Merge pull request #276 from MervinPraison/develop

adding autoagents

Dockerfile

@@ -1,6 +1,6 @@
 FROM python:3.11-slim
 WORKDIR /app
 COPY . .
-RUN pip install flask praisonai==2.0.36 gunicorn markdown
+RUN pip install flask praisonai==2.0.37 gunicorn markdown
 EXPOSE 8080
 CMD ["gunicorn", "-b", "0.0.0.0:8080", "api:app"]

agents/auto_agents_example.py

@@ -0,0 +1,11 @@
+from praisonaiagents import AutoAgents
+from praisonaiagents.tools import duckduckgo
+
+agents = AutoAgents(
+    instructions="Search for information about AI Agents",
+    tools=[duckduckgo],
+    process="sequential",
+    verbose=True
+)
+
+agents.start()

agents/praisonaiagents/__init__.py

@@ -6,6 +6,7 @@
 from .agents.agents import PraisonAIAgents
 from .task.task import Task
 from .tools.tools import Tools
+from .agents.autoagents import AutoAgents
 from .main import (
     TaskOutput,
     ReflectionOutput,
@@ -33,6 +34,7 @@
     'Task',
     'TaskOutput',
     'ReflectionOutput',
+    'AutoAgents',
     'display_interaction',
     'display_self_reflection',
     'display_instruction',

agents/praisonaiagents/agent/agent.py

@@ -733,4 +733,8 @@ async def _achat_completion(self, response, tools):
             return None
         except Exception as e:
             display_error(f"Error in _achat_completion: {e}")
-            return None 
+            return None 
+
+    def run(self):
+        """Alias for start() method"""
+        return self.start() 

agents/praisonaiagents/agents/__init__.py

@@ -1,4 +1,5 @@
 """Agents module for managing multiple AI agents"""
 from .agents import PraisonAIAgents
+from .autoagents import AutoAgents
 
-__all__ = ['PraisonAIAgents'] 
+__all__ = ['PraisonAIAgents', 'AutoAgents'] 

agents/praisonaiagents/agents/autoagents.py

@@ -0,0 +1,335 @@
+"""
+AutoAgents - A class for automatically creating and managing AI agents and tasks.
+
+This class provides a simplified interface for creating and running AI agents with tasks.
+It automatically handles agent creation, task setup, and execution flow.
+"""
+
+from .agents import PraisonAIAgents
+from ..agent.agent import Agent
+from ..task.task import Task
+from typing import List, Any, Optional, Dict, Union
+import logging
+import os
+from pydantic import BaseModel, Field
+from ..main import display_instruction, display_tool_call, display_interaction, client
+
+# Define Pydantic models for structured output
+class TaskConfig(BaseModel):
+    name: str
+    description: str
+    expected_output: str
+    tools: List[str]
+
+class AgentConfig(BaseModel):
+    name: str
+    role: str
+    goal: str
+    backstory: str
+    tools: List[str]
+    tasks: List[TaskConfig]
+
+class AutoAgentsConfig(BaseModel):
+    main_instruction: str
+    process_type: str
+    agents: List[AgentConfig]
+
+class AutoAgents(PraisonAIAgents):
+    def __init__(
+        self,
+        instructions: str,
+        tools: Optional[List[Any]] = None,
+        verbose: bool = False,
+        process: str = "sequential",
+        manager_llm: Optional[str] = None,
+        max_retries: int = 5,
+        completion_checker: Optional[Any] = None,
+        allow_code_execution: bool = False,
+        memory: bool = True,
+        markdown: bool = True,
+        self_reflect: bool = False,
+        max_reflect: int = 3,
+        min_reflect: int = 1,
+        llm: Optional[str] = None,
+        function_calling_llm: Optional[str] = None,
+        respect_context_window: bool = True,
+        code_execution_mode: str = "safe",
+        embedder_config: Optional[Dict[str, Any]] = None,
+        knowledge_sources: Optional[List[Any]] = None,
+        use_system_prompt: bool = True,
+        cache: bool = True,
+        allow_delegation: bool = False,
+        step_callback: Optional[Any] = None,
+        system_template: Optional[str] = None,
+        prompt_template: Optional[str] = None,
+        response_template: Optional[str] = None,
+        max_rpm: Optional[int] = None,
+        max_execution_time: Optional[int] = None,
+        max_iter: int = 20,
+        reflect_llm: Optional[str] = None,
+        base_url: Optional[str] = None,
+        api_key: Optional[str] = None,
+        max_agents: int = 3  # New parameter for maximum number of agents
+    ):
+        """Initialize AutoAgents with configuration for automatic agent and task creation."""
+        if max_agents < 1:
+            raise ValueError("max_agents must be at least 1")
+        if max_agents > 10:
+            raise ValueError("max_agents cannot exceed 10")
+
+        # Store all configuration parameters first
+        self.instructions = instructions
+        self.tools = tools or []
+        self.verbose = verbose  # Store verbose parameter
+        self.max_agents = max_agents  # Store max_agents parameter
+        self.allow_code_execution = allow_code_execution
+        self.memory = memory
+        self.markdown = markdown
+        self.self_reflect = self_reflect
+        self.max_reflect = max_reflect
+        self.min_reflect = min_reflect
+        self.llm = llm or os.getenv('OPENAI_MODEL_NAME', 'gpt-4o')
+        self.function_calling_llm = function_calling_llm
+        self.respect_context_window = respect_context_window
+        self.code_execution_mode = code_execution_mode
+        self.embedder_config = embedder_config
+        self.knowledge_sources = knowledge_sources
+        self.use_system_prompt = use_system_prompt
+        self.cache = cache
+        self.allow_delegation = allow_delegation
+        self.step_callback = step_callback
+        self.system_template = system_template
+        self.prompt_template = prompt_template
+        self.response_template = response_template
+        self.max_rpm = max_rpm
+        self.max_execution_time = max_execution_time
+        self.max_iter = max_iter
+        self.reflect_llm = reflect_llm
+
+        # Display initial instruction
+        if self.verbose:
+            display_instruction(f"🎯 Main Task: {self.instructions}")
+            display_instruction(f"📊 Maximum Agents: {self.max_agents}")
+            if self.tools:
+                tool_names = [t.__name__ if hasattr(t, '__name__') else str(t) for t in self.tools]
+                display_tool_call(f"🛠️  Available Tools: {', '.join(tool_names)}")
+
+        # Generate agents and tasks configuration
+        config = self._generate_config()
+        agents, tasks = self._create_agents_and_tasks(config)
+
+        # Display agents and their tasks
+        if self.verbose:
+            self._display_agents_and_tasks(agents, tasks)
+
+        # Initialize parent class with generated agents and tasks
+        super().__init__(
+            agents=agents,
+            tasks=tasks,
+            verbose=verbose,
+            completion_checker=completion_checker,
+            max_retries=max_retries,
+            process=process,
+            manager_llm=manager_llm
+        )
+
+    def _display_agents_and_tasks(self, agents: List[Agent], tasks: List[Task]):
+        """Display the created agents and their assigned tasks"""
+        display_instruction("\n🤖 Generated Agents and Tasks:")
+
+        # Create a mapping of agents to their tasks
+        agent_tasks = {}
+        for task in tasks:
+            if task.agent not in agent_tasks:
+                agent_tasks[task.agent] = []
+            agent_tasks[task.agent].append(task)
+
+        # Display each agent and their tasks
+        for agent in agents:
+            agent_tools = [t.__name__ if hasattr(t, '__name__') else str(t) for t in agent.tools]
+            display_interaction(
+                f"\n👤 Agent: {agent.name}",
+                f"""Role: {agent.role}
+Goal: {agent.goal}
+Tools: {', '.join(agent_tools)}"""
+            )
+
+            # Display tasks for this agent
+            if agent in agent_tasks:
+                for i, task in enumerate(agent_tasks[agent], 1):
+                    task_tools = [t.__name__ if hasattr(t, '__name__') else str(t) for t in task.tools]
+                    display_instruction(
+                        f"""  📋 Task {i}: {task.name}
+     Description: {task.description}
+     Expected Output: {task.expected_output}
+     Tools: {', '.join(task_tools)}"""
+                    )
+
+    def _get_available_tools(self) -> List[str]:
+        """Get list of available tools"""
+        if not self.tools:
+            return []
+        return [t.__name__ if hasattr(t, '__name__') else str(t) for t in self.tools]
+
+    def _get_tool_by_name(self, tool_name: str) -> Optional[Any]:
+        """Get tool object by its name"""
+        for tool in self.tools:
+            if (hasattr(tool, '__name__') and tool.__name__ == tool_name) or str(tool) == tool_name:
+                return tool
+        return None
+
+    def _assign_tools_to_agent(self, agent_config: AgentConfig) -> List[Any]:
+        """
+        Assign appropriate tools to an agent based on its role and tasks.
+        
+        Args:
+            agent_config: The agent configuration containing role and required tools
+            
+        Returns:
+            List of tool objects assigned to this agent
+        """
+        assigned_tools = []
+        tool_names = set(agent_config.tools)
+
+        # Also look at task requirements
+        for task in agent_config.tasks:
+            tool_names.update(task.tools)
+
+        # Assign tools that match the requirements
+        for tool_name in tool_names:
+            tool = self._get_tool_by_name(tool_name)
+            if tool:
+                assigned_tools.append(tool)
+
+        # If no specific tools matched but we have tools available,
+        # assign all tools to ensure functionality
+        if not assigned_tools and self.tools:
+            assigned_tools = self.tools
+
+        return assigned_tools
+
+    def _generate_config(self) -> AutoAgentsConfig:
+        """Generate the configuration for agents and tasks"""
+        prompt = f"""
+Generate a configuration for AI agents to accomplish this task: "{self.instructions}"
+
+The configuration should include:
+1. A main instruction that clearly states the overall goal
+2. A process type (sequential, workflow, or hierarchical)
+3. A list of maximum {self.max_agents} agents (no more, no less), each with:
+   - Name, role, goal, and backstory
+   - List of required tools from: {self._get_available_tools()}
+   - Only add tools that are needed for the agent to perform its task
+   - Only one task per agent. Add more than one task if absolutely necessary.
+   - List of specific tasks they need to perform
+   - Whether they should self-reflect or allow delegation
+
+Requirements:
+1. Each agent should have clear, focused responsibilities
+2. Tasks should be broken down into manageable steps
+3. Tool selection should be appropriate for each task
+4. The process type should match the task requirements
+5. Generate maximum {self.max_agents} agents to handle this task efficiently
+
+Return the configuration in a structured JSON format matching the AutoAgentsConfig schema.
+"""
+
+        try:
+            response = client.beta.chat.completions.parse(
+                model=self.llm,
+                response_format=AutoAgentsConfig,
+                messages=[
+                    {"role": "system", "content": "You are a helpful assistant designed to generate AI agent configurations."},
+                    {"role": "user", "content": prompt}
+                ]
+            )
+
+            # Ensure we have exactly max_agents number of agents
+            if len(response.choices[0].message.parsed.agents) > self.max_agents:
+                response.choices[0].message.parsed.agents = response.choices[0].message.parsed.agents[:self.max_agents]
+            elif len(response.choices[0].message.parsed.agents) < self.max_agents:
+                logging.warning(f"Generated {len(response.choices[0].message.parsed.agents)} agents, expected {self.max_agents}")
+
+            return response.choices[0].message.parsed
+        except Exception as e:
+            logging.error(f"Error generating configuration: {e}")
+            raise
+
+    def _create_agents_and_tasks(self, config: AutoAgentsConfig) -> tuple[List[Agent], List[Task]]:
+        """Create agents and tasks from configuration"""
+        agents = []
+        tasks = []
+
+        for agent_config in config.agents:
+            # Get appropriate tools for this agent
+            agent_tools = self._assign_tools_to_agent(agent_config)
+
+            # Create the agent with all parameters
+            agent = Agent(
+                name=agent_config.name,
+                role=agent_config.role,
+                goal=agent_config.goal,
+                backstory=agent_config.backstory,
+                tools=agent_tools,  # Use assigned tools
+                verbose=self.verbose >= 1,
+                allow_code_execution=self.allow_code_execution,
+                memory=self.memory,
+                markdown=self.markdown,
+                self_reflect=self.self_reflect,
+                max_reflect=self.max_reflect,
+                min_reflect=self.min_reflect,
+                llm=self.llm,
+                function_calling_llm=self.function_calling_llm,
+                respect_context_window=self.respect_context_window,
+                code_execution_mode=self.code_execution_mode,
+                embedder_config=self.embedder_config,
+                knowledge_sources=self.knowledge_sources,
+                use_system_prompt=self.use_system_prompt,
+                cache=self.cache,
+                allow_delegation=self.allow_delegation,
+                step_callback=self.step_callback,
+                system_template=self.system_template,
+                prompt_template=self.prompt_template,
+                response_template=self.response_template,
+                max_rpm=self.max_rpm,
+                max_execution_time=self.max_execution_time,
+                max_iter=self.max_iter,
+                reflect_llm=self.reflect_llm
+            )
+            agents.append(agent)
+
+            # Create tasks for this agent
+            for task_config in agent_config.tasks:
+                # Get task-specific tools
+                task_tools = [self._get_tool_by_name(t) for t in task_config.tools]
+                task_tools = [t for t in task_tools if t]  # Remove None values
+
+                # If no specific tools matched, use agent's tools
+                if not task_tools:
+                    task_tools = agent_tools
+
+                task = Task(
+                    name=task_config.name,
+                    description=task_config.description,
+                    expected_output=task_config.expected_output,
+                    agent=agent,
+                    tools=task_tools  # Use task-specific tools
+                )
+                tasks.append(task)
+
+        return agents, tasks
+
+    async def astart(self):
+        """
+        Async version of start() method.
+        Creates tasks based on the instructions, then starts execution.
+        Returns the task status and results dictionary.
+        """
+        return await super().astart()
+
+    def start(self):
+        """
+        Creates tasks based on the instructions, then starts execution.
+        Returns the task status and results dictionary.
+        """
+        return super().start()