def test_subsample_array():
t = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
# this will modify the original array and shuffle it
r = utils.subsample_array(t, 5)
assert len(r) == 5
assert set(r).issubset(t)
def _subsample_trackers(
self, incoming_trackers: List[TrackerWithCachedStates],
max_number_of_trackers: int) -> List[TrackerWithCachedStates]:
"""Subsample the list of trackers to retrieve a random subset."""
# if flows get very long and have a lot of forks we
# get into trouble by collecting too many trackers
# hence the sub sampling
if max_number_of_trackers is not None:
return utils.subsample_array(incoming_trackers,
max_number_of_trackers,
rand=self.config.rand)
else:
return incoming_trackers
def _create_start_trackers_for_augmentation(
self, story_end_trackers: List[TrackerWithCachedStates]
) -> TrackerLookupDict:
"""This is where the augmentation magic happens.
We will reuse all the trackers that reached the
end checkpoint `None` (which is the end of a
story) and start processing all steps again. So instead
of starting with a fresh tracker, the second and
all following phases will reuse a couple of the trackers
that made their way to a story end.
We need to do some cleanup before processing them again.
"""
next_active_trackers = defaultdict(list)
if self.config.use_story_concatenation:
ending_trackers = utils.subsample_array(
story_end_trackers,
self.config.augmentation_factor,
rand=self.config.rand,
)
for t in ending_trackers:
# this is a nasty thing - all stories end and
# start with action listen - so after logging the first
# actions in the next phase the trackers would
# contain action listen followed by action listen.
# to fix this we are going to "undo" the last action listen
# tracker should be copied,
# otherwise original tracker is updated
aug_t = t.copy()
aug_t.is_augmented = True
aug_t.update(ActionReverted())
next_active_trackers[STORY_START].append(aug_t)
return next_active_trackers
def test_subsample_array_read_only():
t = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
r = utils.subsample_array(t, 5, can_modify_incoming_array=False)
assert len(r) == 5
assert set(r).issubset(t)