schrodinger.stepper.stepper module

Framework for writing computational workflows and running them in a highly distributed manner. Each step of the workflow is either a “mapping” operation (see MapStep) or “reducing” operation (see ReduceStep). These steps can then be chained together using the `Chain class.

For a more complete introduction, see WordCount tutorial:

For documentation on specific stepper features, see the following feature list. You can ctrl+f the feature tag to jump to the relevant docstrings.

Feature

Tag

MapStep

_map_step_

ReduceStep

_reduce_step_

Chain

_chain_

Settings

_settings_

Serialization

_serialization_

File Handling

_file_handling_

Generic Steps

_generics_

Custom Workflow

_custom_workflows_

Double Batching

_dbl_batching_

#=============================================================================== # Running stepper with custom, undistributed workflows <_custom_workflows_> #=============================================================================== To run steps that aren’t defined in the core suite: The script should be executed inside the working directory and import steps from a local package in the working directory.

Working dir contents:

script.py
my_lib/
    __init__.py
    steps.py

Minimal code in script.py if it needs to run under job control:

from schrodinger.job import launchapi
from schrodinger.ui.qt.appframework2 import application
from my_lib.steps import MyStep

def get_job_spec_from_args(argv):
    jsb = launchapi.JobSpecificationArgsBuilder(argv)
    jsb.setInputFile(__file__)
    jsb.setInputDirectory('my_lib')
    return jsb.getJobSpec()

def main():
    step = MyStep()
    set.getOutputs()

if __name__ == '__main__':
    application.run_application(main)

#=============================================================================== # Generic Steps <_generics_> #=============================================================================== Steps that can be run on any data type can be created by using the GENERIC sentinel type as the input or output type of a step. The step will then receive all inputs as serialized strings and should return all string outputs.

Example:

>>> class DeduplicateStep(ReduceStep):
...     Input = Output = GENERIC
...
...     def reduceFunction(self, inps):
...         yield from sorted(set(inps))
...
>>> class IdentityStep(MapStep):
...     Input = Output = int
...
...     def mapFunction(self, value):
...         yield value
...
>>> class DeduplicatedIdentityChain(Chain):
...     def buildChain(self):
...         self.addStep(IdentityStep())
...         self.addStep(DeduplicateStep())
...         self.addStep(IdentityStep())
...
>>> chain = DeduplicatedIdentityChain()
>>> chain.setInputs([1, 2, 3, 1, 2, 3])
>>> chain.getOutputs()
[1, 2, 3]

These steps can be chained together with any steps, regardless of the input or output types of the other steps.

NOTE::

No validation is done to guarantee that the input type of the step after a generic step is the same as the output type of the precedeing step. The behavior for this circumstance is undefined and will likely resulted in an error.

#=============================================================================== # Double Batching <_dbl_batching_> #=============================================================================== Job launch speeds at the time of writing is about one job per 3 or 4 seconds. This rate becomes insufficient once we need more than a few hundred workers. To get around this, stepper employs a pattern we coin “double batching”, where we create subjobs whose sole purpose is to themselves create the subjobs that actually run the steps.

NOTE:: We use double-batching for the PubSub implementation of stepper as well as the file-based implementation. The literal meaning of “double-batching” doesn’t apply as well to the PubSub implementation but the general pattern of launching subjobs to launch more subjobs still applies.

#=============================================================================== # Environment variables and global settings #===============================================================================

Settings:

  • SCHRODINGER_STEPPER_DEBUG

    Set to 1 to have most files brought back from a workflow run. Set to 2 to have _all_ files brought back.

  • SCHRODINGER_STEPPER_LOG_LEVEL

    Set to an integer or name representing the log level to use for the logger. Valid log names are in decreasing order of verbosity NOTSET, DEBUG, INFO, WARN, ERROR, FATAL, and CRITICAL. Set to DEBUG to see log messages from cloud clients.

  • SCHRODINGER_GCP_KEY

    Expected when running stepper with pubsub and bigquery. Should be a path to the gcp service key. See

    for more information on generating gcp service keys.

  • SCHRODINGER_CLOUD_VENV

    A path to the python executable of a virtualenv that can be used for interfacing with GCP or AWS. This virtualenv should be accessible on all nodes that a stepper workflow will be run on in addition to having all necessary python client libraries.

  • SCHRODINGER_GCP_ENABLED

    Optional setting to signify that the current machine is a GCP compute engine. Setting this will make SCHRODINGER_GCP_KEY also optional as credentials will be inferred from the GCP compute engine. SCHRODINGER_GCP_KEY will have precedence over the compute engine’s default credentials if set.

  • SCHRODINGER_CLOUD_WORKER_TIMEOUT

    Optional debug setting. If set, pubsub workers will timeout after SCHRODINGER_CLOUD_WORKER_TIMEOUT minutes.

  • SCHRODINGER_GCP_NUM_PUBSUB_WORKERS

    Sets the default number of pubsub workers that will be used. If not set, one will be used. Note that this value can still be overridden by a workflow’s configuration.

  • SCHRODINGER_STEPPER_ERROR_THRESHOLD

    Sets the batching error threshold. If not set, 0.9 will be used. Note that this value can still be overridden by a workflow’s configuration.

schrodinger.stepper.stepper.get_debug_level()
class schrodinger.stepper.stepper.ElapsedFormatter

Bases: object

A stepper logging formatter that includes how much time has elapsed since a start time in all messages. FORMATTER.start() is safe to call multiple times, the start time will be set to the time of the earliest call.

Example usage:

logger.debug("Message before debug.")
# Message before debug.
FORMATTER.start()
time.wait(1)
logger.debug("My message")
# <STEPPER> DEBUG +00:00:01: My message
__init__()
start()
format(record)
schrodinger.stepper.stepper.ichunked(iterable, n)

Reimplementation of more_itertools.ichunked that does not cache n items of iterable at a time.

Breaks iterable into sub-iterables with n elements each.

Note that unlike more_itertools.ichunked, an error will be raised if you try to iterate over a chunk before its previous chunk has been consumed.

class schrodinger.stepper.stepper.StepperFile(value='', *args, **kwargs)

Bases: schrodinger.tasks.tasks.TaskFile

class schrodinger.stepper.stepper.StepperFolder(value='', *args, **kwargs)

Bases: schrodinger.tasks.tasks.TaskFolder

class schrodinger.stepper.stepper.StepTaskInput(*args, _param_type=<object object>, **kwargs)

Bases: schrodinger.models.parameters.CompoundParam

dehydrated_step: schrodinger.stepper.stepper._DehydratedStep

See _BaseStep._dehydrateStep for documentation.

debug_mode: bool

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
misc_input_filenames: List[schrodinger.tasks.tasks.TaskFile]

A Param to represent lists. Values of this param will have a mutated signal that will be emitted whenever any mutation method is called.

The constructor optionally takes a item_class keyword argument to specify what type of class the items in the list will be. This information will be used for jsonifying the list if specified.

misc_input_folders: List[schrodinger.tasks.tasks.TaskFolder]

A Param to represent lists. Values of this param will have a mutated signal that will be emitted whenever any mutation method is called.

The constructor optionally takes a item_class keyword argument to specify what type of class the items in the list will be. This information will be used for jsonifying the list if specified.

debug_modeChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

debug_modeReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

dehydrated_stepChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

dehydrated_stepReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

misc_input_filenamesChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

misc_input_filenamesReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

misc_input_foldersChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

misc_input_foldersReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

class schrodinger.stepper.stepper.StepTaskOutput(*args, _param_type=<object object>, **kwargs)

Bases: schrodinger.models.parameters.CompoundParam

output_file: schrodinger.tasks.tasks.TaskFile

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
run_info: dict

A Param to represent dictionaries. Values of this param will have a mutated signal that will be emitted whenever any mutation method is called.

The constructor optionally takes a value_class keyword argument to specify what type of class the values will be. This information will be used for jsonifying the dictionary if specified. (Note that non-string keys are not currently supported for jsonification. This may change in the future. See PANEL-13029).

misc_output_filenames: List[schrodinger.tasks.tasks.TaskFile]

A Param to represent lists. Values of this param will have a mutated signal that will be emitted whenever any mutation method is called.

The constructor optionally takes a item_class keyword argument to specify what type of class the items in the list will be. This information will be used for jsonifying the list if specified.

misc_output_filenamesChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

misc_output_filenamesReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

output_fileChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

output_fileReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

run_infoChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

run_infoReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

class schrodinger.stepper.stepper.StepTaskMixin(*args, step=None, **kwargs)

Bases: schrodinger.models.parameters.CompoundParamMixin

This class must be mixed in with a subclass of AbstractComboTask. The resulting task class may be used to run any step as a task, provided the input, output, and settings classes are all JSONable.

input: schrodinger.stepper.stepper.StepTaskInput
output: schrodinger.stepper.stepper.StepTaskOutput
DEFAULT_TASKDIR_SETTING = 1
__init__(*args, step=None, **kwargs)
addLicenseReservation(license, num_tokens=1)
setStep(step)
getStepClass()
mainFunction()
class schrodinger.stepper.stepper.StepSubprocessTask(*args, _param_type=<object object>, **kwargs)

Bases: schrodinger.stepper.stepper.StepTaskMixin, schrodinger.tasks.tasks.ComboSubprocessTask

calling_contextChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

calling_contextReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

failure_infoChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

failure_infoReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

input: StepTaskInput
inputChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

inputReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

max_progressChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

max_progressReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

nameChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

nameReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

output: StepTaskOutput
outputChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

outputReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

progressChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

progressReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

progress_stringChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

progress_stringReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

statusChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

statusReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

class schrodinger.stepper.stepper.StepJobTask(*args, _param_type=<object object>, **kwargs)

Bases: schrodinger.stepper.stepper.StepTaskMixin, schrodinger.tasks.jobtasks.ComboJobTask

input: schrodinger.stepper.stepper.StepTaskInput
output: schrodinger.stepper.stepper.StepTaskOutput
isCanceled() bool

Whether the job underlying the task is canceled.

If there is no job associated with the task, then this returns False.

mainFunction()
calling_contextChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

calling_contextReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

failure_infoChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

failure_infoReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

inputChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

inputReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

job_configChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

job_configReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

max_progressChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

max_progressReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

nameChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

nameReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

outputChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

outputReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

progressChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

progressReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

progress_stringChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

progress_stringReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

statusChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

statusReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

class schrodinger.stepper.stepper.Topic(*args, _param_type=<object object>, **kwargs)

Bases: schrodinger.models.parameters.CompoundParam

An abstraction over a container of messages in a Publish/Subscribe model

Variables
  • name – the name of the topic

  • num_uploaded_msgs – the number of messages that were uploaded to this topic; it is not reflective of the number of messages currently in the topic.

name: str

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
num_uploaded_msgs: int

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
nameChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

nameReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

num_uploaded_msgsChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

num_uploaded_msgsReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

class schrodinger.stepper.stepper.BatchSettings(*args, _param_type=<object object>, **kwargs)

Bases: schrodinger.models.parameters.CompoundParam

size: int

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
task_class: type

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
hostname: str

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
use_pubsub: bool

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
num_pubsub_workers: int

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
error_threshold: float

Base class for all Param classes. A Param is a descriptor for storing data, which means that a single Param instance will manage the data values for multiple instances of the class that owns it. Example:

class Coord(CompoundParam):
    x: int
    y: int

An instance of the Coord class can be created normally, and Params can be accessed as normal attributes:

coord = Coord()
coord.x = 4

When a Param value is set, the valueChanged signal is emitted. Params can be serialized and deserialized to and from JSON. Params can also be nested:

class Atom(CompoundParam):
    coord: Coord
    element: str
initializeValue()

Override to dynamically set up the default value of the param. Useful for default values that are determined at runtime. This is called any time the param is reset.

error_thresholdChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

error_thresholdReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

hostnameChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

hostnameReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

num_pubsub_workersChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

num_pubsub_workersReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

sizeChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

sizeReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

task_classChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

task_classReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

use_pubsubChanged

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

use_pubsubReplaced

pyqtSignal(*types, name: str = …, revision: int = …, arguments: Sequence = …) -> PYQT_SIGNAL

types is normally a sequence of individual types. Each type is either a type object or a string that is the name of a C++ type. Alternatively each type could itself be a sequence of types each describing a different overloaded signal. name is the optional C++ name of the signal. If it is not specified then the name of the class attribute that is bound to the signal is used. revision is the optional revision of the signal that is exported to QML. If it is not specified then 0 is used. arguments is the optional sequence of the names of the signal’s arguments.

class schrodinger.stepper.stepper.Serializer

Bases: object

<_serialization_> A class for defining special serialization for some datatype. Serialization by default uses the json protocol, but if a specialized protocol is wanted instead, users can subclass this class to do so.

Subclasses should:

  • Define DataType. This is the class that this serializer can

    encode/decode.

  • Define toString(self, output), which defines how to serialize

    an output.

  • Define fromString(self, input_str), which defines how to

    deserialize an input.

This can then be used as the InputSerializer or OutputSerializer for any step.

Here’s an example for defining an int that’s serialized in base-two as opposed to base-ten:

class IntBaseTwoSerializer(Serializer):
    DataType = int

    def toString(self, output):
        return bin(output) # 7 -> '0b111'

    def fromString(self, input_str):
        return int(input_str[2:], 2) # '0b111' -> 7

This can then be used anywhere you’d use an int as the output or input in a step. For example:

class SquaringStep(MapStep):
    Input = int
    InputSerializer = IntBaseTwoSerializer
    Output = int
    OutputSerializer = IntBaseTwoSerializer

    def mapFunction(self, inp):
        yield inp**2

Now, any time that a SquaringStep would read its inputs from a file or write its outputs to a file, it’ll do so using using a base-two representation.

DataType = NotImplemented
serialize(items, fname)

Write items to a file named fname.

deserialize(fname)

Read in items from fname. :type fname: str :rtype: iterable[self.DataType]

fromString(input_str)
toString(output)
exception schrodinger.stepper.stepper.ValidationIssue(source_step, msg)

Bases: RuntimeError

__init__(source_step, msg)
exception schrodinger.stepper.stepper.SettingsError(source_step, msg)

Bases: schrodinger.stepper.stepper.ValidationIssue

Used in conjunction with _BaseStep.validateSettings to report an error with settings. Constructed with the step with the invalid settings and an error message, e.g. SettingsError(bad_step, "Step does not have required settings.")

exception schrodinger.stepper.stepper.SettingsWarning(source_step, msg)

Bases: schrodinger.stepper.stepper.ValidationIssue

Used in conjunction with _BaseStep.validateSettings to report a warning with settings. Constructed with the step with the invalid settings and an error message, e.g. SettingsWarning(bad_step, "Step setting FOO should ideally be positive")

exception schrodinger.stepper.stepper.ResourceError(source_step, msg)

Bases: schrodinger.stepper.stepper.ValidationIssue

Used in conjunction with _BaseStep.validateSettings to report an error with a resource setting. Constructed with the step with the invalid setting and an error message, e.g., ResourceError(bad_step, "Step setting 'file' has not been set.")

exception schrodinger.stepper.stepper.LocalResourceError(source_step, msg)

Bases: schrodinger.stepper.stepper.ResourceError

A ResourceError specifically for local StepperFile and StepperFolder validations, i.e., resources that are on a job submission host and may have to be transferred to compute resources

exception schrodinger.stepper.stepper.StaticResourceError(source_step, msg)

Bases: schrodinger.stepper.stepper.ResourceError

A ResourceError specifically for static StepperFile and StepperFolder validations, i.e., resources that are not necessarily available on a job submission host

class schrodinger.stepper.stepper.PubsubEnabledStepMixin(*args, **kwargs)

Bases: object

A mixin that allows a step to be run using PubSub.

Steps with this mixin will have batch settings that have a use_pubsub flag and a num_pubsub_workers integer. Flipping use_pubsub to on will have the step load up all its inputs into a pubsub topic before spinning up num_pubsub_workers subjobs that will all take from the input topic, run the step’s computation on it, and upload it to an output topic.

Calling my_pubsub_step.getOutputs() will return all the outputs from the output topic, so to a user this will all be implementation detail.

__init__(*args, **kwargs)
property topic_prefix
property topic_suffix
outputs(*args, **kwargs)
usingPubsub()
getInputTopic() Optional[schrodinger.stepper.stepper.Topic]
setInputTopic(inp_topic: Optional[schrodinger.stepper.stepper.Topic])
getOutputTopic() Optional[schrodinger.stepper.stepper.Topic]
setOutputTopic(outp_topic: Optional[schrodinger.stepper.stepper.Topic])
initializeTopics()
class schrodinger.stepper.stepper.UnbatchedReduceStep(settings=None, config=None, step_id=None, metrics_logger_depth=None, _run_info=None, **kwargs)

Bases: schrodinger.stepper.stepper._BaseStep

” An unbatchable ReduceStep. See ReduceStep for more information.

reduceFunction(inputs)
class schrodinger.stepper.stepper.ReduceStep(*args, **kwargs)

Bases: schrodinger.stepper.stepper._BatchableStepMixin, schrodinger.stepper.stepper.UnbatchedReduceStep

<_reduce_step_> A computational step that performs a function on a collection of inputs to produce output items.

To construct a ReduceStep:

  • Implement reduceFunction

  • Define Input (the type expected by the mapFunction)

  • Define Output (the type of item produced by the mapFunction)

  • Define Settings (data class for any settings needed by the mapFunction)

reduceFunction(inputs)

The main computation for this step. This function should take in a iterable of inputs and return an iterable of outputs.

Example:

def reduceFunction(self, words):
    # Find all unique words
    seen_words = set()
    for word in words:
        if word not in seen_words:
            seen_words.add(word)
            yield word
class schrodinger.stepper.stepper.UnbatchedMapStep(settings=None, config=None, step_id=None, metrics_logger_depth=None, _run_info=None, **kwargs)

Bases: schrodinger.stepper.stepper.UnbatchedReduceStep

<_unbatchability_> An unbatchable MapStep. See MapStep for more information.

reduceFunction(inputs)
mapFunction(input)
class schrodinger.stepper.stepper.MapStep(*args, **kwargs)

Bases: schrodinger.stepper.stepper._BatchableStepMixin, schrodinger.stepper.stepper.UnbatchedMapStep

<_map_step_> A computational step that performs a function on input items from an input source to produce output items.

To construct a MapStep:

  • Implement mapFunction

  • Define Input (the type expected by the mapFunction)

  • Optionally define a InputSerializer (see Serializer for more info.)

  • Define Output (the type of item produced by the mapFunction)

  • Optionally define a OutputSerializer (see Serializer for more info.)

  • Define Settings (data class for any settings needed by the mapFunction)

mapFunction(input)

The main computation for this step. This function should take in a single input item and return an iterable of outputs. This allows a single output to produce multiple ouputs (e.g. enumeration).

The output may be yielded as a generator, in order to reduce memory usage.

If only a single output is produced for each input, return it as a single-element list.

Parameters

input

this will be a single input item from the input source. Implementer is encouraged to use a more descriptive, context- specific variable name. Example:

def mapFunction(self, starting_smiles):

class schrodinger.stepper.stepper.UnbatchedChain(*args, **kwargs)

Bases: schrodinger.stepper.stepper.UnbatchedReduceStep

property Input
property Output
property InputSerializer

The default serializer that simply uses json.loads and json.dumps

property OutputSerializer

The default serializer that simply uses json.loads and json.dumps

setInputs(inputs: Iterable[Any], starting_step_id: str = None)

Set the inputs for the chain. If starting_step_id is specified, then all steps before the specified starting step will be skipped. This is useful for resuming a chain’s computation.

setInputFile(input_file: str, starting_step_id: str = None)

Set the input file for the chain. If starting_step_id is specified, then all steps before the specified starting step will be skipped. This is useful for resuming a chain’s computation.

setStartingStep(starting_step: str)
validateSettings()

Check whether the chain settings are valid and return a list of SettingsError and SettingsWarning to report any invalid settings. Default implementation simply returns problems from all child steps.

Return type

list[TaskError or TaskWarning]

getResources(param_type, resource_type)

Get the stepper resources in the settings for the chain as well as for every step in the chain that are instances of param_type and have a resource_type attribute that is resource_type.

Note does not work for list/set/tuple subparams in the settings.

Parameters
  • param_type (tasks._TaskResource) – the resource parameter type

  • resource_type (ResourceType) – the type of resource to get

Returns

the set of stepper resources of resource_type

Return type

set of tasks._TaskResource

__init__(*args, **kwargs)

See class docstring for info on the different constructor arguments.

__len__()
addStep(step)
report(prefix='')

Report the workflow steps and their settings (recursively).

Parameters

prefix (str) – the text to start each line with

validateChain()

Checks that the declaration of the chain is internally consistent - i.e. that each step is valid and each step’s Input class matches the preceding step’s Output class.

reduceFunction(inputs)
buildChain()

This method must be implemented by subclasses to build the chain. The chain is built by modifying self.steps. The chain’s composition may be dependent on self.settings.

class schrodinger.stepper.stepper.Chain(*args, **kwargs)

Bases: schrodinger.stepper.stepper._BatchableStepMixin, schrodinger.stepper.stepper.UnbatchedChain

<_chain_> Run a series of steps. The steps must be created by overriding buildChain.

getLicenseRequirements()
schrodinger.stepper.stepper.get_all_steps_and_chains(step: schrodinger.stepper.stepper._BaseStep) Set[schrodinger.stepper.stepper._BaseStep]

Given a step, return a set of all steps it contains and itself. For example, given a chain A with the following topology:

    A
|-------|
B       C
    |-------|
    D       E

this method will return:

A -> set([A, B, C, D, E])
B -> set[B])
C -> set([C, D, E])