profile_runner.py•2.17 kB
import asyncio
import cProfile
import os
import sys
# Add src to sys.path to allow importing from adaptive_graph_of_thoughts
sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "src")))
from adaptive_graph_of_thoughts.config import (
settings, # Use the global settings instance
)
from adaptive_graph_of_thoughts.domain.services.got_processor import GoTProcessor
async def main(): # Renamed back to main as per instruction
"""
Runs a profiling session by processing a sample query with GoTProcessor asynchronously.
Initializes the GoTProcessor with global settings, submits a predefined query about climate change and ocean acidification, and prints a truncated summary of the final answer from the session data.
"""
print("Initializing GoTProcessor...")
processor = GoTProcessor(settings=settings)
query = "What is the relationship between climate change and ocean acidification?"
operational_params = {
"include_reasoning_trace": False,
"include_graph_state": False,
}
initial_context = {}
session_id = "profile-session-001"
print(f"Processing query: '{query}' with session ID: {session_id}")
session_data = await processor.process_query(
query=query,
session_id=session_id,
operational_params=operational_params,
initial_context=initial_context,
)
print(
f"Query processing complete. Final answer (summary): {session_data.final_answer[:200]}..."
)
if __name__ == "__main__":
profiler_output_file = "output.prof"
print(f"Starting profiling, output will be saved to {profiler_output_file}")
profiler = cProfile.Profile()
profiler.enable()
asyncio.run(main()) # main() is now the workload
profiler.disable()
print(f"Profiling complete. Saving stats to {profiler_output_file}")
profiler.dump_stats(profiler_output_file)
print("\nTo analyze the profile data, use pstats. Example:")
print("import pstats; from pstats import SortKey")
print(f"stats = pstats.Stats('{profiler_output_file}')")
print("stats.strip_dirs().sort_stats(SortKey.CUMULATIVE).print_stats(20)")