qcentroid-agent-cli 0.3.20

Library to interact with qcentroid agent api

qcentroid-agent-cli

Client library to interact with QCentroid Agent API.

Functions

Functions:

• obtain status, and context
• obtain input data
• send output data
• set status
• send execution logs

Install

pip install qcentroid-agent-cli

Use

Simple example

As easy as this:

from qcentroid_agent_cli import QCentroidSolverClient
API_BASE_URL="https://api.qcentroid.xyz"
SOLVER_API_KEY="1234-4567-8910"  # Get your solver API_KEY in the platform dashboard
SOLVER_ID="123"

def main():
    
    # Get the solver details
    solver = QCentroidSolverClient(API_BASE_URL, SOLVER_API_KEY, SOLVER_ID)

    # Request a queued job
    job = solver.obtainJob()
    
    # Notify start of job execution
    job.start()
    
    # Retrieve the job input data
    input_data = job.obtainInputData()
    output_data = {} 

    #
    # TODO: Add your solver code here and generate output_data
    #

    # Send the solver output data and execution logs to the platform
    job.sendOutputData(output_data)
    job.sendExecutionLog(logs)

    # End of job execution
    
    
if __name__ == "__main__":
    main() 

Example for external agents:

import requests
from qcentroid_agent_cli import QCentroidSolverClient
API_BASE_URL="https://api.qcentroid.xyz"
SOLVER_API_KEY="1234-4567-8910"  # Get your solver API_KEY in the platform dashboard
SOLVER_ID="123"

def main():
    exit = False
    print("QCentroid Agent usage example")
    print("Starting...")
    
    # Initialize the agent and get the solver details and a valid access token
    solver = QCentroidSolverClient(API_BASE_URL, SOLVER_API_KEY, SOLVER_ID)

    print("Solver initialization successful.")

    # Loop to request queued jobs until any exit condition you want to set
    while not exit:
        try:
            print("Checking for pending jobs...")
            # Request a queued job (the oldest one will be returned)
            job = solver.obtainJob()

            if job :
                print("New job received.")
                # There is a job to be processed!
                try:
                    print("Processing job...")
                    # Notify the platform we're starting to process this job
                    job.start()
                    # Retrieve the input data
                    input_data = job.obtainInputData()
                    output_data = {} 
                    
                    #
                    # TODO: add your solver code here and generate output_data
                    #

                    print("Job processed successfully.")
                    # Send the solver output data to the platform
                    job.sendOutputData(output_data)
                    # Send the solver execution logs to check them thorugh the platform dashboard
                    # TODO: job.sendExecutionLog(logs)
                    
                    job.end()              
                except Exception as e:
                    # Job execution has failed, notify the platform about the error
                    print("Error during job execution.")
                    job.error(e)

            else:        
                # No queued jobs. Wait for 1 minute and check again
                print("No pending jobs. Waiting for 1 minute...")
                time.sleep(60)
            
        except requests.RequestException as e:
            # Error in an API request
            # Whether parameters are incorrect (URL, api-key or solver_id), or there are connectivity issues
            print(f"QCentroid Agent: API request failed: {e}")
            exit=True
            
        except Exception as e:
            # Any other errors
            print(f"QCentroid Agent error: {e}")
            exit=True
            
    print("End.")


if __name__ == "__main__":
    main()

Versioning

Update manually on main the pyproject.toml the version field to match the next release tag. Launch a new version on Releases section selecting a new tag matching the version. Create the release. The release will be published on pypi.org