The source code and dockerfile for the GSW2024 AI Lab.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
This repo is archived. You can view files and clone it, but cannot push or open issues/pull-requests.
 
 
 
 
 
 

535 lines
23 KiB

import sys
if sys.version_info[0] == 2:
raise ImportError('Python 2.x is not supported for stormpy.')
from . import core
from .core import *
from . import storage
from .storage import *
from ._config import *
from .logic import *
from .exceptions import *
from pycarl import Variable # needed for building parametric models
__version__ = "unknown"
try:
from ._version import __version__
except ImportError:
# We're running in a tree that doesn't have a _version.py, so we don't know what our version is.
pass
core._set_up("")
def _convert_sparse_model(model, parametric=False):
"""
Convert (parametric) model in sparse representation into model corresponding to exact model type.
:param model: Sparse model.
:param parametric: Flag indicating if the model is parametric.
:return: Model corresponding to exact model type.
"""
if parametric:
assert model.supports_parameters
if model.model_type == ModelType.DTMC:
return model._as_sparse_pdtmc()
elif model.model_type == ModelType.MDP:
return model._as_sparse_pmdp()
elif model.model_type == ModelType.POMDP:
return model._as_sparse_ppomdp()
elif model.model_type == ModelType.CTMC:
return model._as_sparse_pctmc()
elif model.model_type == ModelType.MA:
return model._as_sparse_pma()
else:
raise StormError("Not supported parametric model constructed")
else:
assert not model.supports_parameters
if model.model_type == ModelType.DTMC:
return model._as_sparse_dtmc()
elif model.model_type == ModelType.MDP:
return model._as_sparse_mdp()
elif model.model_type == ModelType.POMDP:
return model._as_sparse_pomdp()
elif model.model_type == ModelType.CTMC:
return model._as_sparse_ctmc()
elif model.model_type == ModelType.MA:
return model._as_sparse_ma()
else:
raise StormError("Not supported non-parametric model constructed")
def _convert_symbolic_model(model, parametric=False):
"""
Convert (parametric) model in symbolic representation into model corresponding to exact model type.
:param model: Symbolic model.
:param parametric: Flag indicating if the model is parametric.
:return: Model corresponding to exact model type.
"""
if parametric:
assert model.supports_parameters
if model.model_type == ModelType.DTMC:
return model._as_symbolic_pdtmc()
elif model.model_type == ModelType.MDP:
return model._as_symbolic_pmdp()
elif model.model_type == ModelType.CTMC:
return model._as_symbolic_pctmc()
elif model.model_type == ModelType.MA:
return model._as_symbolic_pma()
else:
raise StormError("Not supported parametric model constructed")
else:
assert not model.supports_parameters
if model.model_type == ModelType.DTMC:
return model._as_symbolic_dtmc()
elif model.model_type == ModelType.MDP:
return model._as_symbolic_mdp()
elif model.model_type == ModelType.CTMC:
return model._as_symbolic_ctmc()
elif model.model_type == ModelType.MA:
return model._as_symbolic_ma()
else:
raise StormError("Not supported non-parametric model constructed")
def build_model(symbolic_description, properties=None):
"""
Build a model in sparse representation from a symbolic description.
:param symbolic_description: Symbolic model description to translate into a model.
:param List[Property] properties: List of properties that should be preserved during the translation. If None, then all properties are preserved.
:return: Model in sparse representation.
"""
return build_sparse_model(symbolic_description, properties=properties)
def build_parametric_model(symbolic_description, properties=None):
"""
Build a parametric model in sparse representation from a symbolic description.
:param symbolic_description: Symbolic model description to translate into a model.
:param List[Property] properties: List of properties that should be preserved during the translation. If None, then all properties are preserved.
:return: Parametric model in sparse representation.
"""
return build_sparse_parametric_model(symbolic_description, properties=properties)
def build_sparse_model(symbolic_description, properties=None):
"""
Build a model in sparse representation from a symbolic description.
:param symbolic_description: Symbolic model description to translate into a model.
:param List[Property] properties: List of properties that should be preserved during the translation. If None, then all properties are preserved.
:return: Model in sparse representation.
"""
if not symbolic_description.undefined_constants_are_graph_preserving:
raise StormError("Program still contains undefined constants")
if properties:
formulae = [(prop.raw_formula if isinstance(prop, Property) else prop) for prop in properties]
intermediate = core._build_sparse_model_from_symbolic_description(symbolic_description, formulae)
else:
intermediate = core._build_sparse_model_from_symbolic_description(symbolic_description)
return _convert_sparse_model(intermediate, parametric=False)
def build_sparse_parametric_model(symbolic_description, properties=None):
"""
Build a parametric model in sparse representation from a symbolic description.
:param symbolic_description: Symbolic model description to translate into a model.
:param List[Property] properties: List of properties that should be preserved during the translation. If None, then all properties are preserved.
:return: Parametric model in sparse representation.
"""
if not symbolic_description.undefined_constants_are_graph_preserving:
raise StormError("Program still contains undefined constants")
if properties:
formulae = [(prop.raw_formula if isinstance(prop, Property) else prop) for prop in properties]
intermediate = core._build_sparse_parametric_model_from_symbolic_description(symbolic_description, formulae)
else:
intermediate = core._build_sparse_parametric_model_from_symbolic_description(symbolic_description)
return _convert_sparse_model(intermediate, parametric=True)
def build_symbolic_model(symbolic_description, properties=None):
"""
Build a model in symbolic representation from a symbolic description.
:param symbolic_description: Symbolic model description to translate into a model.
:param List[Property] properties: List of properties that should be preserved during the translation. If None, then all properties are preserved.
:return: Model in symbolic representation.
"""
if not symbolic_description.undefined_constants_are_graph_preserving:
raise StormError("Program still contains undefined constants")
if properties:
formulae = [(prop.raw_formula if isinstance(prop, Property) else prop) for prop in properties]
intermediate = core._build_symbolic_model_from_symbolic_description(symbolic_description, formulae)
else:
intermediate = core._build_symbolic_model_from_symbolic_description(symbolic_description)
return _convert_symbolic_model(intermediate, parametric=False)
def build_symbolic_parametric_model(symbolic_description, properties=None):
"""
Build a parametric model in symbolic representation from a symbolic description.
:param symbolic_description: Symbolic model description to translate into a model.
:param List[Property] properties: List of properties that should be preserved during the translation. If None, then all properties are preserved.
:return: Parametric model in symbolic representation.
"""
if not symbolic_description.undefined_constants_are_graph_preserving:
raise StormError("Program still contains undefined constants")
if properties:
formulae = [(prop.raw_formula if isinstance(prop, Property) else prop) for prop in properties]
intermediate = core._build_symbolic_parametric_model_from_symbolic_description(symbolic_description, formulae)
else:
intermediate = core._build_symbolic_parametric_model_from_symbolic_description(symbolic_description)
return _convert_symbolic_model(intermediate, parametric=True)
def build_model_from_drn(file, options = DirectEncodingParserOptions()):
"""
Build a model in sparse representation from the explicit DRN representation.
:param String file: DRN file containing the model.
:param DirectEncodingParserOptions: Options for the parser.
:return: Model in sparse representation.
"""
intermediate = core._build_sparse_model_from_drn(file, options)
return _convert_sparse_model(intermediate, parametric=False)
def build_parametric_model_from_drn(file, options = DirectEncodingParserOptions()):
"""
Build a parametric model in sparse representation from the explicit DRN representation.
:param String file: DRN file containing the model.
:param DirectEncodingParserOptions: Options for the parser.
:return: Parametric model in sparse representation.
"""
intermediate = core._build_sparse_parametric_model_from_drn(file, options)
return _convert_sparse_model(intermediate, parametric=True)
def perform_bisimulation(model, properties, bisimulation_type):
"""
Perform bisimulation on model.
:param model: Model.
:param properties: Properties to preserve during bisimulation.
:param bisimulation_type: Type of bisimulation (weak or strong).
:return: Model after bisimulation.
"""
return perform_sparse_bisimulation(model, properties, bisimulation_type)
def perform_sparse_bisimulation(model, properties, bisimulation_type):
"""
Perform bisimulation on model in sparse representation.
:param model: Model.
:param properties: Properties to preserve during bisimulation.
:param bisimulation_type: Type of bisimulation (weak or strong).
:return: Model after bisimulation.
"""
formulae = [(prop.raw_formula if isinstance(prop, Property) else prop) for prop in properties]
if model.supports_parameters:
return core._perform_parametric_bisimulation(model, formulae, bisimulation_type)
else:
return core._perform_bisimulation(model, formulae, bisimulation_type)
def perform_symbolic_bisimulation(model, properties):
"""
Perform bisimulation on model in symbolic representation.
:param model: Model.
:param properties: Properties to preserve during bisimulation.
:return: Model after bisimulation.
"""
formulae = [(prop.raw_formula if isinstance(prop, Property) else prop) for prop in properties]
bisimulation_type = BisimulationType.STRONG
if model.supports_parameters:
return core._perform_symbolic_parametric_bisimulation(model, formulae, bisimulation_type)
else:
return core._perform_symbolic_bisimulation(model, formulae, bisimulation_type)
def model_checking(model, property, only_initial_states=False, extract_scheduler=False, force_fully_observable=False, environment=Environment(), shield_expression=None):
"""
Perform model checking on model for property.
:param model: Model.
:param property: Property to check for.
:param only_initial_states: If True, only results for initial states are computed, otherwise for all states.
:param extract_scheduler: If True, try to extract a scheduler
:return: Model checking result.
:rtype: CheckResult
"""
if model.is_sparse_model:
return check_model_sparse(model, property, only_initial_states=only_initial_states,
extract_scheduler=extract_scheduler, force_fully_observable=force_fully_observable, environment=environment, shield_expression=shield_expression)
else:
assert (model.is_symbolic_model)
if extract_scheduler:
raise StormError("Model checking based on dd engine does not support extracting schedulers right now.")
return check_model_dd(model, property, only_initial_states=only_initial_states,
environment=environment)
def check_model_sparse(model, property, only_initial_states=False, extract_scheduler=False, force_fully_observable=False, environment=Environment(), shield_expression=None):
"""
Perform model checking on model for property.
:param model: Model.
:param property: Property to check for.
:param only_initial_states: If True, only results for initial states are computed, otherwise for all states.
:param extract_scheduler: If True, try to extract a scheduler
:param force_fully_observable: If True, treat a POMDP as an MDP
:return: Model checking result.
:rtype: CheckResult
"""
if isinstance(property, Property):
formula = property.raw_formula
else:
formula = property
if force_fully_observable:
if model.is_partially_observable:
# Note that casting a model to a fully observable model wont work with python/pybind, so we actually have other access points
if model.supports_parameters:
raise NotImplementedError("")
elif model.is_exact:
task = core.ExactCheckTask(formula, only_initial_states)
task.set_produce_schedulers(extract_scheduler)
return core._exact_model_checking_fully_observable(model, task, environment=environment)
else:
task = core.CheckTask(formula, only_initial_states)
task.set_produce_schedulers(extract_scheduler)
return core._model_checking_fully_observable(model, task, environment=environment)
else:
raise RuntimeError("Forcing models that are fully observable is not possible")
if model.supports_parameters:
task = core.ParametricCheckTask(formula, only_initial_states)
task.set_produce_schedulers(extract_scheduler)
return core._parametric_model_checking_sparse_engine(model, task, environment=environment)
else:
if model.is_exact:
task = core.ExactCheckTask(formula, only_initial_states)
task.set_produce_schedulers(extract_scheduler)
return core._exact_model_checking_sparse_engine(model, task, environment=environment)
else:
task = core.CheckTask(formula, only_initial_states)
task.set_produce_schedulers(extract_scheduler)
if shield_expression is not None:
task.set_shielding_expression(shield_expression)
return core._model_checking_sparse_engine(model, task, environment=environment)
def check_model_dd(model, property, only_initial_states=False, environment=Environment()):
"""
Perform model checking using dd engine.
:param model: Model.
:param property: Property to check for.
:param only_initial_states: If True, only results for initial states are computed, otherwise for all states.
:return: Model checking result.
:rtype: CheckResult
"""
if isinstance(property, Property):
formula = property.raw_formula
else:
formula = property
if model.supports_parameters:
task = core.ParametricCheckTask(formula, only_initial_states)
return core._parametric_model_checking_dd_engine(model, task, environment=environment)
else:
task = core.CheckTask(formula, only_initial_states)
return core._model_checking_dd_engine(model, task, environment=environment)
def check_model_hybrid(model, property, only_initial_states=False, environment=Environment()):
"""
Perform model checking using hybrid engine.
:param model: Model.
:param property: Property to check for.
:param only_initial_states: If True, only results for initial states are computed, otherwise for all states.
:return: Model checking result.
:rtype: CheckResult
"""
if isinstance(property, Property):
formula = property.raw_formula
else:
formula = property
if model.supports_parameters:
task = core.ParametricCheckTask(formula, only_initial_states)
return core._parametric_model_checking_hybrid_engine(model, task, environment=environment)
else:
task = core.CheckTask(formula, only_initial_states)
return core._model_checking_hybrid_engine(model, task, environment=environment)
def transform_to_sparse_model(model):
"""
Transform model in symbolic representation into model in sparse representation.
:param model: Symbolic model.
:return: Sparse model.
"""
if model.supports_parameters:
return core._transform_to_sparse_parametric_model(model)
else:
return core._transform_to_sparse_model(model)
def transform_to_discrete_time_model(model, properties):
"""
Transform continuous-time model to discrete time model.
:param model: Continuous-time model.
:param properties: List of properties to transform as well.
:return: Tuple (Discrete-time model, converted properties).
"""
formulae = [(prop.raw_formula if isinstance(prop, Property) else prop) for prop in properties]
if model.supports_parameters:
return core._transform_to_discrete_time_parametric_model(model, formulae)
else:
return core._transform_to_discrete_time_model(model, formulae)
def eliminate_non_markovian_chains(ma, properties, label_behavior):
"""
Eliminate chains of non-Markovian states if possible.
:param ma: Markov automaton.
:param properties: List of properties to transform as well.
:param label_behavior: Behavior of labels while elimination.
:return: Tuple (converted MA, converted properties).
"""
formulae = [(prop.raw_formula if isinstance(prop, Property) else prop) for prop in properties]
if ma.supports_parameters:
return core._eliminate_non_markovian_chains_parametric(ma, formulae, label_behavior)
else:
return core._eliminate_non_markovian_chains(ma, formulae, label_behavior)
def prob01min_states(model, eventually_formula):
assert type(eventually_formula) == logic.EventuallyFormula
labelform = eventually_formula.subformula
labelprop = core.Property("label-prop", labelform)
phiStates = BitVector(model.nr_states, True)
psiStates = model_checking(model, labelprop).get_truth_values()
return compute_prob01min_states(model, phiStates, psiStates)
def prob01max_states(model, eventually_formula):
assert type(eventually_formula) == logic.EventuallyFormula
labelform = eventually_formula.subformula
labelprop = core.Property("label-prop", labelform)
phiStates = BitVector(model.nr_states, True)
psiStates = model_checking(model, labelprop).get_truth_values()
return compute_prob01min_states(model, phiStates, psiStates)
def compute_prob01_states(model, phi_states, psi_states):
"""
Compute prob01 states for properties of the form phi_states until psi_states
:param SparseDTMC model:
:param BitVector phi_states:
:param BitVector psi_states: Target states
"""
if model.model_type != ModelType.DTMC:
raise StormError("Prob 01 is only defined for DTMCs -- model must be a DTMC")
if model.supports_parameters:
return core._compute_prob01states_rationalfunc(model, phi_states, psi_states)
else:
return core._compute_prob01states_double(model, phi_states, psi_states)
def compute_prob01min_states(model, phi_states, psi_states):
if model.model_type == ModelType.DTMC:
return compute_prob01_states(model, phi_states, psi_states)
if model.supports_parameters:
return core._compute_prob01states_min_rationalfunc(model, phi_states, psi_states)
else:
return core._compute_prob01states_min_double(model, phi_states, psi_states)
def compute_prob01max_states(model, phi_states, psi_states):
if model.model_type == ModelType.DTMC:
return compute_prob01_states(model, phi_states, psi_states)
if model.supports_parameters:
return core._compute_prob01states_max_rationalfunc(model, phi_states, psi_states)
else:
return core._compute_prob01states_max_double(model, phi_states, psi_states)
def topological_sort(model, forward=True, initial=[]):
"""
:param model:
:param forward:
:return:
"""
matrix = model.transition_matrix if forward else model.backward_transition_matrix
if isinstance(model, storage._SparseParametricModel):
return storage._topological_sort_rf(matrix, initial)
elif isinstance(model, storage._SparseModel):
return storage._topological_sort_double(matrix, initial)
else:
raise StormError("Unknown kind of model.")
def construct_submodel(model, states, actions, keep_unreachable_states=True, options=SubsystemBuilderOptions()):
"""
:param model: The model
:param states: Which states should be preserved
:param actions: Which actions should be preserved
:param keep_unreachable_states: If False, run a reachability analysis.
:return: A model with fewer states/actions
"""
if isinstance(model, storage._SparseModel):
return core._construct_subsystem_double(model, states, actions, keep_unreachable_states, options)
else:
raise NotImplementedError()
def parse_properties(properties, context=None, filters=None):
"""
:param properties: A string with the pctl properties
:param context: A symbolic model that gives meaning to variables and constants.
:param filters: filters, if applicable.
:return: A list of properties
"""
if context is None:
return core.parse_properties_without_context(properties, filters)
elif type(context) == core.SymbolicModelDescription:
if context.is_prism_program():
return core.parse_properties_for_prism_program(properties, context.as_prism_program(), filters)
else:
core.parse_properties_for_jani_program(properties, context.as_jani_model(), filters)
elif type(context) == storage.PrismProgram:
return core.parse_properties_for_prism_program(properties, context, filters)
elif type(context) == storage.JaniModel:
core.parse_properties_for_jani_model(properties, context, filters)
else:
raise StormError("Unclear context. Please pass a symbolic model description")
def export_to_drn(model, file, options=DirectEncodingOptions()):
"""
Export a model to DRN format
:param model: The model
:param file: A path
:param options: DirectEncodingOptions
:return:
"""
if model.supports_parameters:
return core._export_parametric_to_drn(model, file, options)
if model.is_exact:
return core._export_exact_to_drn(model, file, options)
return core._export_to_drn(model, file, options)