ai-mirow-projects
/
swarms
mirror of https://github.com/kyegomez/swarms
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Update heavy_swarm.py

Browse Source
pull/1089/head
CI-DEV 2 weeks ago committed by GitHub
parent 145443a5e0
commit ab4391c103
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 107 additions and 106 deletions
Show Stats Download Patch File Download Diff File

213
swarms/structs/heavy_swarm.py
Unescape View File

@ -10,22 +10,15 @@ from typing import Dict, List, Optional
from loguru import logger from loguru import logger
from rich.console import Console from rich.console import Console
from rich.panel import Panel from rich.panel import Panel
from rich.progress import ( from rich.progress import (Progress, SpinnerColumn, TextColumn,
Progress, TimeElapsedColumn)
SpinnerColumn,
TextColumn,
TimeElapsedColumn,
)
from rich.table import Table from rich.table import Table
from swarms.structs.agent import Agent from swarms.structs.agent import Agent
from swarms.structs.conversation import Conversation from swarms.structs.conversation import Conversation
from swarms.tools.tool_type import tool_type
from swarms.utils.formatter import formatter from swarms.utils.formatter import formatter
from swarms.utils.history_output_formatter import ( from swarms.utils.history_output_formatter import history_output_formatter
history_output_formatter,
)
from swarms.utils.litellm_wrapper import LiteLLM from swarms.utils.litellm_wrapper import LiteLLM
from swarms.tools.tool_type import tool_type
RESEARCH_AGENT_PROMPT = """ RESEARCH_AGENT_PROMPT = """
You are an expert Research Agent with exceptional capabilities in: You are an expert Research Agent with exceptional capabilities in:
@ -144,7 +137,8 @@ You approach analysis with:
- Actionable insights - Actionable insights
- Clear communication - Clear communication
Provide precise, data-driven analysis with clear implications and recommendations.""" Provide precise, data-driven analysis with clear implications and
recommendations."""
ALTERNATIVES_AGENT_PROMPT = """ ALTERNATIVES_AGENT_PROMPT = """
You are an expert Alternatives Agent with exceptional capabilities in: You are an expert Alternatives Agent with exceptional capabilities in:
@ -218,7 +212,8 @@ You approach alternatives generation with:
- Implementation focus - Implementation focus
- Value optimization - Value optimization
Provide innovative, practical, and well-evaluated alternative approaches and solutions. Provide innovative, practical, and well-evaluated alternative approaches
and solutions.
""" """
@ -304,7 +299,8 @@ You approach verification with:
- Quality focus - Quality focus
- Practical realism - Practical realism
Provide thorough, objective verification with clear feasibility assessments and risk evaluations.""" Provide thorough, objective verification with clear feasibility
assessments and risk evaluations."""
SYNTHESIS_AGENT_PROMPT = """ SYNTHESIS_AGENT_PROMPT = """
You are an expert Synthesis Agent with advanced capabilities in: You are an expert Synthesis Agent with advanced capabilities in:
@ -378,35 +374,54 @@ You approach synthesis with:
- Value optimization - Value optimization
- Implementation focus - Implementation focus
Provide comprehensive, integrated analysis with clear, actionable recommendations and detailed implementation guidance.""" Provide comprehensive, integrated analysis with clear, actionable
recommendations and detailed implementation guidance."""
schema = { schema = {
"type": "function", "type": "function",
"function": { "function": {
"name": "generate_specialized_questions", "name": "generate_specialized_questions",
"description": "Generate 4 specialized questions for different agent roles to comprehensively analyze a given task", "description": (
"Generate 4 specialized questions for different agent roles to "
"comprehensively analyze a given task"
),
"parameters": { "parameters": {
"type": "object", "type": "object",
"properties": { "properties": {
"thinking": { "thinking": {
"type": "string", "type": "string",
"description": "Your reasoning process for how to break down this task into 4 specialized questions for different agent roles", "description": (
"Your reasoning process for how to break down this task "
"into 4 specialized questions for different agent roles"
),
}, },
"research_question": { "research_question": {
"type": "string", "type": "string",
"description": "A detailed research question for the Research Agent to gather comprehensive background information and data", "description": (
"A detailed research question for the Research Agent to "
"gather comprehensive background information and data"
),
}, },
"analysis_question": { "analysis_question": {
"type": "string", "type": "string",
"description": "An analytical question for the Analysis Agent to examine patterns, trends, and insights", "description": (
"An analytical question for the Analysis Agent to examine "
"patterns, trends, and insights"
),
}, },
"alternatives_question": { "alternatives_question": {
"type": "string", "type": "string",
"description": "A strategic question for the Alternatives Agent to explore different approaches, options, and solutions", "description": (
"A strategic question for the Alternatives Agent to explore "
"different approaches, options, and solutions"
),
}, },
"verification_question": { "verification_question": {
"type": "string", "type": "string",
"description": "A verification question for the Verification Agent to validate findings, check accuracy, and assess feasibility", "description": (
"A verification question for the Verification Agent to "
"validate findings, check accuracy, and assess feasibility"
),
}, },
}, },
"required": [ "required": [
@ -425,9 +440,11 @@ schema = [schema]
class HeavySwarm: class HeavySwarm:
""" """
HeavySwarm is a sophisticated multi-agent orchestration system that decomposes complex tasks HeavySwarm is a sophisticated multi-agent orchestration system that
into specialized questions and executes them using four specialized agents: Research, Analysis, decomposes complex tasks into specialized questions and executes them
Alternatives, and Verification. The results are then synthesized into a comprehensive response. using four specialized agents: Research, Analysis, Alternatives, and
Verification. The results are then synthesized into a comprehensive
response.
This swarm architecture provides robust task analysis through: This swarm architecture provides robust task analysis through:
- Intelligent question generation for specialized agent roles - Intelligent question generation for specialized agent roles
@ -445,12 +462,13 @@ class HeavySwarm:
5. Optional iterative refinement through multiple loops 5. Optional iterative refinement through multiple loops
Key Features: Key Features:
- **Multi-loop Execution**: The max_loops parameter enables iterative refinement where - **Multi-loop Execution**: The max_loops parameter enables iterative
each subsequent loop builds upon the context and results from previous loops refinement where each subsequent loop builds upon the context and
- **Context Preservation**: Conversation history is maintained across all loops, results from previous loops
allowing for deeper analysis and refinement - **Context Preservation**: Conversation history is maintained across
- **Iterative Refinement**: Each loop can refine, improve, or complete aspects all loops, allowing for deeper analysis and refinement
of the analysis based on previous results - **Iterative Refinement**: Each loop can refine, improve, or complete
aspects of the analysis based on previous results
Attributes: Attributes:
name (str): Name identifier for the swarm instance name (str): Name identifier for the swarm instance
@ -485,7 +503,10 @@ class HeavySwarm:
def __init__( def __init__(
self, self,
name: str = "HeavySwarm", name: str = "HeavySwarm",
description: str = "A swarm of agents that can analyze a task and generate specialized questions for each agent role", description: str = (
"A swarm of agents that can analyze a task and generate "
"specialized questions for each agent role"
),
timeout: int = 300, timeout: int = 300,
aggregation_strategy: str = "synthesis", aggregation_strategy: str = "synthesis",
loops_per_agent: int = 1, loops_per_agent: int = 1,
@ -512,27 +533,28 @@ class HeavySwarm:
timeout (int, optional): Maximum execution time per agent in seconds. Defaults to 300. timeout (int, optional): Maximum execution time per agent in seconds. Defaults to 300.
aggregation_strategy (str, optional): Strategy for aggregating results. Currently only aggregation_strategy (str, optional): Strategy for aggregating results. Currently only
'synthesis' is supported. Defaults to "synthesis". 'synthesis' is supported. Defaults to "synthesis".
loops_per_agent (int, optional): Number of execution loops each agent should perform. loops_per_agent (int, optional): Number of execution loops each
Must be greater than 0. Defaults to 1. agent should perform. Must be greater than 0. Defaults to 1.
question_agent_model_name (str, optional): Language model for question generation. question_agent_model_name (str, optional): Language model for
Defaults to "gpt-4o-mini". question generation. Defaults to "gpt-4o-mini".
worker_model_name (str, optional): Language model for specialized worker agents. worker_model_name (str, optional): Language model for specialized
Defaults to "gpt-4o-mini". worker agents. Defaults to "gpt-4o-mini".
verbose (bool, optional): Enable detailed logging and debug output. Defaults to False. verbose (bool, optional): Enable detailed logging and debug output. Defaults to False.
max_workers (int, optional): Maximum concurrent workers for parallel execution. max_workers (int, optional): Maximum concurrent workers for
Defaults to 90% of CPU count. parallel execution. Defaults to 90% of CPU count.
show_dashboard (bool, optional): Enable rich dashboard with progress visualization. show_dashboard (bool, optional): Enable rich dashboard with
Defaults to False. progress visualization. Defaults to False.
agent_prints_on (bool, optional): Enable individual agent output printing. agent_prints_on (bool, optional): Enable individual agent
Defaults to False. output printing. Defaults to False.
output_type (str, optional): Output format type for conversation history. output_type (str, optional): Output format type for conversation
Defaults to "dict-all-except-first". history. Defaults to "dict-all-except-first".
worker_tools (tool_type, optional): Tools available to worker agents for enhanced functionality. worker_tools (tool_type, optional): Tools available to worker
Defaults to None. agents for enhanced functionality. Defaults to None.
random_loops_per_agent (bool, optional): Enable random number of loops per agent (1-10 range). random_loops_per_agent (bool, optional): Enable random number of
Defaults to False. loops per agent (1-10 range). Defaults to False.
max_loops (int, optional): Maximum number of execution loops for the entire swarm. max_loops (int, optional): Maximum number of execution loops for
Each loop builds upon previous results for iterative refinement. Defaults to 1. the entire swarm. Each loop builds upon previous results for
iterative refinement. Defaults to 1.
Raises: Raises:
ValueError: If loops_per_agent is 0 or negative ValueError: If loops_per_agent is 0 or negative
@ -619,17 +641,11 @@ class HeavySwarm:
info_table.add_row("Swarm Name", self.name) info_table.add_row("Swarm Name", self.name)
info_table.add_row("Description", self.description) info_table.add_row("Description", self.description)
info_table.add_row("Timeout", f"{self.timeout}s") info_table.add_row("Timeout", f"{self.timeout}s")
info_table.add_row( info_table.add_row("Loops per Agent", str(self.loops_per_agent))
"Loops per Agent", str(self.loops_per_agent) info_table.add_row("Question Model", self.question_agent_model_name)
)
info_table.add_row(
"Question Model", self.question_agent_model_name
)
info_table.add_row("Worker Model", self.worker_model_name) info_table.add_row("Worker Model", self.worker_model_name)
info_table.add_row("Max Workers", str(self.max_workers)) info_table.add_row("Max Workers", str(self.max_workers))
info_table.add_row( info_table.add_row("Aggregation Strategy", self.aggregation_strategy)
"Aggregation Strategy", self.aggregation_strategy
)
# Display dashboard with professional Arasaka styling # Display dashboard with professional Arasaka styling
self.console.print( self.console.print(
@ -670,9 +686,7 @@ class HeavySwarm:
if self.show_dashboard: if self.show_dashboard:
with Progress( with Progress(
SpinnerColumn(), SpinnerColumn(),
TextColumn( TextColumn("[progress.description]{task.description}"),
"[progress.description]{task.description}"
),
transient=True, transient=True,
console=self.console, console=self.console,
) as progress: ) as progress:
@ -831,7 +845,9 @@ class HeavySwarm:
"Build upon the insights from the previous loop to provide deeper analysis." "Build upon the insights from the previous loop to provide deeper analysis."
) )
if self.verbose: if self.verbose:
logger.info("Subsequent loop: Building upon previous results") logger.info(
"Subsequent loop: Building upon previous results"
)
# Question generation with dashboard # Question generation with dashboard
try: try:
@ -849,7 +865,9 @@ class HeavySwarm:
total=100, total=100,
) )
progress.update(task_gen, advance=30) progress.update(task_gen, advance=30)
questions = self.execute_question_generation(loop_task) questions = self.execute_question_generation(
loop_task
)
progress.update( progress.update(
task_gen, task_gen,
advance=70, advance=70,
@ -857,7 +875,9 @@ class HeavySwarm:
) )
time.sleep(0.5) time.sleep(0.5)
else: else:
questions = self.execute_question_generation(loop_task) questions = self.execute_question_generation(
loop_task
)
self.conversation.add( self.conversation.add(
role="Question Generator Agent", role="Question Generator Agent",
@ -889,7 +909,9 @@ class HeavySwarm:
) )
agent_results = self._execute_agents_parallel( agent_results = self._execute_agents_parallel(
questions=questions, agents=self.agents, img=img questions=questions,
agents=self.agents,
img=img,
) )
except Exception as e: except Exception as e:
@ -979,7 +1001,9 @@ class HeavySwarm:
logger.success("Task iteration completed successfully") logger.success("Task iteration completed successfully")
except Exception as e: except Exception as e:
error_msg = f"Failed to execute loop {current_loop + 1}: {str(e)}" error_msg = (
f"Failed to execute loop {current_loop + 1}: {str(e)}"
)
logger.error(error_msg) logger.error(error_msg)
logger.error(f"Traceback: {traceback.format_exc()}") logger.error(f"Traceback: {traceback.format_exc()}")
return error_msg return error_msg
@ -1005,7 +1029,6 @@ class HeavySwarm:
if self.verbose: if self.verbose:
logger.success("HeavySwarm execution completed") logger.success("HeavySwarm execution completed")
return history_output_formatter( return history_output_formatter(
conversation=self.conversation, conversation=self.conversation,
type=self.output_type, type=self.output_type,
@ -1165,9 +1188,7 @@ class HeavySwarm:
""" """
if self.show_dashboard: if self.show_dashboard:
return self._execute_agents_with_dashboard( return self._execute_agents_with_dashboard(questions, agents, img)
questions, agents, img
)
else: else:
return self._execute_agents_basic(questions, agents, img) return self._execute_agents_basic(questions, agents, img)
@ -1265,14 +1286,10 @@ class HeavySwarm:
} }
# Collect results as they complete # Collect results as they complete
for future in concurrent.futures.as_completed( for future in concurrent.futures.as_completed(future_to_agent):
future_to_agent
):
agent_type = future_to_agent[future] agent_type = future_to_agent[future]
try: try:
agent_name, result = future.result( agent_name, result = future.result(timeout=self.timeout)
timeout=self.timeout
)
results[agent_name.lower()] = result results[agent_name.lower()] = result
except concurrent.futures.TimeoutError: except concurrent.futures.TimeoutError:
logger.error( logger.error(
@ -1285,9 +1302,7 @@ class HeavySwarm:
logger.error( logger.error(
f"❌ Exception in {agent_type} Agent: {str(e)}" f"❌ Exception in {agent_type} Agent: {str(e)}"
) )
results[agent_type.lower()] = ( results[agent_type.lower()] = f"Exception: {str(e)}"
f"Exception: {str(e)}"
)
return results return results
@ -1478,16 +1493,12 @@ class HeavySwarm:
) as executor: ) as executor:
# Submit all agent tasks # Submit all agent tasks
future_to_agent = { future_to_agent = {
executor.submit( executor.submit(execute_agent_with_progress, task): task[1]
execute_agent_with_progress, task
): task[1]
for task in agent_tasks for task in agent_tasks
} }
# Collect results as they complete # Collect results as they complete
for future in concurrent.futures.as_completed( for future in concurrent.futures.as_completed(future_to_agent):
future_to_agent
):
agent_key = future_to_agent[future] agent_key = future_to_agent[future]
try: try:
agent_name, result = future.result( agent_name, result = future.result(
@ -1570,11 +1581,11 @@ class HeavySwarm:
- Use bullet points, numbered lists, and section headings where appropriate for clarity and readability. - Use bullet points, numbered lists, and section headings where appropriate for clarity and readability.
You may reference the conversation history for additional context: You may reference the conversation history for additional context:
\n\n \n\n
{self.conversation.return_history_as_string()} {self.conversation.return_history_as_string()}
\n\n \n\n
Please present your synthesis in the following structure: Please present your synthesis in the following structure:
@ -1640,12 +1651,8 @@ class HeavySwarm:
return { return {
"thinking": arguments.get("thinking", ""), "thinking": arguments.get("thinking", ""),
"research_question": arguments.get( "research_question": arguments.get("research_question", ""),
"research_question", "" "analysis_question": arguments.get("analysis_question", ""),
),
"analysis_question": arguments.get(
"analysis_question", ""
),
"alternatives_question": arguments.get( "alternatives_question": arguments.get(
"alternatives_question", "" "alternatives_question", ""
), ),
@ -1664,9 +1671,7 @@ class HeavySwarm:
"function_name": tool_call.function.name, "function_name": tool_call.function.name,
} }
def execute_question_generation( def execute_question_generation(self, task: str) -> Dict[str, str]:
self, task: str
) -> Dict[str, str]:
""" """
Execute the question generation using the schema with a language model. Execute the question generation using the schema with a language model.
@ -1682,7 +1687,7 @@ class HeavySwarm:
You are an expert task analyzer. Your job is to break down the following task into 4 specialized questions for different agent roles: You are an expert task analyzer. Your job is to break down the following task into 4 specialized questions for different agent roles:
1. Research Agent: Focuses on gathering information, data, and background context 1. Research Agent: Focuses on gathering information, data, and background context
2. Analysis Agent: Focuses on examining patterns, trends, and deriving insights 2. Analysis Agent: Focuses on examining patterns, trends, and deriving insights
3. Alternatives Agent: Focuses on exploring different approaches and solutions 3. Alternatives Agent: Focuses on exploring different approaches and solutions
4. Verification Agent: Focuses on validating findings and checking feasibility 4. Verification Agent: Focuses on validating findings and checking feasibility
@ -1756,12 +1761,8 @@ class HeavySwarm:
return { return {
"research_question": result.get("research_question", ""), "research_question": result.get("research_question", ""),
"analysis_question": result.get("analysis_question", ""), "analysis_question": result.get("analysis_question", ""),
"alternatives_question": result.get( "alternatives_question": result.get("alternatives_question", ""),
"alternatives_question", "" "verification_question": result.get("verification_question", ""),
),
"verification_question": result.get(
"verification_question", ""
),
} }
def get_questions_as_list(self, task: str) -> List[str]: def get_questions_as_list(self, task: str) -> List[str]:

Write Preview
Loading…
Cancel
Save