make sense."""
+        return all(key in results for key in ["mean", "max", "min", "count"])
+    
+    # Add a new code tool to both agents
+    agent_remote.add_code_tool(validator)
+    agent_local.add_code_tool(validator)
+    
+    print("\n" + "="*80)
+    print("🔧 Testing with dynamically added tools")
+    
+    # Test both agents with the new tool
+    validation_prompt = "Now validate the previous analysis results using the validator tool."
+    
+    response2_remote = await agent_remote.run(validation_prompt)
+    print("Remote Agent Validation Response:")
+    print(response2_remote)
+    
+    response2_local = await agent_local.run(validation_prompt)
+    print("Local Agent Validation Response:")
+    print(response2_local)
+    
+    await agent_remote.close()
+    await agent_local.close()
+
+
+if __name__ == "__main__":
+    import asyncio
+    asyncio.run(run_example()) 

tinyagent/code_agent/tools/__init__.py

@@ -0,0 +1,3 @@
+from .example_tools import get_weather, get_traffic
+
+__all__ = ["get_weather", "get_traffic"] 

tinyagent/code_agent/tools/example_tools.py

@@ -0,0 +1,41 @@
+from tinyagent import tool
+
+# Global variables to track state across calls
+weather_global = '-'
+traffic_global = '-'
+
+
+@tool(name="get_weather", description="Get the weather for a given city.")
+def get_weather(city: str) -> str:
+    """Get the weather for a given city.
+    Args:
+        city: The city to get the weather for
+
+    Returns:
+        The weather for the given city
+    """
+    import random
+    global weather_global
+    output = f"Last time weather was checked was {weather_global}"
+    weather_global = random.choice(['sunny', 'cloudy', 'rainy', 'snowy'])
+    output += f"\n\nThe weather in {city} is now {weather_global}"
+
+    return output
+
+
+@tool(name="get_traffic", description="Get the traffic for a given city.")
+def get_traffic(city: str) -> str:
+    """Get the traffic for a given city.
+    Args:
+        city: The city to get the traffic for
+
+    Returns:
+        The traffic for the given city
+    """
+    import random
+    global traffic_global
+    output = f"Last time traffic was checked was {traffic_global}"
+    traffic_global = random.choice(['light', 'moderate', 'heavy', 'blocked'])
+    output += f"\n\nThe traffic in {city} is now {traffic_global}"
+
+    return output