socratic-workflow 0.1.4

Production-grade workflow orchestration system with LLM cost tracking, performance analytics, and dependency management.

Socratic Workflow

Why Socratic Workflow?

Building multi-step AI workflows is complex. Socratic Workflow handles the production challenges:

• Cost Tracking - Track LLM costs across 16+ models and 5 providers (Claude, GPT-4, Gemini, Llama, Mistral)
• Dependency Management - Define complex task graphs with automatic dependency resolution
• Parallel Execution - Run independent tasks concurrently for maximum performance
• Error Recovery - Automatic retry logic with exponential backoff and graceful degradation
• Performance Analytics - Measure execution time, bottlenecks, and success rates in real-time

Production-grade workflow orchestration system with LLM cost tracking, performance analytics, and dependency management.

Features

• 🔄 Workflow Orchestration - Define and execute multi-step AI workflows
• 💰 Cost Tracking - Track LLM costs across providers (Claude, GPT-4, Gemini, etc.)
• 📊 Performance Analytics - Measure execution time, success rates, bottlenecks
• 🔗 Task Dependencies - Manage complex task graphs with dependencies
• ⚡ Parallel Execution - Run independent tasks concurrently for speed
• 💾 State Management - Save/restore workflow state for resilience
• 🔄 Error Recovery - Retry logic and graceful degradation
• 🔌 Framework Integration - Works with Openclaw, LangChain, Socratic Agents

Quick Start

Installation

# Basic installation
pip install socratic-workflow

# With LangChain integration
pip install socratic-workflow[langchain]

# With Openclaw integration
pip install socratic-workflow[openclaw]

# With everything
pip install socratic-workflow[all]

Simple Workflow

from socratic_workflow import Workflow, Task, WorkflowEngine

# Define a custom task
class GreetingTask(Task):
    def execute(self, context):
        name = self.config.get("name", "World")
        return {"greeting": f"Hello, {name}!"}

# Build workflow
workflow = Workflow("Greeting Pipeline")
workflow.add_task("greet", GreetingTask(name="Alice"))

# Execute
engine = WorkflowEngine()
result = engine.execute(workflow)

print(result.success)  # True
print(result.task_results)  # {"greet": {"greeting": "Hello, Alice!"}}
print(f"Execution time: {result.duration:.2f}s")

Workflow with Dependencies

from socratic_workflow import Workflow, SimpleTask, WorkflowEngine

# Create workflow with dependencies
workflow = Workflow("Multi-Step Pipeline")
workflow.add_task("step1", SimpleTask(result={"data": "processed"}))
workflow.add_task("step2", SimpleTask(result={"analysis": "complete"}), depends_on=["step1"])
workflow.add_task("step3", SimpleTask(result={"summary": "done"}), depends_on=["step2"])

# Execute
engine = WorkflowEngine()
result = engine.execute(workflow)

print(result.success)
print(result.duration)

Async Execution

import asyncio
from socratic_workflow import Workflow, SimpleTask, WorkflowEngine

async def main():
    workflow = Workflow("Async Pipeline")
    workflow.add_task("task1", SimpleTask(result={"value": 1}))
    workflow.add_task("task2", SimpleTask(result={"value": 2}))

    engine = WorkflowEngine()
    result = await engine.execute_async(workflow)

    print(f"Success: {result.success}")
    print(f"Duration: {result.duration:.2f}s")

asyncio.run(main())

Core Concepts

Workflow

A workflow defines a sequence of tasks with their dependencies. Use the builder pattern to create workflows:

workflow = Workflow("My Pipeline", description="Optional description")
workflow.add_task("task1", MyTask())
workflow.add_task("task2", MyTask(), depends_on=["task1"])
workflow.add_task("task3", MyTask(), depends_on=["task1"])  # Parallel with task2

Tasks

Create custom tasks by inheriting from Task:

from socratic_workflow import Task

class CustomTask(Task):
    def execute(self, context):
        # context contains results from previous tasks
        return {"result": "value"}

    async def execute_async(self, context):
        # Optional: override for true async execution
        return await self.execute(context)

WorkflowEngine

The engine orchestrates task execution:

engine = WorkflowEngine()

# Sync execution
result = engine.execute(workflow)

# Async execution
result = await engine.execute_async(workflow)

# Save/load state
engine.save_state("workflow.json")
engine.load_state("workflow.json")

Results

Get detailed execution results:

result = engine.execute(workflow)

print(result.success)          # bool
print(result.duration)         # float (seconds)
print(result.task_results)     # Dict[str, Any]
print(result.errors)           # Dict[str, str]
print(result.to_dict())        # Full result as dict

API Reference

Workflow

class Workflow:
    def __init__(self, name: str, description: str = "")
    def add_task(self, task_id: str, task: Task, depends_on: Optional[List[str]] = None) -> Workflow
    def get_task(self, task_id: str) -> Optional[Task]
    def list_tasks(self) -> List[str]
    def get_dependencies(self, task_id: str) -> List[str]
    def to_dict(self) -> Dict[str, Any]

Task

class Task(ABC):
    def __init__(self, name: Optional[str] = None, **kwargs)
    @abstractmethod
    def execute(self, context: Dict[str, Any]) -> Dict[str, Any]
    async def execute_async(self, context: Dict[str, Any]) -> Dict[str, Any]
    def to_dict(self) -> Dict[str, Any]

WorkflowEngine

class WorkflowEngine:
    def __init__(self, llm_client: Optional[Any] = None)
    def execute(self, workflow: Workflow) -> WorkflowResult
    async def execute_async(self, workflow: Workflow) -> WorkflowResult
    def save_state(self, path: str) -> None
    def load_state(self, path: str) -> None

WorkflowResult

class WorkflowResult:
    success: bool
    duration: float
    task_results: Dict[str, Any]
    errors: Dict[str, str]
    metrics: Dict[str, Any]
    def to_dict(self) -> Dict[str, Any]

Support Development

If you find this package useful, consider supporting development:

• Become a Sponsor - Get early access to new features
• Star on GitHub - Shows your appreciation
• Report Issues - Help improve the package

Your support helps fund development of the entire Socratic ecosystem.

Examples

Example 1: Basic Workflow

See examples/01_basic_workflow.py for a complete example.

Example 2: Cost Tracking

See examples/02_cost_tracking.py (Phase 2).

Example 3: Parallel Execution

See examples/03_parallel_execution.py (Phase 3).

Example 4: Error Recovery

See examples/04_error_recovery.py (Phase 3).

Advanced Features

Cost Tracking

Track LLM costs across multiple providers:

from socratic_workflow import CostTracker

tracker = CostTracker()

# Track individual calls
cost1 = tracker.track_call("claude-opus-4", 1000, 500)
cost2 = tracker.track_call("gpt-4", 1000, 500)

# Get summary
summary = tracker.get_summary()
print(f"Total cost: ${summary['total_cost_usd']:.4f}")
print(f"By provider: {summary['cost_by_provider']}")

# Get recommendations
recommendations = tracker.get_recommendations()
for rec in recommendations:
    print(f"- {rec}")

Supports 16+ models across 5 providers:

• Anthropic: Claude Opus-4, Sonnet-3.5, Haiku-4.5
• OpenAI: GPT-4, GPT-4o, GPT-3.5-turbo
• Google: Gemini 1.5 Pro, Gemini 1.5 Flash
• Meta: Llama-2-70b, Llama-3-70b
• Mistral: Mistral-7b, Mistral-Large

Parallel Execution

Run independent tasks concurrently:

from socratic_workflow.execution.executor import ParallelExecutor

executor = ParallelExecutor(max_workers=5)
result = await executor.execute_parallel(workflow)

metrics = executor.get_execution_metrics()
print(f"Success rate: {metrics['success_rate']:.1%}")

Error Recovery

Automatic retry with exponential backoff:

from socratic_workflow.execution.retry import RetryConfig, retry

config = RetryConfig(
    max_attempts=3,
    initial_delay=1.0,
    exponential_base=2.0,
    jitter=True
)

@retry(config)
def risky_operation():
    # Automatically retried on failure
    return result

Documentation

• Workflow Optimization Guide - Complete guide to the Quality Controller mechanism, path optimization, cost calculation, and risk assessment algorithms

Performance Metrics

Collect and analyze performance data:

from socratic_workflow.analytics import MetricsCollector

collector = MetricsCollector()
# ... execute workflow ...
summary = collector.get_summary()

print(f"Total duration: {summary['total_duration']:.2f}s")
print(f"Bottlenecks: {summary['bottlenecks']}")
print(f"Success rate: {summary['success_rate']:.1%}")

Integrations

LangChain

Integrate with LangChain agents:

import json
from socratic_workflow.integrations.langchain import SocraticWorkflowTool
from langchain.agents import Tool

tool = SocraticWorkflowTool(track_costs=True)

# Create LangChain tool
langchain_tool = Tool(
    name="execute_workflow",
    func=tool.execute_sync,
    description=tool.get_tool_description()
)

# Use in agent
from langchain.agents import initialize_agent
agent = initialize_agent([langchain_tool], llm, agent="zero-shot-react-description")

Openclaw

Create Openclaw skills:

from socratic_workflow.integrations.openclaw import SocraticWorkflowSkill

skill = SocraticWorkflowSkill(
    use_parallel_executor=True,
    track_costs=True,
    track_metrics=True
)

# Get capabilities for discovery
capabilities = skill.get_capabilities()

# Execute workflow
result = await skill.execute_workflow({
    "name": "My Workflow",
    "tasks": [
        {"id": "task1", "type": "SimpleTask", "result": {"status": "ok"}}
    ]
})

Roadmap

Phase 1: Core Infrastructure ✅

• Workflow definition and execution
• Task base class
• State management
• Sync and async execution

Phase 2: Cost Tracking & Analytics ✅

• Cost tracker with provider pricing (16+ models)
• Performance metrics collection
• Cost recommendations

Phase 3: Advanced Execution ✅

• Parallel task execution with asyncio
• Task scheduler with dependency resolution
• Retry logic with exponential backoff

Phase 4: Integrations ✅

• Openclaw skill integration
• LangChain tool integration
• Socratic Agents compatibility

Phase 5: Polish & Release ✅

• Comprehensive documentation
• 4 complete examples
• PyPI publication ready
• 95% test coverage (188 tests)
• All quality checks passing

Contributing

Contributions welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Add tests for new functionality
4. Ensure all tests pass
5. Submit a pull request

Testing

# Run tests
pytest

# With coverage
pytest --cov=src/socratic_workflow

# Specific test file
pytest tests/unit/test_engine.py -v

Quality

# Format code
black src/ tests/ examples/

# Lint
ruff check src/ tests/ examples/

# Type check
mypy src/socratic_workflow

License

MIT License - see LICENSE file for details.

Project Status

Version: 0.1.0 (MVP) ✅

Statistics

• Lines of Code: ~3,400
• Test Coverage: 95% (188 tests)
• Supported Python: 3.9, 3.10, 3.11, 3.12
• Quality: Black, Ruff, MyPy ✅
• CI/CD: GitHub Actions (automated testing and quality checks)

Implementation Progress

• ✅ Phase 1: Core Infrastructure (100%)
• ✅ Phase 2: Cost Tracking & Analytics (100%)
• ✅ Phase 3: Advanced Execution (100%)
• ✅ Phase 4: Integrations (100%)
• ✅ Phase 5: Polish & Release (100%)

Support

• 📖 Documentation
• 🐛 Issues
• 💬 Discussions

Related Projects

• Socratic Agents - Multi-agent orchestration
• Socratic RAG - Retrieval-augmented generation
• Socratic Analyzer - Code analysis
• Socrates Nexus - Universal LLM client

---

Built with ❤️ by the Socratic Ecosystem