schrodinger.active_learning.mq_worker module

class schrodinger.active_learning.mq_worker.LigandMLBackend(url, model, descending)

Bases: object

__init__(url, model, descending)

start()

isRunning()

Returns

True if the backend is running

Return type

bool

getCrashLog()

kill()

class schrodinger.active_learning.mq_worker.Worker(options)

Bases: object

__init__(options)

run()

mainLoop()

checkDeadlines()

sendToDriver(msg)

handleGET(msg, address)

handlePING(msg, address)

handlePONG(msg, address)

handleSTOP(msg, address)

handleDONE(msg, address)

Handle DONE message from backend. If no predictions were generated, sends empty results to driver.

handleREADY(msg, address)

handleCSV(msg, address)

Receives inputs from driver, writes out, sends path to backend

writeCSVFile(filepath, msg)

writeModelFiles(files)

recv()

Listen for a message, if any, from the driver or backend. :return: message and address. Only messages from the backend socket have an address, messages from other sources will have address None. If there are no messages, return (None, None) :rtype: tuple

startBackend()

Launch backend as a subprocess on the same host as the worker.

sendToBackend(msg, address=None)

Send a message through the backend socket.

stopBackend()

Stop the backend, by sending it a STOP message. (The message is only likely to be received if the worker is idle, but that’s the normal case at the end of a job.)

cleanUp()

getInputs()

Tries 10 times to request models from driver, or from peer if driver is not available.

initSockets()

schrodinger.active_learning.mq_worker.my_uuid()

Return a base64-encoded random UUID to use as worker ID or ping ID.

schrodinger.active_learning.mq_worker.get_backend_args()

schrodinger.active_learning.mq_worker.run_dj(args)

Launch a number of workers using JobDJ. Used when -NJOBS > 1 on the command line.

Parameters

args (Argparse.Namespace) – command-line arguments

schrodinger.active_learning.mq_worker.submit_jobdj(args, subjobs)

schrodinger.active_learning.mq_worker.combine_log_files(subjobs, logger)

Concatenates the contents of subjob log files.

Parameters

• subjobs (list(str)) – subjob names

• logger (Logger) – logger to which concatenated log files are to be written

schrodinger.active_learning.mq_worker.parse_args(argv=None)

schrodinger.active_learning.mq_worker.sigterm_handler(signum, frame)

schrodinger.active_learning.mq_worker.main()

schrodinger.active_learning.mq_worker.get_job_spec_from_args(argv: argparse.Namespace) → schrodinger.job.launchapi.JobSpecificationArgsBuilder

Parse command-line arguments, excluding the first three elements:

sys.argv[0] # “run” sys.argv[1] # “python3” sys.argv[2] # “-m”