def test_timeseries_segments(client, timeseries):
"""
No valid timeseries data to test against, but we can test the API calls.
"""
channels = [
TimeSeriesChannel(name='ch 1', rate=256, unit='uV'),
TimeSeriesChannel(name='ch 2', rate=111.123, unit='uV'),
]
timeseries.add_channels(*channels)
for ch in timeseries.channels:
ch.end = ch.start + 5e6
ch.update()
ch.segments()
ch.segments(start=ch.start + 1e6)
ch.segments(stop=ch.end - 1e6)
ch.segments(stop=ch.end - 1e6, gap_factor=3)
ch.segments(stop=ch.end - 1e6, gap_factor=33.33)
ch.segments(stop=ch.end - 1e6, gap_factor=0.5)
timeseries.segments()
timeseries.segments(gap_factor=3)
timeseries.segments(gap_factor=33.33)
timeseries.segments(gap_factor=0.5)
# TODO: API should soon error when value is non-numeric.
# Uncomment below when API is changed.
with pytest.raises(Exception):
timeseries.segments(gap_factor="should be int")
with pytest.raises(Exception):
ch.segments(gap_factor="should be int")
def test_timeseries_annotations_check_that_channels_belong_to_series(
timeseries, timeseries2):
ch = TimeSeriesChannel(name="test_channel",
rate=256,
unit="uV",
start=0,
end=1e6)
timeseries.add_channels(ch)
print(timeseries, timeseries2)
ch2 = TimeSeriesChannel(name="test_channel_2",
rate=256,
unit="uV",
start=0,
end=1e6)
timeseries2.add_channels(ch2)
layer = TimeSeriesAnnotationLayer(name="a_layer",
time_series_id=timeseries.id,
description="a layer")
timeseries.add_layer(layer)
with pytest.raises(Exception):
layer.annotations(channels=[ch, ch2])
with pytest.raises(Exception):
next(layer.iter_annotations(channels=[ch, ch2]))
with pytest.raises(Exception):
layer.annotation_counts(0, 1e6, period="0.1s", channels=[ch, ch2])
def test_timeseries_channels(client, timeseries):
num_channels = 10
assert timeseries.exists
ts2 = client.get(timeseries.id)
chs = []
for i in range(num_channels):
# init
chname = 'Channel-{}'.format(i)
ch = TimeSeriesChannel(name=chname, rate=256, unit='uV')
assert not ch.exists
# create
timeseries.add_channels(ch)
assert ch.exists
assert ch.name == chname
chs.append(ch)
# use separate request to get channel
ch.insert_property('key', 'value')
ch2 = [x for x in ts2.channels if x.id == ch.id][0]
assert ch2.get_property('key') is not None
assert ch2.name == ch.name
assert ch2.type == ch.type
assert ch2.rate == ch.rate
assert ch2.unit == ch.unit
del ch2
# update channel
ch.name = '{}-updated'.format(ch.name)
ch.rate = 200
ch.update()
# use separate request to confirm name change
ch2 = [x for x in ts2.channels if x.id == ch.id][0]
assert ch2.name == ch.name
assert ch2.rate == ch.rate
# ensure correct number of channels
channels = timeseries.channels
assert len(channels) == num_channels
# get channels (via API)
ch_ids = [x.id for x in channels]
# check
for ch in chs:
assert ch.id in ch_ids
# TODO: remove channel
ch = channels[0]
timeseries.remove_channels(ch)
channels = timeseries.channels
assert len(channels) == num_channels - 1
assert ch not in channels
assert ch.id not in [x.id for x in channels]
def test_timeseries_annotations_can_get_more_than_default_limit(timeseries):
ch = TimeSeriesChannel(name="test_channel",
rate=256,
unit="uV",
start=0,
end=2e6)
timeseries.add_channels(ch)
layer = TimeSeriesAnnotationLayer(
name="limit_layer",
time_series_id=timeseries.id,
description="layer with many small annotations",
)
timeseries.add_layer(layer)
for i in range(200):
annotation = TimeSeriesAnnotation(
label="annotation_{}".format(i),
channel_ids=ch.id,
start=i * 1e4,
end=(i + 1) * 1e4,
)
layer.add_annotations(annotation)
# Should retrieve all annotations
assert len(layer.annotations()) == 200
# All of these annotations should be in the first 10-second window
annots = next(layer.iter_annotations())
assert len(annots) == 200
layer.delete()
def test_stream_upload(use_dev, client, dataset):
# can only run against dev server :-(
if not use_dev: return
# generate data
freq = 100
df = generate_dataframe(minutes=1, freq=100)
# create timeseries
ts = TimeSeries('My Test TimeSeries')
dataset.add(ts)
assert ts.exists
print(ts)
# create channels
channels = [TimeSeriesChannel(c, rate=freq) for c in df.columns]
ts.add_channels(*channels)
for ch in channels:
assert ch.exists
assert ch.rate == freq
assert ch.name in df.columns
print("channels =", ts.channels)
# stream data up
ts.stream_data(df)
# wait a bit -- let server write data
time.sleep(10)
# retrieve data
for chunk in ts.get_data_iter():
assert len(chunk) > 0
assert chunk.columns == df.columns
assert chunk.index[-1] == df.index[-1]
# make sure channel times were updated
for ch in ts.channels:
assert ch.start_datetime == df.index[0]
assert ch.end_datetime == df.index[-1]
def test_timeseries_annotations(client, timeseries):
assert timeseries.exists
print 'layers = ', timeseries.layers
#Create Layer
layer1 = TimeSeriesAnnotationLayer(name="test_layer", time_series_id = timeseries.id, description="test_description")
a = layer1.as_dict()
assert a['name'] == 'test_layer'
assert a['description'] == 'test_description'
# Add Layer
timeseries.add_layer(layer1)
assert layer1.exists
# Get Layer
layer1b = timeseries.get_layer('test_layer')
assert layer1b.exists
assert layer1b.name == "test_layer"
assert layer1.id == layer1b.id
assert layer1._api.timeseries is not None
# Add another layer
layer2 = timeseries.add_layer('test_layer2','test_description2')
assert layer2.exists
layer2copy = timeseries.add_layer('test_layer2')
assert layer2copy.id == layer2.id
# Get Layer
layer2b = timeseries.get_layer('test_layer2')
assert layer2b.exists
assert layer2b.name == "test_layer2"
assert layer2.id == layer2b.id
assert layer2._api.timeseries is not None
layers = timeseries.layers
assert len(layers)==2
assert layers[0].name == "test_layer"
assert layers[1].name == 'test_layer2'
#Create channels
ch = TimeSeriesChannel(
name='test_channel',
rate=256,
unit='uV',
start=1,
end=60*1e6)
#Create annotation over one channel
# create
timeseries.add_channels(ch)
assert ch.exists
assert ch.id in [x.id for x in timeseries.channels]
annot = TimeSeriesAnnotation(label = 'test_label', channel_ids = timeseries.channels[0].id,start = timeseries.channels[0].start,end=timeseries.channels[0].start+1*1e6)
#Add Annotation
layer1.add_annotations(annot)
assert annot.exists
# get annotations
annotb = layer1.annotations()
print "annotations = ", annotb
assert annotb[0].label == annot.label
#Create annotation over multiple channels
ch2 = TimeSeriesChannel(
name='test_channel',
rate=256,
unit='uV',
start=1,
end=60*1e6)
timeseries.add_channels(ch2)
channels = timeseries.channels
ch_ids = [x.id for x in channels]
assert ch2.exists
assert ch2.id in ch_ids
assert ch.id in ch_ids
for ch in channels:
assert ch.rate == 256
assert ch.exists
#add annotation over two channels
channel_ids = [timeseries.channels[x].id for x in range(len(timeseries.channels))]
annot2 = layer1.insert_annotation(annotation='test_label2', channel_ids = channel_ids,
start =timeseries.channels[0].start+1*1e6,
end = timeseries.channels[0].start+2*1e6)
assert annot2.exists
annot_gen = layer1.iter_annotations(1)
annot = annot_gen.next()
assert annot[0].label == 'test_label'
next_annot= annot_gen.next()
assert next_annot[0].label == 'test_label2'
### TEST DELETION
annot3 = TimeSeriesAnnotation(label= 'test_label3', channel_ids= channel_ids,start = timeseries.channels[0].start+1*1e6,end=timeseries.channels[0].start+2*1e6)
#layer1.add_annotations([annot2,annot3])
annot3 = timeseries.add_annotations(layer=layer1,annotations=annot3)
assert annot3.exists
annot3.delete()
assert not annot3.exists
annot4 = timeseries.insert_annotation(layer=layer1,annotation='test_label3',start= timeseries.channels[0].start+1*1e6,end=timeseries.channels[0].start+2*1e6)
assert annot4.exists
annot4.delete()
assert not annot4.exists
annot5 = timeseries.insert_annotation(layer='test_layer4',annotation='test_label3',start=timeseries.channels[0].start+1*1e6,end=timeseries.channels[0].start+2*1e6)
assert annot5.exists
annot5.delete()
assert not annot5.exists
layer1.add_annotations([annot2,annot3])
assert annot2.exists
assert annot3.exists
#test datetime input
annot4 = timeseries.insert_annotation(layer='test_layer4',annotation='test_label3',start= datetime.datetime.utcfromtimestamp((timeseries.channels[0].start+1*1e6)/1e6),end=datetime.datetime.utcfromtimestamp(timeseries.channels[0].start+2*1e6))
assert annot4.exists
annot4.delete()
assert not annot4.exists
layer = timeseries.get_layer('test_layer4')
assert layer.exists
layer.delete()
assert not layer.exists
# delete annotations
annot[0].delete()
assert not annot[0].exists
assert timeseries.exists
timeseries.delete_layer(layer1)
assert not layer1.exists
assert layer2.exists
layer2.delete()
assert not layer2.exists
def test_timeseries_annotations(client, timeseries):
assert timeseries.exists
print("layers = ", timeseries.layers)
# Create Layer
layer1 = TimeSeriesAnnotationLayer(name="test_layer",
time_series_id=timeseries.id,
description="test_description")
a = layer1.as_dict()
assert a["name"] == "test_layer"
assert a["description"] == "test_description"
# Add Layer
timeseries.add_layer(layer1)
assert layer1.exists
# Get Layer
layer1b = timeseries.get_layer("test_layer")
assert layer1b.exists
assert layer1b.name == "test_layer"
assert layer1.id == layer1b.id
assert layer1._api.timeseries is not None
# Add another layer
layer2 = timeseries.add_layer("test_layer2", "test_description2")
assert layer2.exists
layer2copy = timeseries.add_layer("test_layer2")
assert layer2copy.id == layer2.id
# Get Layer
layer2b = timeseries.get_layer("test_layer2")
assert layer2b.exists
assert layer2b.name == "test_layer2"
assert layer2.id == layer2b.id
assert layer2._api.timeseries is not None
layers = timeseries.layers
assert len(layers) == 2
assert set([layers[0].name,
layers[1].name]) == set(["test_layer", "test_layer2"])
# Create channels
ch = TimeSeriesChannel(name="test_channel",
rate=256,
unit="uV",
start=1,
end=60 * 1e6)
# Create annotation over one channel
# create
timeseries.add_channels(ch)
assert ch.exists
assert ch.id in [x.id for x in timeseries.channels]
annot = TimeSeriesAnnotation(
label="test_label",
channel_ids=timeseries.channels[0].id,
start=timeseries.channels[0].start,
end=timeseries.channels[0].start + 1 * 1e6,
)
# Add Annotation
layer1.add_annotations(annot)
assert annot.exists
# get annotations
annotb = layer1.annotations()
assert annotb[0].label == annot.label
annotc = client._api.timeseries.get_annotation(timeseries, layer1, annot)
assert annotc.label == annot.label
# Create annotation over multiple channels
ch2 = TimeSeriesChannel(name="test_channel",
rate=256,
unit="uV",
start=1,
end=60 * 1e6)
timeseries.add_channels(ch2)
channels = timeseries.channels
ch_ids = [x.id for x in channels]
assert ch2.exists
assert ch2.id in ch_ids
assert ch.id in ch_ids
for ch in channels:
assert ch.rate == 256
assert ch.exists
# add annotation over two channels
channel_ids = [
timeseries.channels[x].id for x in range(len(timeseries.channels))
]
annot2 = layer1.insert_annotation(
annotation="test_label2",
channel_ids=channel_ids,
start=timeseries.channels[0].start + 1 * 1e6,
end=timeseries.channels[0].start + 2 * 1e6,
)
assert annot2.exists
annot_gen = layer1.iter_annotations(1)
annot = next(annot_gen)
assert annot[0].label == "test_label"
next_annot = next(annot_gen)
assert next_annot[0].label == "test_label2"
### TEST ANNOTATION COUNTS
layer1_expected_counts = [
{
"start": 0,
"end": 250000,
"value": 1.0
},
{
"start": 250000,
"end": 500000,
"value": 1.0
},
{
"start": 500000,
"end": 750000,
"value": 1.0
},
{
"start": 750000,
"end": 1000000,
"value": 1.0
},
{
"start": 1000000,
"end": 1250000,
"value": 2.0
},
{
"start": 1250000,
"end": 1500000,
"value": 1.0
},
{
"start": 1500000,
"end": 1750000,
"value": 1.0
},
{
"start": 1750000,
"end": 2000000,
"value": 1.0
},
{
"start": 2000000,
"end": 2000001,
"value": 1.0
},
]
def _sort_counts(counts):
return sorted(counts, key=lambda c: c["start"])
assert (_sort_counts(
timeseries.annotation_counts(
start=timeseries.channels[0].start * 1e6,
end=timeseries.channels[0].start + 2 * 1e6,
layers=[layer1],
period="0.25s",
)[str(layer1.id)]) == layer1_expected_counts)
assert (_sort_counts(
layer1.annotation_counts(
start=timeseries.channels[0].start * 1e6,
end=timeseries.channels[0].start + 2 * 1e6,
period="0.25s",
)[str(layer1.id)]) == layer1_expected_counts)
### TEST DELETION
annot3 = TimeSeriesAnnotation(
label="test_label3",
channel_ids=channel_ids,
start=timeseries.channels[0].start + 1 * 1e6,
end=timeseries.channels[0].start + 2 * 1e6,
)
layer1.add_annotations([annot2, annot3])
annot3 = timeseries.add_annotations(layer=layer1, annotations=annot3)
assert annot3.exists
annot3.delete()
assert not annot3.exists
annot4 = timeseries.insert_annotation(
layer=layer1,
annotation="test_label3",
start=timeseries.channels[0].start + 1 * 1e6,
end=timeseries.channels[0].start + 2 * 1e6,
)
assert annot4.exists
annot4.delete()
assert not annot4.exists
annot5 = timeseries.insert_annotation(
layer="test_layer4",
annotation="test_label3",
start=timeseries.channels[0].start + 1 * 1e6,
end=timeseries.channels[0].start + 2 * 1e6,
)
assert annot5.exists
annot5.delete()
assert not annot5.exists
layer1.add_annotations([annot2, annot3])
assert annot2.exists
assert annot3.exists
# test datetime input
annot4 = timeseries.insert_annotation(
layer="test_layer4",
annotation="test_label3",
start=datetime.datetime.utcfromtimestamp(
(timeseries.channels[0].start + 1 * 1e6) / 1e6),
end=datetime.datetime.utcfromtimestamp(timeseries.channels[0].start +
2 * 1e6),
)
assert annot4.exists
annot4.delete()
assert not annot4.exists
layer = timeseries.get_layer("test_layer4")
assert layer.exists
layer.delete()
assert not layer.exists
# delete annotations
annot[0].delete()
assert not annot[0].exists
assert timeseries.exists
timeseries.delete_layer(layer1)
assert not layer1.exists
assert layer2.exists
layer2.delete()
assert not layer2.exists
def test_timeseries_annotations(client, timeseries):
assert timeseries.exists
print('layers = ', timeseries.layers)
#Create Layer
layer1 = TimeSeriesAnnotationLayer(name="test_layer",
time_series_id=timeseries.id,
description="test_description")
a = layer1.as_dict()
assert a['name'] == 'test_layer'
assert a['description'] == 'test_description'
# Add Layer
timeseries.add_layer(layer1)
assert layer1.exists
# Get Layer
layer1b = timeseries.get_layer('test_layer')
assert layer1b.exists
assert layer1b.name == "test_layer"
assert layer1.id == layer1b.id
assert layer1._api.timeseries is not None
# Add another layer
layer2 = timeseries.add_layer('test_layer2', 'test_description2')
assert layer2.exists
layer2copy = timeseries.add_layer('test_layer2')
assert layer2copy.id == layer2.id
# Get Layer
layer2b = timeseries.get_layer('test_layer2')
assert layer2b.exists
assert layer2b.name == "test_layer2"
assert layer2.id == layer2b.id
assert layer2._api.timeseries is not None
layers = timeseries.layers
assert len(layers) == 2
assert layers[0].name == "test_layer"
assert layers[1].name == 'test_layer2'
#Create channels
ch = TimeSeriesChannel(name='test_channel',
rate=256,
unit='uV',
start=1,
end=60 * 1e6)
#Create annotation over one channel
# create
timeseries.add_channels(ch)
assert ch.exists
assert ch.id in [x.id for x in timeseries.channels]
annot = TimeSeriesAnnotation(label='test_label',
channel_ids=timeseries.channels[0].id,
start=timeseries.channels[0].start,
end=timeseries.channels[0].start + 1 * 1e6)
#Add Annotation
layer1.add_annotations(annot)
assert annot.exists
# get annotations
annotb = layer1.annotations()
assert annotb[0].label == annot.label
annotc = client._api.timeseries.get_annotation(timeseries, layer1, annot)
assert annotc.label == annot.label
#Create annotation over multiple channels
ch2 = TimeSeriesChannel(name='test_channel',
rate=256,
unit='uV',
start=1,
end=60 * 1e6)
timeseries.add_channels(ch2)
channels = timeseries.channels
ch_ids = [x.id for x in channels]
assert ch2.exists
assert ch2.id in ch_ids
assert ch.id in ch_ids
for ch in channels:
assert ch.rate == 256
assert ch.exists
#add annotation over two channels
channel_ids = [
timeseries.channels[x].id for x in range(len(timeseries.channels))
]
annot2 = layer1.insert_annotation(
annotation='test_label2',
channel_ids=channel_ids,
start=timeseries.channels[0].start + 1 * 1e6,
end=timeseries.channels[0].start + 2 * 1e6)
assert annot2.exists
annot_gen = layer1.iter_annotations(1)
annot = next(annot_gen)
assert annot[0].label == 'test_label'
next_annot = next(annot_gen)
assert next_annot[0].label == 'test_label2'
### TEST ANNOTATION COUNTS
layer1_expected_counts = [{
'start': 0,
'end': 250000,
'value': 1.0
}, {
'start': 250000,
'end': 500000,
'value': 1.0
}, {
'start': 500000,
'end': 750000,
'value': 1.0
}, {
'start': 750000,
'end': 1000000,
'value': 1.0
}, {
'start': 1000000,
'end': 1250000,
'value': 2.0
}, {
'start': 1250000,
'end': 1500000,
'value': 1.0
}, {
'start': 1500000,
'end': 1750000,
'value': 1.0
}, {
'start': 1750000,
'end': 2000000,
'value': 1.0
}, {
'start': 2000000,
'end': 2000001,
'value': 1.0
}]
def _sort_counts(counts):
return sorted(counts, key=lambda c: c['start'])
assert _sort_counts(
timeseries.annotation_counts(
start=timeseries.channels[0].start * 1e6,
end=timeseries.channels[0].start + 2 * 1e6,
layers=[layer1],
period="0.25s")[str(layer1.id)]) == layer1_expected_counts
assert _sort_counts(
layer1.annotation_counts(start=timeseries.channels[0].start * 1e6,
end=timeseries.channels[0].start + 2 * 1e6,
period="0.25s")[str(
layer1.id)]) == layer1_expected_counts
### TEST DELETION
annot3 = TimeSeriesAnnotation(label='test_label3',
channel_ids=channel_ids,
start=timeseries.channels[0].start + 1 * 1e6,
end=timeseries.channels[0].start + 2 * 1e6)
layer1.add_annotations([annot2, annot3])
annot3 = timeseries.add_annotations(layer=layer1, annotations=annot3)
assert annot3.exists
annot3.delete()
assert not annot3.exists
annot4 = timeseries.insert_annotation(
layer=layer1,
annotation='test_label3',
start=timeseries.channels[0].start + 1 * 1e6,
end=timeseries.channels[0].start + 2 * 1e6)
assert annot4.exists
annot4.delete()
assert not annot4.exists
annot5 = timeseries.insert_annotation(
layer='test_layer4',
annotation='test_label3',
start=timeseries.channels[0].start + 1 * 1e6,
end=timeseries.channels[0].start + 2 * 1e6)
assert annot5.exists
annot5.delete()
assert not annot5.exists
layer1.add_annotations([annot2, annot3])
assert annot2.exists
assert annot3.exists
# test datetime input
annot4 = timeseries.insert_annotation(
layer='test_layer4',
annotation='test_label3',
start=datetime.datetime.utcfromtimestamp(
(timeseries.channels[0].start + 1 * 1e6) / 1e6),
end=datetime.datetime.utcfromtimestamp(timeseries.channels[0].start +
2 * 1e6))
assert annot4.exists
annot4.delete()
assert not annot4.exists
layer = timeseries.get_layer('test_layer4')
assert layer.exists
layer.delete()
assert not layer.exists
# delete annotations
annot[0].delete()
assert not annot[0].exists
assert timeseries.exists
timeseries.delete_layer(layer1)
assert not layer1.exists
assert layer2.exists
layer2.delete()
assert not layer2.exists