summaryrefslogtreecommitdiff
path: root/lib/automata.py
diff options
context:
space:
mode:
authorDaniel Friesel <derf@finalrewind.org>2019-02-21 12:04:50 +0100
committerDaniel Friesel <derf@finalrewind.org>2019-02-21 12:04:50 +0100
commit46e69c4f402414d93e6efb2a7138e1fc13bb66d3 (patch)
tree88d19091947a1fa76adde5963a954e1433bf8db8 /lib/automata.py
parentcc3bac282f011310cdb7e29e797a9f766d860faf (diff)
automata: .states -> .state (it's a dict)
Diffstat (limited to 'lib/automata.py')
-rwxr-xr-xlib/automata.py20
1 files changed, 10 insertions, 10 deletions
diff --git a/lib/automata.py b/lib/automata.py
index 3913670..fbfaf36 100755
--- a/lib/automata.py
+++ b/lib/automata.py
@@ -249,7 +249,7 @@ class PTA:
parameters -- names of PTA parameters
initial_param_values -- initial value for each parameter
"""
- self.states = dict([[state_name, State(state_name)] for state_name in state_names])
+ self.state = dict([[state_name, State(state_name)] for state_name in state_names])
self.parameters = parameters.copy()
if initial_param_values:
self.initial_param_values = initial_param_values.copy()
@@ -258,7 +258,7 @@ class PTA:
self.transitions = []
if not 'UNINITIALIZED' in state_names:
- self.states['UNINITIALIZED'] = State('UNINITIALIZED')
+ self.state['UNINITIALIZED'] = State('UNINITIALIZED')
@classmethod
def from_json(cls, json_input: dict):
@@ -272,7 +272,7 @@ class PTA:
if key in json_input:
kwargs[key] = json_input[key]
pta = cls(**kwargs)
- for name, state in json_input['states'].items():
+ for name, state in json_input['state'].items():
power_function = _json_function_to_analytic_function(state, 'power', pta.parameters)
pta.add_state(name, power = _json_get_static(state, 'power'), power_function = power_function)
for transition in json_input['transitions']:
@@ -308,7 +308,7 @@ class PTA:
ret = {
'parameters' : self.parameters,
'initial_param_values' : self.initial_param_values,
- 'states' : dict([[state.name, state.to_json()] for state in self.states.values()]),
+ 'state' : dict([[state.name, state.to_json()] for state in self.state.values()]),
'transitions' : [trans.to_json() for trans in self.transitions]
}
return ret
@@ -322,7 +322,7 @@ class PTA:
if 'power_function' in kwargs and type(kwargs['power_function']) != AnalyticFunction:
kwargs['power_function'] = AnalyticFunction(kwargs['power_function'],
self.parameters, 0)
- self.states[state_name] = State(state_name, **kwargs)
+ self.state[state_name] = State(state_name, **kwargs)
def add_transition(self, orig_state: str, dest_state: str, function_name: str, **kwargs):
"""
@@ -334,8 +334,8 @@ class PTA:
function_name -- function name
kwargs -- see Transition() documentation
"""
- orig_state = self.states[orig_state]
- dest_state = self.states[dest_state]
+ orig_state = self.state[orig_state]
+ dest_state = self.state[dest_state]
for key in ('duration_function', 'energy_function', 'timeout_function'):
if key in kwargs and type(kwargs[key]) != AnalyticFunction:
kwargs[key] = AnalyticFunction(kwargs[key], self.parameters, 0)
@@ -352,12 +352,12 @@ class PTA:
depth -- search depth
orig_state -- initial state for depth-first search
"""
- return self.states[orig_state].dfs(depth, **kwargs)
+ return self.state[orig_state].dfs(depth, **kwargs)
def simulate(self, trace: list, orig_state: str = 'UNINITIALIZED'):
total_duration = 0.
total_energy = 0.
- state = self.states[orig_state]
+ state = self.state[orig_state]
param_dict = dict([[self.parameters[i], self.initial_param_values[i]] for i in range(len(self.parameters))])
for function in trace:
function_name = function[0]
@@ -383,7 +383,7 @@ class PTA:
return total_energy, total_duration, state, param_dict
def update(self, static_model, param_model):
- for state in self.states.values():
+ for state in self.state.values():
if state.name != 'UNINITIALIZED':
state.power = static_model(state.name, 'power')
if param_model(state.name, 'power'):