|
|
@ -1,7 +1,7 @@ |
|
|
|
import stormpy |
|
|
|
import stormpy.logic |
|
|
|
from helpers.helper import get_example_path |
|
|
|
import pytest |
|
|
|
from configurations import numpy_avail |
|
|
|
|
|
|
|
|
|
|
|
class TestSparseModelComponents: |
|
|
@ -25,7 +25,7 @@ class TestSparseModelComponents: |
|
|
|
nr_states = 13 |
|
|
|
nr_choices = 13 |
|
|
|
|
|
|
|
# TransitionMatrix |
|
|
|
# transition_matrix |
|
|
|
builder = stormpy.SparseMatrixBuilder(rows=0, columns=0, entries=0, force_dimensions=False, |
|
|
|
has_custom_row_grouping=False, row_groups=0) |
|
|
|
|
|
|
@ -49,7 +49,7 @@ class TestSparseModelComponents: |
|
|
|
# Build transition matrix, update number of rows and columns |
|
|
|
transition_matrix = builder.build(nr_states, nr_states) |
|
|
|
|
|
|
|
# StateLabeling |
|
|
|
# state_labeling |
|
|
|
state_labeling = stormpy.storage.StateLabeling(nr_states) |
|
|
|
state_labels = {'init', 'one', 'two', 'three', 'four', 'five', 'six', 'done', 'deadlock'} |
|
|
|
for label in state_labels: |
|
|
@ -62,16 +62,16 @@ class TestSparseModelComponents: |
|
|
|
state_labeling.add_label_to_state('four', 10) |
|
|
|
state_labeling.add_label_to_state('five', 11) |
|
|
|
state_labeling.add_label_to_state('six', 12) |
|
|
|
# Set the labeling of states given in a bit vector, where length = nr_states |
|
|
|
|
|
|
|
state_labeling.set_states('done', stormpy.BitVector(nr_states, [7, 8, 9, 10, 11, 12])) |
|
|
|
|
|
|
|
# RewardModels |
|
|
|
# reward_models |
|
|
|
reward_models = {} |
|
|
|
# Create a vector representing the state-action rewards |
|
|
|
action_reward = [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0] |
|
|
|
reward_models['coin_flips'] = stormpy.SparseRewardModel(optional_state_action_reward_vector=action_reward) |
|
|
|
|
|
|
|
# StateValuations |
|
|
|
# state_valuations |
|
|
|
manager = stormpy.ExpressionManager() |
|
|
|
var_s = manager.create_integer_variable(name='s') |
|
|
|
var_d = manager.create_integer_variable(name='d') |
|
|
@ -87,29 +87,29 @@ class TestSparseModelComponents: |
|
|
|
|
|
|
|
state_valuations = v_builder.build(13) |
|
|
|
|
|
|
|
# todo choice origins: |
|
|
|
# choice origins |
|
|
|
prism_program = stormpy.parse_prism_program(get_example_path("dtmc", "die.pm")) |
|
|
|
index_to_identifier_mapping = [1, 2, 3, 4, 5, 6, 7, 8, 8, 8, 8, 8, 8] |
|
|
|
id_to_command_set_mapping = [stormpy.FlatSet() for _ in range(9)] |
|
|
|
for i in range(1, 8): # 0: no origin |
|
|
|
id_to_command_set_mapping[i].insert(i - 1) |
|
|
|
id_to_command_set_mapping[8].insert(7) |
|
|
|
# |
|
|
|
choice_origins = stormpy.PrismChoiceOrigins(prism_program, index_to_identifier_mapping, |
|
|
|
id_to_command_set_mapping) |
|
|
|
|
|
|
|
# Construct Components |
|
|
|
components = stormpy.SparseModelComponents(transition_matrix=transition_matrix, state_labeling=state_labeling, |
|
|
|
reward_models=reward_models) |
|
|
|
components.choice_origins = choice_origins |
|
|
|
|
|
|
|
components.state_valuations = state_valuations |
|
|
|
|
|
|
|
# Build DTMC |
|
|
|
dtmc = stormpy.storage.SparseDtmc(components) |
|
|
|
|
|
|
|
assert type(dtmc) is stormpy.SparseDtmc |
|
|
|
assert not dtmc.supports_parameters |
|
|
|
|
|
|
|
# test transition matrix |
|
|
|
# Test transition matrix |
|
|
|
assert dtmc.nr_choices == nr_choices |
|
|
|
assert dtmc.nr_states == nr_states |
|
|
|
assert dtmc.nr_transitions == 20 |
|
|
@ -119,10 +119,10 @@ class TestSparseModelComponents: |
|
|
|
for state in dtmc.states: |
|
|
|
assert len(state.actions) <= 1 |
|
|
|
|
|
|
|
# test state_labeling |
|
|
|
# Test state_labeling |
|
|
|
assert dtmc.labeling.get_labels() == {'init', 'deadlock', 'done', 'one', 'two', 'three', 'four', 'five', 'six'} |
|
|
|
|
|
|
|
# test reward_models |
|
|
|
# Test reward_models |
|
|
|
assert len(dtmc.reward_models) == 1 |
|
|
|
assert not dtmc.reward_models["coin_flips"].has_state_rewards |
|
|
|
assert dtmc.reward_models["coin_flips"].has_state_action_rewards |
|
|
@ -130,10 +130,10 @@ class TestSparseModelComponents: |
|
|
|
assert reward == 1.0 or reward == 0.0 |
|
|
|
assert not dtmc.reward_models["coin_flips"].has_transition_rewards |
|
|
|
|
|
|
|
# test choice_labeling |
|
|
|
# Test choice_labeling |
|
|
|
assert not dtmc.has_choice_labeling() |
|
|
|
|
|
|
|
# test state_valuations |
|
|
|
# Test state_valuations |
|
|
|
assert dtmc.has_state_valuations() |
|
|
|
assert dtmc.state_valuations |
|
|
|
value_s = [None] * nr_states |
|
|
@ -144,7 +144,162 @@ class TestSparseModelComponents: |
|
|
|
assert value_s == [0, 1, 2, 3, 4, 5, 6, 7, 7, 7, 7, 7, 7] |
|
|
|
assert value_d == [0, 0, 0, 0, 0, 0, 0, 1, 2, 3, 4, 5, 6] |
|
|
|
|
|
|
|
# todo choice_origins more tests |
|
|
|
# Test choice_origins |
|
|
|
assert dtmc.has_choice_origins() |
|
|
|
assert dtmc.choice_origins is components.choice_origins |
|
|
|
assert dtmc.choice_origins.get_number_of_identifiers() == 9 |
|
|
|
|
|
|
|
@numpy_avail |
|
|
|
def test_build_ctmc_from_model_components(self): |
|
|
|
import numpy as np |
|
|
|
|
|
|
|
nr_states = 12 |
|
|
|
nr_choices = 12 |
|
|
|
|
|
|
|
# Build transition_matrix |
|
|
|
transitions = np.array([ |
|
|
|
[0, 0.5, 0.5, 200, 0, 0, 0, 0, 0, 0, 0, 0], |
|
|
|
[0, 0, 0, 0, 0.5, 200, 0, 0, 0, 0, 0, 0], |
|
|
|
[0, 0, 0, 0, 0.5, 0, 200, 0, 0, 0, 0, 0], |
|
|
|
[200, 0, 0, 0, 0, 0, 0.5, 0.5, 0, 0, 0, 0], |
|
|
|
[0, 0, 0, 0, 0, 0, 0, 0, 200, 0, 0, 0], |
|
|
|
[0, 0, 0, 1, 0, 0, 0, 0, 0.5, 0, 0, 0], |
|
|
|
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0.5, 200, 0], |
|
|
|
[0, 200, 0, 0, 0, 0, 0, 0, 0, 0.5, 0, 0], |
|
|
|
[0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0], |
|
|
|
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 200], |
|
|
|
[1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.5], |
|
|
|
[0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], dtype='float64') |
|
|
|
|
|
|
|
transition_matrix = stormpy.build_sparse_matrix(transitions) |
|
|
|
|
|
|
|
# state_labeling |
|
|
|
state_labeling = stormpy.storage.StateLabeling(nr_states) |
|
|
|
# Add labels |
|
|
|
state_labels = {'init', 'deadlock', 'target'} |
|
|
|
for label in state_labels: |
|
|
|
state_labeling.add_label(label) |
|
|
|
|
|
|
|
# Add labels to states |
|
|
|
state_labeling.add_label_to_state('init', 0) |
|
|
|
state_labeling.set_states('target', stormpy.BitVector(nr_states, [5, 8])) |
|
|
|
|
|
|
|
# reward_models |
|
|
|
reward_models = {} |
|
|
|
# vector representing state-action rewards |
|
|
|
action_reward = [0.0, 0.0, 0.0, 0.0, 0.0, 2 / 3, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0] |
|
|
|
reward_models['served'] = stormpy.SparseRewardModel(optional_state_action_reward_vector=action_reward) |
|
|
|
|
|
|
|
# vector representing state rewards |
|
|
|
state_reward = [0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0] |
|
|
|
reward_models['waiting'] = stormpy.SparseRewardModel(optional_state_reward_vector=state_reward) |
|
|
|
|
|
|
|
# choice_labeling |
|
|
|
choice_labeling = stormpy.storage.ChoiceLabeling(nr_choices) |
|
|
|
choice_labels = {'loop1a', 'loop1b', 'serve1', 'loop2a', 'loop2b', 'serve2'} |
|
|
|
# Add labels |
|
|
|
for label in choice_labels: |
|
|
|
choice_labeling.add_label(label) |
|
|
|
|
|
|
|
choice_labeling.set_choices('loop1a', stormpy.BitVector(nr_choices, [0, 2])) |
|
|
|
choice_labeling.set_choices('loop1b', stormpy.BitVector(nr_choices, [1, 4])) |
|
|
|
choice_labeling.set_choices('serve1', stormpy.BitVector(nr_choices, [5, 8])) |
|
|
|
choice_labeling.set_choices('loop2a', stormpy.BitVector(nr_choices, [3, 7])) |
|
|
|
choice_labeling.set_choices('loop2b', stormpy.BitVector(nr_choices, [6, 9])) |
|
|
|
choice_labeling.set_choices('serve2', stormpy.BitVector(nr_choices, [10, 11])) |
|
|
|
|
|
|
|
# state exit rates |
|
|
|
exit_rates = [201.0, 200.5, 200.5, 201.0, 200.0, 1.5, 200.5, 200.5, 1.0, 200.0, 1.5, 1.0] |
|
|
|
|
|
|
|
# state_valuations |
|
|
|
manager = stormpy.ExpressionManager() |
|
|
|
var_s = manager.create_integer_variable(name='s') |
|
|
|
var_a = manager.create_integer_variable(name='a') |
|
|
|
var_s1 = manager.create_integer_variable(name='s1') |
|
|
|
var_s2 = manager.create_integer_variable(name='s2') |
|
|
|
v_builder = stormpy.StateValuationsBuilder() |
|
|
|
v_builder.add_variable(var_s) |
|
|
|
v_builder.add_variable(var_a) |
|
|
|
v_builder.add_variable(var_s1) |
|
|
|
v_builder.add_variable(var_s2) |
|
|
|
|
|
|
|
v_builder.add_state(state=0, boolean_values=[], integer_values=[1, 0, 0, 0], rational_values=[]) |
|
|
|
v_builder.add_state(state=1, boolean_values=[], integer_values=[1, 0, 1, 0], rational_values=[]) |
|
|
|
v_builder.add_state(state=2, boolean_values=[], integer_values=[1, 0, 0, 1], rational_values=[]) |
|
|
|
v_builder.add_state(state=3, boolean_values=[], integer_values=[2, 0, 0, 0], rational_values=[]) |
|
|
|
v_builder.add_state(state=4, boolean_values=[], integer_values=[1, 0, 1, 1], rational_values=[]) |
|
|
|
v_builder.add_state(state=5, boolean_values=[], integer_values=[1, 1, 1, 0], rational_values=[]) |
|
|
|
v_builder.add_state(state=6, boolean_values=[], integer_values=[2, 0, 0, 1], rational_values=[]) |
|
|
|
v_builder.add_state(state=7, boolean_values=[], integer_values=[2, 0, 1, 0], rational_values=[]) |
|
|
|
v_builder.add_state(state=8, boolean_values=[], integer_values=[1, 1, 1, 1], rational_values=[]) |
|
|
|
v_builder.add_state(state=9, boolean_values=[], integer_values=[2, 0, 1, 1], rational_values=[]) |
|
|
|
v_builder.add_state(state=10, boolean_values=[], integer_values=[2, 1, 0, 1], rational_values=[]) |
|
|
|
v_builder.add_state(state=11, boolean_values=[], integer_values=[2, 1, 1, 1], rational_values=[]) |
|
|
|
|
|
|
|
state_valuations = v_builder.build(nr_states) |
|
|
|
|
|
|
|
# set rate_transitions to True: the transition values are interpreted as rates |
|
|
|
components = stormpy.SparseModelComponents(transition_matrix=transition_matrix, state_labeling=state_labeling, |
|
|
|
reward_models=reward_models, rate_transitions=True) |
|
|
|
components.choice_labeling = choice_labeling |
|
|
|
components.exit_rates = exit_rates |
|
|
|
components.state_valuations = state_valuations |
|
|
|
|
|
|
|
ctmc = stormpy.storage.SparseCtmc(components) |
|
|
|
assert type(ctmc) is stormpy.SparseCtmc |
|
|
|
assert not ctmc.supports_parameters |
|
|
|
|
|
|
|
# Test transition matrix |
|
|
|
assert ctmc.nr_choices == nr_choices |
|
|
|
assert ctmc.nr_states == nr_states |
|
|
|
assert ctmc.nr_transitions == 22 |
|
|
|
assert ctmc.transition_matrix.nr_columns == nr_choices |
|
|
|
assert ctmc.transition_matrix.nr_rows == nr_states |
|
|
|
for e in ctmc.transition_matrix: |
|
|
|
assert e.value() == 0.5 or e.value() == 0 or e.value() == 200 or e.value() == 1.0 |
|
|
|
for state in ctmc.states: |
|
|
|
assert len(state.actions) <= 3 |
|
|
|
|
|
|
|
# Test state_labeling |
|
|
|
assert ctmc.labeling.get_labels() == {'target', 'init', 'deadlock'} |
|
|
|
|
|
|
|
# Test reward_models |
|
|
|
assert len(ctmc.reward_models) == 2 |
|
|
|
assert not ctmc.reward_models["served"].has_state_rewards |
|
|
|
assert ctmc.reward_models["served"].has_state_action_rewards |
|
|
|
assert ctmc.reward_models["served"].state_action_rewards == [0.0, 0.0, 0.0, 0.0, 0.0, 0.6666666666666666, 0.0, 0.0, 1.0, 0.0, |
|
|
|
0.0, 0.0] |
|
|
|
assert not ctmc.reward_models["served"].has_transition_rewards |
|
|
|
|
|
|
|
assert ctmc.reward_models["waiting"].has_state_rewards |
|
|
|
assert not ctmc.reward_models["waiting"].has_state_action_rewards |
|
|
|
assert ctmc.reward_models["waiting"].state_rewards == [0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0] |
|
|
|
assert not ctmc.reward_models["waiting"].has_transition_rewards |
|
|
|
|
|
|
|
# Test choice_labeling |
|
|
|
assert ctmc.has_choice_labeling() |
|
|
|
assert ctmc.choice_labeling.get_labels() == {'loop1a', 'loop1b', 'serve1', 'loop2a', 'loop2b', 'serve2'} |
|
|
|
|
|
|
|
# Test state_valuations |
|
|
|
assert ctmc.has_state_valuations() |
|
|
|
assert ctmc.state_valuations |
|
|
|
value_s = [None] * nr_states |
|
|
|
value_a = [None] * nr_states |
|
|
|
value_s1 = [None] * nr_states |
|
|
|
value_s2 = [None] * nr_states |
|
|
|
for s in range(0, ctmc.nr_states): |
|
|
|
value_s[s] = ctmc.state_valuations.get_integer_value(s, var_s) |
|
|
|
value_a[s] = ctmc.state_valuations.get_integer_value(s, var_a) |
|
|
|
value_s1[s] = ctmc.state_valuations.get_integer_value(s, var_s1) |
|
|
|
value_s2[s] = ctmc.state_valuations.get_integer_value(s, var_s2) |
|
|
|
assert value_s == [1,1,1,2,1,1,2,2,1,2,2,2] |
|
|
|
assert value_a == [0,0,0,0,0,1,0,0,1,0,1,1] |
|
|
|
assert value_s1 == [0,1,0,0,1,1,0,1,1,1,0,1] |
|
|
|
assert value_s2 == [0,0,1,0,1,0,1,0,1,1,1,1] |
|
|
|
|
|
|
|
# Test choice_origins |
|
|
|
assert not ctmc.has_choice_origins() |
|
|
|
|
|
|
|
# Test exit_rates |
|
|
|
assert ctmc.exit_rates == [201.0, 200.5, 200.5, 201.0, 200.0, 1.5, 200.5, 200.5, 1.0, 200.0, 1.5, 1.0] |
|
|
|
|