def test_interval_storage_for_has_entry(self):
interval_storage = IntervalStorage()
self.assertFalse(interval_storage.has_entry(0.5))
self.assertFalse(interval_storage.has_entry(1.5))
interval_storage.add_interval("1", 0.3, 0.7)
self.assertTrue(interval_storage.has_entry(0.5))
self.assertFalse(interval_storage.has_entry(0.8))
self.assertFalse(interval_storage.has_entry(0.2))
self.assertFalse(interval_storage.has_entry(-0.2))
def test_interval_storage_for_has_entry(self):
interval_storage = IntervalStorage()
self.assertFalse(interval_storage.has_entry(0.5))
self.assertFalse(interval_storage.has_entry(1.5))
interval_storage.add_interval("1", 0.3, 0.7)
self.assertTrue(interval_storage.has_entry(0.5))
self.assertFalse(interval_storage.has_entry(0.8))
self.assertFalse(interval_storage.has_entry(0.2))
self.assertFalse(interval_storage.has_entry(-0.2))
class OptimizationAlgorithmStageHandler(object):
def __init__(self, settings, stages=None):
self.settings = settings
if stages == None:
stages = self.settings.get("optimization_algorithm.default_stages")
for key in stages.iterkeys():
if not issubclass(key, OptimizationAlgorithmStage):
raise ValueError(
"OptimizationAlgorithmStageHandler must contain OptimizationAlgorithmStage subclasses."
)
self.stage_storage = IntervalStorage(stages)
self.stages_used = []
def add_stage(self, stage, start_percentage, end_percentage):
self.stage_storage.add_interval(stage, start_percentage,
end_percentage)
def has_stage(self, current_step, total_steps=None):
step_percentage = self._get_step_percentage(current_step, total_steps)
return self.stage_storage.has_entry(step_percentage)
def get_stage(self, current_step, total_steps=None):
step_percentage = self._get_step_percentage(current_step, total_steps)
return self.stage_storage.get_entry(step_percentage)
# TODO: its possible that an optimization step won't get run if stage == None and
# on_step doesn't run. Need some way to fix that.
def run_stage(self, current_step, total_steps=None, payload=None):
stage = None
if self.has_stage(current_step, total_steps):
stage = self.get_stage(current_step, total_steps)
# If we've started using a new stage, check to see if we have to run the
# `after_stage` method.
if (stage == None) or not self._is_used_stage(stage):
if len(self.stages_used) > 0:
last_stage, has_run_after_stage = self.stages_used[-1]
if not has_run_after_stage:
last_stage(self.settings).after_stage(payload)
self.stages_used[-1] = (last_stage, True)
# If this is the first time we've seen the stage, then run the `before_stage`
# method
if (stage != None) and not self._is_used_stage(stage):
stage(self.settings).before_stage(payload)
self.stages_used.append((stage, False))
# Run the stage if it exists
if stage != None:
stage(self.settings).on_step(payload=payload)
# If this is the last step, then we need to run the after_stage
total_steps = self._get_total_steps(total_steps)
if stage != None and (current_step >= total_steps - 1):
print "Running after stage"
stage(self.settings).after_stage(payload)
def _is_used_stage(self, stage):
for used_stage, _ in self.stages_used:
if stage == used_stage:
return True
return False
def _get_step_percentage(self, current_step, total_steps=None):
total_steps = self._get_total_steps(total_steps)
return float(current_step) / total_steps
def _get_total_steps(self, total_steps=None):
if total_steps == None:
return self.settings.get(
"optimization_algorithm.finishing_criteria.max_steps")
else:
return total_steps
class OptimizationAlgorithmStageHandler(object):
def __init__(self, settings, stages=None):
self.settings = settings
if stages == None:
stages = self.settings.get("optimization_algorithm.default_stages")
for key in stages.iterkeys():
if not issubclass(key, OptimizationAlgorithmStage):
raise ValueError("OptimizationAlgorithmStageHandler must contain OptimizationAlgorithmStage subclasses.")
self.stage_storage = IntervalStorage(stages)
self.stages_used = []
def add_stage(self, stage, start_percentage, end_percentage):
self.stage_storage.add_interval(stage, start_percentage, end_percentage)
def has_stage(self, current_step, total_steps=None):
step_percentage = self._get_step_percentage(current_step, total_steps)
return self.stage_storage.has_entry(step_percentage)
def get_stage(self, current_step, total_steps=None):
step_percentage = self._get_step_percentage(current_step, total_steps)
return self.stage_storage.get_entry(step_percentage)
# TODO: its possible that an optimization step won't get run if stage == None and
# on_step doesn't run. Need some way to fix that.
def run_stage(self, current_step, total_steps=None, payload=None):
stage = None
if self.has_stage(current_step, total_steps):
stage = self.get_stage(current_step, total_steps)
# If we've started using a new stage, check to see if we have to run the
# `after_stage` method.
if (stage == None) or not self._is_used_stage(stage):
if len(self.stages_used) > 0:
last_stage, has_run_after_stage = self.stages_used[-1]
if not has_run_after_stage:
last_stage(self.settings).after_stage(payload)
self.stages_used[-1] = (last_stage, True)
# If this is the first time we've seen the stage, then run the `before_stage`
# method
if (stage != None) and not self._is_used_stage(stage):
stage(self.settings).before_stage(payload)
self.stages_used.append((stage, False))
# Run the stage if it exists
if stage != None:
stage(self.settings).on_step(payload=payload)
# If this is the last step, then we need to run the after_stage
total_steps = self._get_total_steps(total_steps)
if stage != None and (current_step >= total_steps - 1):
print "Running after stage"
stage(self.settings).after_stage(payload)
def _is_used_stage(self, stage):
for used_stage, _ in self.stages_used:
if stage == used_stage:
return True
return False
def _get_step_percentage(self, current_step, total_steps=None):
total_steps = self._get_total_steps(total_steps)
return float(current_step) / total_steps
def _get_total_steps(self, total_steps=None):
if total_steps == None:
return self.settings.get("optimization_algorithm.finishing_criteria.max_steps")
else:
return total_steps