tests/utils/debug_workflow.py

"""
Debug script to test patent workflow execution
"""
import asyncio
import sys
from pathlib import Path
from loguru import logger

# Configure logger for debugging
logger.remove()
logger.add(sys.stdout, level="DEBUG")

async def test_document_analysis():
    """Test just the document analysis step"""
    from src.llm.langchain_ollama_client import LangChainOllamaClient
    from src.agents.scenario1 import DocumentAnalysisAgent

    logger.info("=" * 70)
    logger.info("Testing Document Analysis Agent")
    logger.info("=" * 70)

    # Initialize LLM client
    logger.info("Initializing LLM client...")
    llm_client = LangChainOllamaClient()

    # Initialize agent
    logger.info("Initializing DocumentAnalysisAgent...")
    agent = DocumentAnalysisAgent(llm_client=llm_client, memory_agent=None)

    # Test with a patent file
    patent_path = "/home/mhamdan/SPARKNET/Dataset/Microsoft July 2006.pdf"
    logger.info(f"Testing with patent: {patent_path}")

    try:
        logger.info("Starting patent analysis...")
        analysis = await agent.analyze_patent(patent_path)

        logger.success("✅ Analysis completed!")
        logger.info(f"Patent Title: {analysis.title}")
        logger.info(f"TRL Level: {analysis.trl_level}")
        logger.info(f"Key Innovations: {len(analysis.key_innovations)}")
        logger.info(f"Technical Domains: {analysis.technical_domains}")

        return analysis

    except Exception as e:
        logger.error(f"❌ Analysis failed: {e}")
        import traceback
        traceback.print_exc()
        return None

async def test_full_workflow():
    """Test the full workflow"""
    from src.llm.langchain_ollama_client import LangChainOllamaClient
    from src.agents.planner_agent import create_planner_agent
    from src.agents.critic_agent import create_critic_agent
    from src.agents.memory_agent import create_memory_agent
    from src.workflow.langgraph_workflow import create_workflow
    from src.workflow.langgraph_state import ScenarioType

    logger.info("=" * 70)
    logger.info("Testing Full Workflow")
    logger.info("=" * 70)

    # Initialize components
    logger.info("Initializing LLM client...")
    llm_client = LangChainOllamaClient()

    logger.info("Initializing agents...")
    planner = create_planner_agent(llm_client)
    critic = create_critic_agent(llm_client)
    memory = create_memory_agent(llm_client)

    logger.info("Creating workflow...")
    workflow = create_workflow(
        llm_client=llm_client,
        planner_agent=planner,
        critic_agent=critic,
        memory_agent=memory
    )

    # Test workflow
    patent_path = "/home/mhamdan/SPARKNET/Dataset/Microsoft July 2006.pdf"
    logger.info(f"Testing workflow with patent: {patent_path}")

    try:
        result = await workflow.run(
            task_description=f"Analyze patent: Microsoft July 2006.pdf and create valorization roadmap",
            scenario=ScenarioType.PATENT_WAKEUP,
            input_data={"patent_path": patent_path},
            task_id="debug_test_001"
        )

        logger.success("✅ Workflow completed!")
        logger.info(f"Success: {result.success}")
        logger.info(f"Quality Score: {result.quality_score}")
        logger.info(f"Iterations: {result.iterations_used}")
        logger.info(f"Execution Time: {result.execution_time_seconds:.2f}s")

        return result

    except Exception as e:
        logger.error(f"❌ Workflow failed: {e}")
        import traceback
        traceback.print_exc()
        return None

if __name__ == "__main__":
    import argparse

    parser = argparse.ArgumentParser(description="Debug SPARKNET workflow")
    parser.add_argument("--test", choices=["doc", "full"], default="doc",
                       help="Test to run: 'doc' for document analysis only, 'full' for full workflow")
    args = parser.parse_args()

    if args.test == "doc":
        asyncio.run(test_document_analysis())
    else:
        asyncio.run(test_full_workflow())