async def get_shaped_timeseries(self, query: ShapeQuery) -> ModelResult:
logger.debug(
"\n--->>> Shape Query Called successfully on %s Model!! <<<---" % self.name)
# logger.debug("Spatial Component is: \n%s" % str(query.spatial))
# logger.debug("Temporal Component is: \n%s" % str(query.temporal))
#
# logger.debug("\nDerived LAT1: %s\nDerived LON1: %s\nDerived LAT2: %s\nDerived LON2: %s" %
# query.spatial.expanded(0.05))
sr = await (self.get_shaped_resultcube(query))
sr.load()
dps = []
for r in sr['time']:
t = r['time'].values
o = sr.sel(time=t)
p = self.outputs['readings']['prefix']
df = o[p].to_dataframe()
df = df[p]
# TODO - This is a quick hack to massage the datetime format into a markup suitable for D3 & ngx-charts!
m = df.mean()
dps.append(DataPoint(observation_time=str(t).replace('.000000000', '.000Z'),
value=m,
mean=m,
minimum=df.min(),
maximum=df.max(),
deviation=0))
asyncio.sleep(1)
return ModelResult(model_name=self.name, data_points=dps)
async def get_datapoint_for_param(b, param):
"""
Takes the mean min and max values for datapoints at a particular time slice.
:param b:
:param param:
:return:
"""
bin_ = b.to_dataframe()
# TODO - This is a quick hack to massage the datetime format into a markup suitable for D3 & ngx-charts!
tvalue = str(b["time"].values).replace('.000000000', '.000Z')
avalue = bin_[param].median()
logger.debug(
"\n>>>> Datapoint creation. (time={}, value={})".format(tvalue, avalue))
asyncio.sleep(1)
return DataPoint(observation_time=tvalue,
value=avalue,
mean=bin_[param].mean(),
minimum=bin_[param].min(),
maximum=bin_[param].max(),
deviation=bin_[param].std())
def dummy_single(v):
# Dummy data for testing...
# value, mean, min, max, std
five_values = [np.random.random_sample() for j in range(5)]
five_values = pd.DataFrame(five_values)
# print(five_values)
return DataPoint(
dt.date(2018, 1, v + 1).strftime("%Y-%m-%dT00:00:00.000Z"),
five_values[0].mean(), five_values[0].mean(), five_values[0].min(),
five_values[0].max(), five_values[0].std())
def dummy_series():
response = []
for i in range(1, 30):
# Dummy data for testing...
# value, mean, min, max, std
five_values = [np.random.random_sample() for j in range(5)]
five_values = pd.DataFrame(five_values)
# print(five_values)
response.append(
DataPoint(
dt.date(2018, 1, i).strftime("%Y-%m-%dT00:00:00.000Z"),
five_values[0].mean(), five_values[0].mean(),
five_values[0].min(), five_values[0].max(),
five_values[0].std()))
return response
def subscribe(self, observer):
if dev.DEBUG:
dps = []
for i in range(3):
dps.append(
DataPoint(observation_time=dt.date(2018, 5, 17),
value=i,
mean=i,
minimum=i,
maximum=i,
deviation=0))
observer.on_next(
ModelResult(model_name='test', data_points=dps))
observer.on_completed()
else:
dps = []
pass
async def get_shaped_timeseries(self, query: ShapeQuery) -> ModelResult:
logger.debug(
"\n--->>> Shape Query Called successfully on %s Model!! <<<---" % self.name)
logger.debug("Spatial Component is: \n%s" % str(query.spatial))
logger.debug("Temporal Component is: \n%s" % str(query.temporal))
logger.debug("\nDerived LAT1: %s\nDerived LON1: %s\nDerived LAT2: %s\nDerived LON2: %s" %
query.spatial.expanded(0.05))
dps = []
try:
sr = await (self.get_shaped_resultcube(query))
sr.load()
# Values in AWRA are in range 0..1
# Normalise to between 0-100%
# sr[self.outputs["readings"]["prefix"]] *= 100
if dev.DEBUG:
logger.debug("Shaped ResultCube is: \n%s" % sr)
for r in sr['time']:
t = r['time'].values
o = sr.sel(time=t)
p = self.outputs['readings']['prefix']
df = o[p].to_dataframe()
df = df[p]
# TODO - This is a quick hack to massage the datetime format into a markup suitable for D3 & ngx-charts!
m = df.median()
dps.append(DataPoint(observation_time=str(t).replace('.000000000', '.000Z'),
value=m,
mean=df.mean(),
minimum=df.min(),
maximum=df.max(),
deviation=0))
except FileNotFoundError:
logger.exception('Files not found for date range.')
asyncio.sleep(1)
return ModelResult(model_name=self.name, data_points=dps)