def setUp(self):
self.course_id = 'foo/bar/baz'
self.username = '******'
self.prev_week_start_date = datetime.date(2014, 3, 18)
self.start_date = datetime.date(2014, 3, 25)
self.date = datetime.date(2014, 4, 1)
self.reduce_key = (self.course_id, self.username)
self.input_record = ModuleEngagementSummaryRecord(
course_id=self.course_id,
username=self.username,
start_date=self.start_date,
end_date=self.date,
problem_attempts=0,
problems_attempted=0,
problems_completed=0,
problem_attempts_per_completed=0.0,
videos_viewed=0,
discussion_contributions=0,
days_active=0,
)
self.range_record = ModuleEngagementSummaryMetricRangeRecord(
course_id=self.course_id,
start_date=self.start_date,
end_date=self.date,
metric='problems_attempted',
range_type='high',
low_value=5.0,
high_value=10.0)
self.task = self.task_class( # pylint: disable=not-callable
date=self.date,
output_root=self.DEFAULT_ARGS['output_root'])
def setUp(self):
super(ModuleEngagementSummaryMetricRangesDataTaskReducerTest, self).setUp()
self.reduce_key = 'foo/bar/baz'
self.input_record = ModuleEngagementSummaryRecord(
course_id='foo/bar/baz',
username='******',
start_date=datetime.date(2014, 3, 25),
end_date=datetime.date(2014, 4, 1),
problem_attempts=0,
problems_attempted=0,
problems_completed=0,
problem_attempts_per_completed=0.0,
videos_viewed=0,
discussion_contributions=0,
days_active=0,
)
def setUp(self):
super(ModuleEngagementSummaryMetricRangesDataTaskReducerTest, self).setUp()
self.reduce_key = 'foo/bar/baz'
self.input_record = ModuleEngagementSummaryRecord(
course_id='foo/bar/baz',
username='******',
start_date=datetime.date(2014, 3, 25),
end_date=datetime.date(2014, 4, 1),
problem_attempts=0,
problems_attempted=0,
problems_completed=0,
problem_attempts_per_completed=0.0,
videos_viewed=0,
discussion_contributions=0,
days_active=0,
)
def setUp(self):
self.course_id = 'foo/bar/baz'
self.username = '******'
self.prev_week_start_date = datetime.date(2014, 3, 18)
self.start_date = datetime.date(2014, 3, 25)
self.date = datetime.date(2014, 4, 1)
self.reduce_key = (self.course_id, self.username)
self.input_record = ModuleEngagementSummaryRecord(
course_id=self.course_id,
username=self.username,
start_date=self.start_date,
end_date=self.date,
problem_attempts=0,
problems_attempted=0,
problems_completed=0,
problem_attempts_per_completed=0.0,
videos_viewed=0,
discussion_contributions=0,
days_active=0,
)
self.range_record = ModuleEngagementSummaryMetricRangeRecord(
course_id=self.course_id,
start_date=self.start_date,
end_date=self.date,
metric='problems_attempted',
range_type='high',
low_value=5.0,
high_value=10.0
)
self.task = self.task_class( # pylint: disable=not-callable
date=self.date,
output_root=self.DEFAULT_ARGS['output_root'],
overwrite_from_date=datetime.date(2014, 4, 1),
)
class ModuleEngagementUserSegmentDataTaskReducerTest(ReducerTestMixin, unittest.TestCase):
"""Base class for test analysis of student engagement summaries"""
task_class = ModuleEngagementUserSegmentDataTask
output_record_type = ModuleEngagementUserSegmentRecord
def setUp(self):
self.course_id = 'foo/bar/baz'
self.username = '******'
self.prev_week_start_date = datetime.date(2014, 3, 18)
self.start_date = datetime.date(2014, 3, 25)
self.date = datetime.date(2014, 4, 1)
self.reduce_key = (self.course_id, self.username)
self.input_record = ModuleEngagementSummaryRecord(
course_id=self.course_id,
username=self.username,
start_date=self.start_date,
end_date=self.date,
problem_attempts=0,
problems_attempted=0,
problems_completed=0,
problem_attempts_per_completed=0.0,
videos_viewed=0,
discussion_contributions=0,
days_active=0,
)
self.range_record = ModuleEngagementSummaryMetricRangeRecord(
course_id=self.course_id,
start_date=self.start_date,
end_date=self.date,
metric='problems_attempted',
range_type='high',
low_value=5.0,
high_value=10.0
)
self.task = self.task_class( # pylint: disable=not-callable
date=self.date,
output_root=self.DEFAULT_ARGS['output_root'],
overwrite_from_date=datetime.date(2014, 4, 1),
)
def initialize_task(self, metric_ranges):
"""Given a list of metric ranges, setup the task by calling init_local"""
metric_ranges_text = '\n'.join([
r.to_separated_values()
for r in metric_ranges
])
self.task.input_local = MagicMock(return_value={
'range_data': FakeTarget(value=metric_ranges_text)
})
self.task.init_local()
def test_init_local(self):
other_course_record = self.range_record.replace(
course_id='another/course/id',
metric='problems_completed'
)
self.initialize_task([
self.range_record,
self.range_record.replace(
range_type='low',
low_value=0.0,
high_value=3.0
),
other_course_record
])
self.assertEqual(dict(self.task.high_metric_ranges), {
self.course_id: {
'problems_attempted': self.range_record
},
'another/course/id': {
'problems_completed': other_course_record
}
})
def test_init_local_empty_input(self):
self.initialize_task([])
self.assertEqual(dict(self.task.high_metric_ranges), {})
def test_output_format(self):
self.initialize_task([
self.range_record,
self.range_record.replace(
metric='problem_attempts_per_completed',
low_value=8.0,
high_value=10.1
)
])
self._check_output_complete_tuple(
[
self.input_record.replace(
problems_attempted=6,
problem_attempts_per_completed=9
).to_separated_values()
],
(
(
'foo/bar/baz',
'test_user',
'2014-03-25',
'2014-04-01',
'highly_engaged',
'problems_attempted'
),
(
'foo/bar/baz',
'test_user',
'2014-03-25',
'2014-04-01',
'struggling',
'problem_attempts_per_completed'
),
)
)
@data(
'problems_attempted',
'problems_completed',
'videos_viewed',
'discussion_contributions'
)
def test_highly_engaged(self, metric):
self.initialize_task([
self.range_record.replace(
metric=metric
)
])
self._check_output_by_record_field(
[
self.input_record.replace(
**{metric: 8}
).to_separated_values()
],
{
'segment': 'highly_engaged'
}
)
@data(
'problem_attempts',
'problem_attempts_per_completed',
)
def test_not_highly_engaged(self, metric):
self.initialize_task([
self.range_record.replace(
metric=metric
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
**{metric: 8}
).to_separated_values()
]
)
self.assert_not_in_segment(output, 'highly_engaged')
def assert_not_in_segment(self, output, segment):
"""Assert that the user was not put into the provided segment."""
for record_tuple in output:
record = self.output_record_type.from_string_tuple(record_tuple)
self.assertNotEqual(record.segment, segment)
def test_highly_engaged_too_low(self):
self.initialize_task([
self.range_record.replace(
metric='problems_completed'
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
problems_completed=0
).to_separated_values()
]
)
self.assert_not_in_segment(output, 'highly_engaged')
def test_highly_engaged_left_closed_interval_bottom(self):
self.initialize_task([
self.range_record.replace(
metric='problems_completed',
low_value=6.0
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
problems_completed=6
).to_separated_values()
]
)
self.assert_in_segment(output, 'highly_engaged')
def assert_in_segment(self, output, segment):
"""Assert that the user was put into the provided segment."""
for record_tuple in output:
record = self.output_record_type.from_string_tuple(record_tuple)
if record.segment == segment:
return True
return False
def test_highly_engaged_left_closed_interval_top(self):
self.initialize_task([
self.range_record.replace(
metric='problems_completed',
high_value=9.0
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
problems_completed=9
).to_separated_values()
]
)
self.assert_not_in_segment(output, 'highly_engaged')
def test_disengaging(self):
self.initialize_task([])
output = self._get_reducer_output(
[
self.input_record.replace(
start_date=self.prev_week_start_date,
end_date=self.start_date,
days_active=1,
).to_separated_values()
]
)
self.assert_in_segment(output, 'disengaging')
def test_not_disengaging_only_recent(self):
self.initialize_task([])
output = self._get_reducer_output(
[
self.input_record.replace(
days_active=1
).to_separated_values()
]
)
self.assert_not_in_segment(output, 'disengaging')
def test_struggling(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed',
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
problem_attempts_per_completed=8.0
).to_separated_values()
]
)
self.assert_in_segment(output, 'struggling')
def test_struggling_infinite_low_high_value(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed',
low_value=float('inf'),
high_value=float('inf'),
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
problem_attempts_per_completed=float('inf')
).to_separated_values()
]
)
self.assert_in_segment(output, 'struggling')
def test_struggling_infinite_high(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed',
high_value=float('inf'),
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
problem_attempts_per_completed=10.0
).to_separated_values()
]
)
self.assert_in_segment(output, 'struggling')
def test_struggling_infinite_high_value(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed',
high_value=float('inf'),
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
problem_attempts_per_completed=float('inf')
).to_separated_values()
]
)
self.assert_in_segment(output, 'struggling')
def test_not_struggling(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed',
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
problem_attempts_per_completed=3.0
).to_separated_values()
]
)
self.assert_not_in_segment(output, 'struggling')
def test_not_struggling_infinite_low(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed',
low_value=float('inf')
)
])
output = self._get_reducer_output(
[
self.input_record.replace(
problem_attempts_per_completed=100000.0
).to_separated_values()
]
)
self.assert_not_in_segment(output, 'struggling')
class ModuleEngagementSummaryMetricRangesDataTaskReducerTest(ReducerTestMixin, unittest.TestCase):
"""Base class for test analysis of student engagement summaries"""
task_class = ModuleEngagementSummaryMetricRangesDataTask
output_record_type = ModuleEngagementSummaryMetricRangeRecord
def setUp(self):
super(ModuleEngagementSummaryMetricRangesDataTaskReducerTest, self).setUp()
self.reduce_key = 'foo/bar/baz'
self.input_record = ModuleEngagementSummaryRecord(
course_id='foo/bar/baz',
username='******',
start_date=datetime.date(2014, 3, 25),
end_date=datetime.date(2014, 4, 1),
problem_attempts=0,
problems_attempted=0,
problems_completed=0,
problem_attempts_per_completed=0.0,
videos_viewed=0,
discussion_contributions=0,
days_active=0,
)
def test_simple_distribution(self):
# [0, 0, 0, 0] (these values are dropped from the set before analyzing)
# [4, 13, 13, 13] (3 records are <= 13, this accounts for 15% of the total 20 non-zero values)
# [15] * 4 (throw in a bunch of data in the "normal" range)
# [50] * 11 (round out the 20 records with some other arbitrary value, note that this will also contain two
# of the three highest values)
# [154] (throw in an outlier - a very high maximum value, this will show the high end of the range, but the
# 85th percentile should be at the 50 value)
values = [4] + ([13] * 3) + ([0] * 4) + ([15] * 4) + ([50] * 11) + [154]
self.assert_ranges(
values,
[
('low', 0, 13.0),
('normal', 13.0, 50.0),
('high', 50.0, 'inf'),
]
)
def assert_ranges(self, values, range_values):
"""Given a list of values, assert that the ranges generated have the min, low, high, and max bounds."""
# Manufacture some records with these values
records = [self.input_record.replace(problem_attempts_per_completed=v).to_separated_values() for v in values]
output = self._get_reducer_output(records)
range_value_map = {rv[0]: rv for rv in range_values}
for record in output:
if record[3] == 'problem_attempts_per_completed':
range_type, low, high = range_value_map[record[4]]
self.assertEqual(
record,
(
'foo/bar/baz',
'2014-03-25',
'2014-04-01',
'problem_attempts_per_completed',
range_type,
str(low),
str(high),
)
)
def test_identical_values(self):
values = [5] * 6
self.assert_ranges(values, [('low', 0, 5.0), ('normal', 5.0, 'inf')])
def test_single_value(self):
self.assert_ranges([1], [('low', 0, 1.0), ('normal', 1.0, 'inf')])
def test_single_infinite_value(self):
values = [float('inf')] * 3
self.assert_ranges(values, [('low', 0, 'inf'), ('normal', 'inf', 'inf')])
def test_infinite_threshold_low_normal(self):
values = [1, float('inf'), float('inf'), float('inf')]
self.assert_ranges(values, [('low', 0, 'inf'), ('normal', 'inf', 'inf')])
def test_infinite_threshold_normal_high(self):
values = [1, 2, 3, float('inf'), float('inf')]
self.assert_ranges(values, [('low', 0, 1.6), ('normal', 1.6, 'inf'), ('high', 'inf', 'inf')])
def test_infinite_value_in_high(self):
values = [1, 2, 3, 4, 5, 6, 7, float('inf')]
self.assert_ranges(values, [('low', 0, 2.05), ('normal', 2.05, 6.95), ('high', 6.95, 'inf')])
def test_no_values(self):
self.assert_ranges([], [('normal', 0, 'inf')])
def test_single_zero_value(self):
values = [0] * 3
self.assert_ranges(values, [('normal', 0, 'inf')])
def test_zeroes_are_normal(self):
values = [1, 0, 0, 0]
self.assert_ranges(values, [('normal', 0, 0.55), ('high', 0.55, 'inf')])
def test_zeroes_are_low(self):
values = [0, 0, 0] + ([1] * 10) + ([2]*4)
self.assert_ranges(values, [('low', 0, 0.4), ('normal', 0.4, 2.0), ('high', 2.0, 'inf')])
class ModuleEngagementUserSegmentDataTaskReducerTest(ReducerTestMixin,
unittest.TestCase):
"""Base class for test analysis of student engagement summaries"""
task_class = ModuleEngagementUserSegmentDataTask
output_record_type = ModuleEngagementUserSegmentRecord
def setUp(self):
self.course_id = 'foo/bar/baz'
self.username = '******'
self.prev_week_start_date = datetime.date(2014, 3, 18)
self.start_date = datetime.date(2014, 3, 25)
self.date = datetime.date(2014, 4, 1)
self.reduce_key = (self.course_id, self.username)
self.input_record = ModuleEngagementSummaryRecord(
course_id=self.course_id,
username=self.username,
start_date=self.start_date,
end_date=self.date,
problem_attempts=0,
problems_attempted=0,
problems_completed=0,
problem_attempts_per_completed=0.0,
videos_viewed=0,
discussion_contributions=0,
days_active=0,
)
self.range_record = ModuleEngagementSummaryMetricRangeRecord(
course_id=self.course_id,
start_date=self.start_date,
end_date=self.date,
metric='problems_attempted',
range_type='high',
low_value=5.0,
high_value=10.0)
self.task = self.task_class( # pylint: disable=not-callable
date=self.date,
output_root=self.DEFAULT_ARGS['output_root'])
def initialize_task(self, metric_ranges):
"""Given a list of metric ranges, setup the task by calling init_local"""
metric_ranges_text = '\n'.join(
[r.to_separated_values() for r in metric_ranges])
self.task.input_local = MagicMock(
return_value={'range_data': FakeTarget(value=metric_ranges_text)})
self.task.init_local()
def test_init_local(self):
other_course_record = self.range_record.replace(
course_id='another/course/id', metric='problems_completed')
self.initialize_task([
self.range_record,
self.range_record.replace(range_type='low',
low_value=0.0,
high_value=3.0), other_course_record
])
self.assertEqual(
dict(self.task.high_metric_ranges), {
self.course_id: {
'problems_attempted': self.range_record
},
'another/course/id': {
'problems_completed': other_course_record
}
})
def test_init_local_empty_input(self):
self.initialize_task([])
self.assertEqual(dict(self.task.high_metric_ranges), {})
def test_output_format(self):
self.initialize_task([
self.range_record,
self.range_record.replace(metric='problem_attempts_per_completed',
low_value=8.0,
high_value=10.1)
])
self._check_output_complete_tuple([
self.input_record.replace(
problems_attempted=6,
problem_attempts_per_completed=9).to_separated_values()
], (
('foo/bar/baz', 'test_user', '2014-03-25', '2014-04-01',
'highly_engaged', 'problems_attempted'),
('foo/bar/baz', 'test_user', '2014-03-25', '2014-04-01',
'struggling', 'problem_attempts_per_completed'),
))
@data('problems_attempted', 'problems_completed', 'videos_viewed',
'discussion_contributions')
def test_highly_engaged(self, metric):
self.initialize_task([self.range_record.replace(metric=metric)])
self._check_output_by_record_field(
[self.input_record.replace(**{
metric: 8
}).to_separated_values()], {'segment': 'highly_engaged'})
@data(
'problem_attempts',
'problem_attempts_per_completed',
)
def test_not_highly_engaged(self, metric):
self.initialize_task([self.range_record.replace(metric=metric)])
output = self._get_reducer_output(
[self.input_record.replace(**{
metric: 8
}).to_separated_values()])
self.assert_not_in_segment(output, 'highly_engaged')
def assert_not_in_segment(self, output, segment):
"""Assert that the user was not put into the provided segment."""
for record_tuple in output:
record = self.output_record_type.from_string_tuple(record_tuple)
self.assertNotEqual(record.segment, segment)
def test_highly_engaged_too_low(self):
self.initialize_task(
[self.range_record.replace(metric='problems_completed')])
output = self._get_reducer_output([
self.input_record.replace(
problems_completed=0).to_separated_values()
])
self.assert_not_in_segment(output, 'highly_engaged')
def test_highly_engaged_left_closed_interval_bottom(self):
self.initialize_task([
self.range_record.replace(metric='problems_completed',
low_value=6.0)
])
output = self._get_reducer_output([
self.input_record.replace(
problems_completed=6).to_separated_values()
])
self.assert_in_segment(output, 'highly_engaged')
def assert_in_segment(self, output, segment):
"""Assert that the user was put into the provided segment."""
for record_tuple in output:
record = self.output_record_type.from_string_tuple(record_tuple)
if record.segment == segment:
return True
return False
def test_highly_engaged_left_closed_interval_top(self):
self.initialize_task([
self.range_record.replace(metric='problems_completed',
high_value=9.0)
])
output = self._get_reducer_output([
self.input_record.replace(
problems_completed=9).to_separated_values()
])
self.assert_not_in_segment(output, 'highly_engaged')
def test_disengaging(self):
self.initialize_task([])
output = self._get_reducer_output([
self.input_record.replace(
start_date=self.prev_week_start_date,
end_date=self.start_date,
days_active=1,
).to_separated_values()
])
self.assert_in_segment(output, 'disengaging')
def test_not_disengaging_only_recent(self):
self.initialize_task([])
output = self._get_reducer_output(
[self.input_record.replace(days_active=1).to_separated_values()])
self.assert_not_in_segment(output, 'disengaging')
def test_struggling(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed', )
])
output = self._get_reducer_output([
self.input_record.replace(
problem_attempts_per_completed=8.0).to_separated_values()
])
self.assert_in_segment(output, 'struggling')
def test_struggling_infinite_low_high_value(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed',
low_value=float('inf'),
high_value=float('inf'),
)
])
output = self._get_reducer_output([
self.input_record.replace(problem_attempts_per_completed=float(
'inf')).to_separated_values()
])
self.assert_in_segment(output, 'struggling')
def test_struggling_infinite_high(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed',
high_value=float('inf'),
)
])
output = self._get_reducer_output([
self.input_record.replace(
problem_attempts_per_completed=10.0).to_separated_values()
])
self.assert_in_segment(output, 'struggling')
def test_struggling_infinite_high_value(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed',
high_value=float('inf'),
)
])
output = self._get_reducer_output([
self.input_record.replace(problem_attempts_per_completed=float(
'inf')).to_separated_values()
])
self.assert_in_segment(output, 'struggling')
def test_not_struggling(self):
self.initialize_task([
self.range_record.replace(
metric='problem_attempts_per_completed', )
])
output = self._get_reducer_output([
self.input_record.replace(
problem_attempts_per_completed=3.0).to_separated_values()
])
self.assert_not_in_segment(output, 'struggling')
def test_not_struggling_infinite_low(self):
self.initialize_task([
self.range_record.replace(metric='problem_attempts_per_completed',
low_value=float('inf'))
])
output = self._get_reducer_output([
self.input_record.replace(
problem_attempts_per_completed=100000.0).to_separated_values()
])
self.assert_not_in_segment(output, 'struggling')
class ModuleEngagementSummaryMetricRangesDataTaskReducerTest(
ReducerTestMixin, unittest.TestCase):
"""Base class for test analysis of student engagement summaries"""
task_class = ModuleEngagementSummaryMetricRangesDataTask
output_record_type = ModuleEngagementSummaryMetricRangeRecord
def setUp(self):
super(ModuleEngagementSummaryMetricRangesDataTaskReducerTest,
self).setUp()
self.reduce_key = 'foo/bar/baz'
self.input_record = ModuleEngagementSummaryRecord(
course_id='foo/bar/baz',
username='******',
start_date=datetime.date(2014, 3, 25),
end_date=datetime.date(2014, 4, 1),
problem_attempts=0,
problems_attempted=0,
problems_completed=0,
problem_attempts_per_completed=0.0,
videos_viewed=0,
discussion_contributions=0,
days_active=0,
)
def test_simple_distribution(self):
# [0, 0, 0, 0] (these values are dropped from the set before analyzing)
# [4, 13, 13, 13] (3 records are <= 13, this accounts for 15% of the total 20 non-zero values)
# [15] * 4 (throw in a bunch of data in the "normal" range)
# [50] * 11 (round out the 20 records with some other arbitrary value, note that this will also contain two
# of the three highest values)
# [154] (throw in an outlier - a very high maximum value, this will show the high end of the range, but the
# 85th percentile should be at the 50 value)
values = [4] + ([13] * 3) + ([0] * 4) + ([15] * 4) + ([50] *
11) + [154]
self.assert_ranges(values, 13.0, 50.0)
def assert_ranges(self, values, low, high):
"""Given a list of values, assert that the ranges generated have the min, low, high, and max bounds."""
# Manufacture some records with these values
records = [
self.input_record.replace(
problem_attempts_per_completed=v).to_separated_values()
for v in values
]
self._check_output_complete_tuple(records, (
(
'foo/bar/baz',
'2014-03-25',
'2014-04-01',
'problem_attempts_per_completed',
'low',
'0',
str(low),
),
(
'foo/bar/baz',
'2014-03-25',
'2014-04-01',
'problem_attempts_per_completed',
'normal',
str(low),
str(high),
),
(
'foo/bar/baz',
'2014-03-25',
'2014-04-01',
'problem_attempts_per_completed',
'high',
str(high),
'inf',
),
))
def test_identical_values(self):
values = [5] * 6
self.assert_ranges(values, 5.0, 5.0)
def test_single_value(self):
self.assert_ranges([1], 1.0, 1.0)
def test_very_small_values(self):
self.assert_ranges(([0.01] * 10) + ([0.09] * 10), 0.01, 0.09)
def test_infinite_value(self):
self.assert_ranges(([1.0] * 19) + [float('inf')], 1.0, 1.0)
class ModuleEngagementSummaryMetricRangesDataTaskReducerTest(ReducerTestMixin, unittest.TestCase):
"""Base class for test analysis of student engagement summaries"""
task_class = ModuleEngagementSummaryMetricRangesDataTask
output_record_type = ModuleEngagementSummaryMetricRangeRecord
def setUp(self):
super(ModuleEngagementSummaryMetricRangesDataTaskReducerTest, self).setUp()
self.reduce_key = 'foo/bar/baz'
self.input_record = ModuleEngagementSummaryRecord(
course_id='foo/bar/baz',
username='******',
start_date=datetime.date(2014, 3, 25),
end_date=datetime.date(2014, 4, 1),
problem_attempts=0,
problems_attempted=0,
problems_completed=0,
problem_attempts_per_completed=0.0,
videos_viewed=0,
discussion_contributions=0,
days_active=0,
)
def test_simple_distribution(self):
# [0, 0, 0, 0] (these values are dropped from the set before analyzing)
# [4, 13, 13, 13] (3 records are <= 13, this accounts for 15% of the total 20 non-zero values)
# [15] * 4 (throw in a bunch of data in the "normal" range)
# [50] * 11 (round out the 20 records with some other arbitrary value, note that this will also contain two
# of the three highest values)
# [154] (throw in an outlier - a very high maximum value, this will show the high end of the range, but the
# 85th percentile should be at the 50 value)
values = [4] + ([13] * 3) + ([0] * 4) + ([15] * 4) + ([50] * 11) + [154]
self.assert_ranges(
values,
[
('low', 0, 13.0),
('normal', 13.0, 50.0),
('high', 50.0, 'inf'),
]
)
def assert_ranges(self, values, range_values):
"""Given a list of values, assert that the ranges generated have the min, low, high, and max bounds."""
# Manufacture some records with these values
records = [self.input_record.replace(problem_attempts_per_completed=v).to_separated_values() for v in values]
output = self._get_reducer_output(records)
range_value_map = {rv[0]: rv for rv in range_values}
for record in output:
if record[3] == 'problem_attempts_per_completed':
range_type, low, high = range_value_map[record[4]]
self.assertEqual(
record,
(
'foo/bar/baz',
'2014-03-25',
'2014-04-01',
'problem_attempts_per_completed',
range_type,
str(low),
str(high),
)
)
def test_identical_values(self):
values = [5] * 6
self.assert_ranges(values, [('low', 0, 5.0), ('normal', 5.0, 'inf')])
def test_single_value(self):
self.assert_ranges([1], [('low', 0, 1.0), ('normal', 1.0, 'inf')])
def test_single_infinite_value(self):
values = [float('inf')] * 3
self.assert_ranges(values, [('low', 0, 'inf'), ('normal', 'inf', 'inf')])
def test_infinite_threshold_low_normal(self):
values = [1, float('inf'), float('inf'), float('inf')]
self.assert_ranges(values, [('low', 0, 'inf'), ('normal', 'inf', 'inf')])
def test_infinite_threshold_normal_high(self):
values = [1, 2, 3, float('inf'), float('inf')]
self.assert_ranges(values, [('low', 0, 1.6), ('normal', 1.6, 'inf'), ('high', 'inf', 'inf')])
def test_infinite_value_in_high(self):
values = [1, 2, 3, 4, 5, 6, 7, float('inf')]
self.assert_ranges(values, [('low', 0, 2.05), ('normal', 2.05, 6.95), ('high', 6.95, 'inf')])
def test_no_values(self):
self.assert_ranges([], [('normal', 0, 'inf')])
def test_single_zero_value(self):
values = [0] * 3
self.assert_ranges(values, [('normal', 0, 'inf')])
def test_zeroes_are_normal(self):
values = [1, 0, 0, 0]
self.assert_ranges(values, [('normal', 0, 0.55), ('high', 0.55, 'inf')])
def test_zeroes_are_low(self):
values = [0, 0, 0] + ([1] * 10) + ([2]*4)
self.assert_ranges(values, [('low', 0, 0.4), ('normal', 0.4, 2.0), ('high', 2.0, 'inf')])
