def process_dataframe(self, dataframe, start, end, sources, resample='T'):
source_boundary = max(
[x.boundary for x in sources] if sources else [timedelta(0)])
fields = [f for f in dataframe.columns if f not in self._HIDDEN_FIELDS]
dataframe = self.sort_dataframe(dataframe)[fields + ['source']]
if dataframe.empty:
return TimeSeries.from_dataframe(
dataframe[fields], source_boundary).dataframe.reset_index()
output_dataframe = None
for src in sources:
src_timeseries = TimeSeries.from_dataframe(
dataframe[dataframe['source'] == src.slug][fields],
source_boundary)
src_dataframe = src_timeseries.query(start, end, 'T')
if output_dataframe is None:
output_dataframe = src_dataframe
continue
output_dataframe = output_dataframe.combine_first(src_dataframe)
dataframe = output_dataframe
dataframe = self.model.postprocess_dataframe(dataframe)
dataframe = dataframe.asfreq(resample)
return dataframe.reset_index()
def test_iters(self):
ts = TimeSeries([1, 2, 3], [4, 5, 6])
times = ts.times()
count = 0
for item in ts:
self.assertEqual(item, ts[times[count]])
count += 1
self.assertEqual(count, len(ts))
count = 0
for item in ts.itervalues():
self.assertEqual(item, ts[times[count]])
count += 1
self.assertEqual(count, len(ts))
count = 0
for item in ts.itertimes():
self.assertEqual(item, times[count])
count += 1
self.assertEqual(count, len(ts))
count = 0
for item in ts.iteritems():
self.assertEqual(item, (times[count], ts[times[count]]))
count += 1
self.assertEqual(count, len(ts))
def test_add(self):
a = TimeSeries([1, 2, 3], [1, 2, 3])
b = TimeSeries([1, 2, 3], [4, 5, 6])
c = TimeSeries([4, 5, 6], [4, 5, 6])
self.assertEqual(a + b, TimeSeries([1, 2, 3], [5, 7, 9]))
with self.assertRaises(ValueError):
d = b + c
def process_dataframe(self, dataframe, start, end, sources, resample='T'):
source_boundary = max([x.boundary for x in sources] if sources
else [timedelta(0)])
fields = [f for f in dataframe.columns if f not in self._HIDDEN_FIELDS]
dataframe = self.sort_dataframe(dataframe)[fields + ['source']]
if dataframe.empty:
return TimeSeries.from_dataframe(
dataframe[fields], source_boundary).dataframe.reset_index()
output_dataframe = None
for src in sources:
src_timeseries = TimeSeries.from_dataframe(
dataframe[dataframe['source'] == src.slug][fields],
source_boundary)
src_dataframe = src_timeseries.query(start, end, 'T')
if output_dataframe is None:
output_dataframe = src_dataframe
continue
output_dataframe = output_dataframe.combine_first(src_dataframe)
dataframe = output_dataframe
dataframe = self.model.postprocess_dataframe(dataframe)
dataframe = dataframe.asfreq(resample)
return dataframe.reset_index()
def test_mul(self):
a = TimeSeries([1, 2, 3], [1, 2, 3])
b = TimeSeries([1, 2, 3], [4, 5, 6])
c = TimeSeries([4, 5, 6], [4, 5, 6])
self.assertEqual(a * b, TimeSeries([1, 2, 3], [4, 10, 18]))
with self.assertRaises(ValueError):
d = b * c
def test_neg(self):
self.assertListEqual(
list(TimeSeries([1, 2, 3], [-1, 2, -4]).__neg__()), [1, -2, 4])
self.assertListEqual(list(TimeSeries([1, 2, 3], [1, 2, 4]).__neg__()),
[-1, -2, -4])
self.assertListEqual(
list(TimeSeries([1, 2, 3], [-4, -6, -7]).__neg__()), [4, 6, 7])
with self.assertRaises(ValueError):
TimeSeries([], []).__neg__()
def test_range_slice_with_step(self):
t = TimeSeries(range(0, 10), range(0, 10))
t_slice = t[0:8:2]
self.assertEqual(t_slice.times, [0, 2, 4, 6])
self.assertEqual(t_slice.values, [0, 2, 4, 6])
t = TimeSeries(range(2, 10, 2), range(2, 10, 2))
t_slice = t[3:7:2]
self.assertEqual(t_slice.times, [3, 5])
self.assertEqual(t_slice.values, [2, 4])
def test_range_slice_with_interpolation_and_step(self):
t = TimeSeries(range(2, 10, 2), range(2, 10, 2), interpolate=True)
t_slice = t[3:8:2]
self.assertEqual(t_slice.times, [3, 5, 7])
self.assertEqual(t_slice.values, [3, 5, 7])
t_slice = t[3:7:2]
self.assertEqual(t_slice.times, [3, 5])
self.assertEqual(t_slice.values, [3, 5])
def test_range_slice_with_first_val(self):
t = TimeSeries(range(2, 10, 2), range(2, 10, 2), first_val=-1)
t_slice = t[0:6]
self.assertEqual(t_slice.times, [0, 2, 4])
self.assertEqual(t_slice.values, [-1, 2, 4])
t_slice = t[0:]
self.assertEqual(t_slice.times, [0, 2, 4, 6, 8])
self.assertEqual(t_slice.values, [-1, 2, 4, 6, 8])
def test_range_slice(self):
t = TimeSeries(range(2, 10, 2), range(2, 10, 2))
t_slice = t[2:]
self.assertEqual(t_slice.times, [2, 4, 6, 8])
self.assertEqual(t_slice.values, [2, 4, 6, 8])
t_slice = t[2:8]
self.assertEqual(t_slice.times, [2, 4, 6])
self.assertEqual(t_slice.values, [2, 4, 6])
t_slice = t[:6]
self.assertEqual(t_slice.times, [2, 4])
self.assertEqual(t_slice.values, [2, 4])
def draw_time_series(self, *args, **kwargs):
"""
Draw a time series visualization on this :class:`~Drawable`.
The arguments and keyword arguments are those supported by the :class:`~timeseries.timeseries.TimeSeries`' :func:`~timeseries.timeseries.TimeSeries.draw` method.
:return: A tuple made up of the drawn plot and label.
:rtype: tuple
"""
self.timeseries = self.timeseries or TimeSeries(self)
return self.timeseries.draw(*args, **kwargs)
def draw_time_series(self, *args, **kwargs):
"""
Draw a time series visualization on this :class:`~Drawable`.
The arguments and keyword arguments are those supported by :meth:`~text.annotation.TextAnnotation.draw` method.
:return: A tuple made up of the drawn plot and label.
:rtype: tuple
"""
self.timeseries = self.timeseries if self.timeseries is not None else TimeSeries(
self)
return self.timeseries.draw(*args, **kwargs)
def test_iters(self):
ts = TimeSeries([1, 2, 3], [4, 5, 6])
times = ts.times()
count = 0
for item in ts:
self.assertEqual(item, ts[times[count]])
count += 1
self.assertEqual(count, len(ts))
count = 0
for item in ts.itervalues():
self.assertEqual(item, ts[times[count]])
count += 1
self.assertEqual(count, len(ts))
count = 0
for item in ts.itertimes():
self.assertEqual(item, times[count])
count += 1
self.assertEqual(count, len(ts))
count = 0
for item in ts.iteritems():
self.assertEqual(item, (times[count], ts[times[count]]))
count += 1
self.assertEqual(count, len(ts))
def random_ts(self, a):
'''
Randomly generate a time series with 100 time-value points. The time of the time series
are 100 equally spaced values between 0 and 1. The values of the time series
are 100 randomly generated values between 0 and a.
Param:
a: int, a multiplicative constant to generate 100 values for the time series
Return:
A time series with 100 time-value points
'''
t = np.arange(0.0, 1.0, 0.01)
v = a * np.random.random(100)
# return ts.TimeSeries(t, v)
return TimeSeries(t, v)
def stand(self, x, m, s):
'''
Standardize a variable x, using its mean m and its standard deviation s.
Param:
x : TimeSeries, the time series variable to standardize
m : float, the time series values' mean
s : float, the time series values' standard deviation
Return:
The standardized time series.
'''
time = x._times
value = x._data
for i in range(len(value)):
value[i] = (value[i] - m) / float(s)
return TimeSeries(time, value)
def test_similarity():
num_ts = 50
num_vp = 20
generate_ts(num_ts)
#with raises(Exception):
generateDB(num_ts, num_vp)
#curr_dir = os.getcwd().split('/')
#print(curr_dir)
file_name = 'vp_origin_idx.dat'
vantage_point_index_list = np.loadtxt(file_name, delimiter=',')
x = np.loadtxt('tsfiles/ts_1.dat', delimiter=',') #####
query_ts = TimeSeries(x[:, 0], x[:, 1])
nearest_ts = vp_similarity_search(query_ts, list(vantage_point_index_list))
#print(nearest_ts)
assert len(nearest_ts) == 10 and nearest_ts[0] == 'ts_1'
def tsmaker(self, m, s, j):
'''
Generate a time series with 100 time-value points. The time of the time series
are 100 equally spaced values between 0 and 1. The values of the time series
are 100 values generated from a normal distribution with mean s and standard
deviation s and random noises from 0 and j.
Param:
m : float, the time series values' mean
s : float, the time series values' standard deviation
j : float, the multiplicative constant to generate noises
Return:
TimeSeries, a time series with 100 time-value points
'''
t = np.arange(0.0, 1.0, 0.01)
v = norm.pdf(t, m, s) + j * np.random.randn(100)
# return meta, ts.TimeSeries(t, v)
return TimeSeries(t, v)
def test_shifted(self):
t = shifted(TimeSeries(range(5), range(5)), 1)
self.assertEqual(t.times, [0, 1, 2, 3])
self.assertEqual(t.values, [1, 2, 3, 4])
def test_creation_from_dict(self):
t = TimeSeries({x: x for x in range(5)})
self.assertEqual(t.times, [0, 1, 2, 3, 4])
self.assertEqual(t.values, [0, 1, 2, 3, 4])
def test_bool(self):
self.assertEqual(bool(TimeSeries([1, 2, 3], [-1, 2, -4])), True)
self.assertEqual(bool(TimeSeries([1, 2, 3], [1, 2, 4])), True)
self.assertEqual(bool(TimeSeries([1, 2, 3], [0, 0, 0])), False)
def test_abs(self):
self.assertEqual(abs(TimeSeries([1, 2, 3], [-1, 2, -4])), np.sqrt(21))
self.assertEqual(abs(TimeSeries([1, 2, 3], [1, 2, 4])), np.sqrt(21))
self.assertEqual(abs(TimeSeries([1, 2, 3], [0, 0, 0])), 0)
with self.assertRaises(ValueError):
abs(TimeSeries([], []))
def test_median(self):
self.assertEqual(TimeSeries([1, 2, 3], [2, 2, 2]).median(), 2)
self.assertEqual(TimeSeries([1, 2, 3], [0, 2, 0]).median(), 0)
self.assertEqual(TimeSeries([1, 2, 3, 4], [0, 2, 2, 0]).median(), 1)
with self.assertRaises(ValueError):
TimeSeries([], []).median()
def test_sincle_slice_with_interpolation(self):
t = TimeSeries(range(2, 6, 2), range(2, 6, 2), interpolate=True)
self.assertEqual(t[0], (0, 0))
self.assertEqual(t[2], (2, 2))
self.assertEqual(t[3], (3, 3))
def test_single_slice(self):
t = TimeSeries(range(2, 6, 2), range(2, 6, 2))
self.assertEqual(t[0], (0, 0))
self.assertEqual(t[1], (1, 0))
self.assertEqual(t[2], (2, 2))
self.assertEqual(t[3], (3, 2))
def test_sincle_slice_with_first_val(self):
t = TimeSeries(range(2, 6, 2), range(2, 6, 2), first_val=-1)
self.assertEqual(t[0], (0, -1))
self.assertEqual(t[1], (1, -1))
self.assertEqual(t[2], (2, 2))
def test_aligned_subtract(self):
t = TimeSeries(range(5), range(5)) - TimeSeries(range(5), range(5))
self.assertEqual(t.times, [0, 1, 2, 3, 4])
self.assertEqual(t.values, [0, 0, 0, 0, 0])
def test_misaligned_add(self):
t = TimeSeries([0, 2, 4], [0, 2, 4]) + TimeSeries([1, 2, 3], [1, 2, 3])
self.assertEqual(t.times, [0, 1, 2, 3, 4])
self.assertEqual(t.values, [0, 1, 4, 5, 7])
def test_creation_from_lists(self):
t = TimeSeries(range(5), range(5))
self.assertEqual(t.times, [0, 1, 2, 3, 4])
self.assertEqual(t.values, [0, 1, 2, 3, 4])
def test_pruned(self):
t = pruned(TimeSeries(range(5), range(5)), 2)
self.assertEqual(t.times, [0, 2, 4])
self.assertEqual(t.values, [0, 2, 4])
def test_tilted(self):
t = tilted(TimeSeries(range(5), range(5)), 8)
self.assertEqual(t.times, [0, 1, 2, 3, 4])
self.assertEqual(t.values, [0, 2, 4, 6, 8])
def test_aligned_add(self):
t = TimeSeries(range(5), range(5)) + TimeSeries(range(5), range(5))
self.assertEqual(t.times, [0, 1, 2, 3, 4])
self.assertEqual(t.values, [0, 2, 4, 6, 8])
def test_range_slice_with_interpolation(self):
t = TimeSeries(range(2, 10, 2), range(2, 10, 2), interpolate=True)
t_slice = t[3:6]
self.assertEqual(t_slice.times, [3, 4])
self.assertEqual(t_slice.values, [3, 4])