def test_standard_profile(self, elev, drag, burst, wind, const, linear):
wind_ds = Mock()
elev_ds = Mock()
const.return_value = 'const'
wind.return_value = 'wind'
linear.return_value = 'linear'
burst.return_value = 'burst'
drag.return_value = 'drag'
elev.return_value = 'elev'
model = models.standard_profile(5.0, 30000.0, 6.0, wind_ds, elev_ds)
const.assert_called_with(5.0)
wind.assert_called_with(wind_ds)
linear.assert_any_call(['const', 'wind'])
burst.assert_called_with(30000.0)
drag.assert_called_with(6.0)
linear.assert_called_with(['drag', 'wind'])
elev.assert_called_with(elev_ds)
assert_equal(model, (('linear', 'burst'), ('linear', 'elev')))
def test_standard_profile(self, elev, drag, burst, wind, const, linear):
wind_ds = Mock()
elev_ds = Mock()
const.return_value = 'const'
wind.return_value = 'wind'
linear.return_value = 'linear'
burst.return_value = 'burst'
drag.return_value = 'drag'
elev.return_value = 'elev'
model = models.standard_profile(5.0, 30000.0, 6.0, wind_ds, elev_ds)
const.assert_called_with(5.0)
wind.assert_called_with(wind_ds)
linear.assert_any_call(['const', 'wind'])
burst.assert_called_with(30000.0)
drag.assert_called_with(6.0)
linear.assert_called_with(['drag', 'wind'])
elev.assert_called_with(elev_ds)
assert_equal(model, (('linear', 'burst'), ('linear', 'elev')))
def run_prediction(req):
"""
Run the prediction.
"""
# Response dict
resp = {
"request": req,
"prediction": [],
}
# Dataset
try:
if req['dataset'] == LATEST_DATASET_KEYWORD:
tawhiri_ds = WindDataset.open_latest(persistent=True)
else:
tawhiri_ds = WindDataset(datetime.fromtimestamp(req['dataset']))
except IOError:
raise InvalidDatasetException("No matching dataset found.")
except ValueError as e:
raise InvalidDatasetException(*e.args)
# Note that hours and minutes are set to 00 as Tawhiri uses hourly datasets
resp['request']['dataset'] = tawhiri_ds.ds_time.strftime(
"%Y-%m-%dT%H:00:00Z")
# Stages
if req['profile'] == PROFILE_STANDARD:
stages = models.standard_profile(req['ascent_rate'],
req['burst_altitude'],
req['descent_rate'], tawhiri_ds,
ruaumoko_ds())
elif req['profile'] == PROFILE_FLOAT:
stages = models.float_profile(req['ascent_rate'],
req['float_altitude'],
req['stop_datetime'], tawhiri_ds)
else:
raise InternalException("No implementation for known profile.")
# Run solver
try:
result = solver.solve(req['launch_datetime'], req['launch_latitude'],
req['launch_longitude'], req['launch_altitude'],
stages)
except Exception as e:
raise PredictionException("Prediction did not complete: '%s'." %
str(e))
# Format trajectory
if req['profile'] == PROFILE_STANDARD:
resp['prediction'] = _parse_stages(["ascent", "descent"], result)
elif req['profile'] == PROFILE_FLOAT:
resp['prediction'] = _parse_stages(["ascent", "float"], result)
else:
raise InternalException("No implementation for known profile.")
# Convert request UNIX timestamps to RFC3339 timestamps
for key in resp['request']:
if "datetime" in key:
resp['request'][key] = _timestamp_to_rfc3339(resp['request'][key])
return resp
import json
import calendar
from tawhiri import solver, models, kml
from tawhiri.dataset import Dataset as WindDataset
from ruaumoko import Dataset as ElevationDataset
lat0 = 52.5563
lng0 = 360 - 3.1970
alt0 = 0.0
t0 = calendar.timegm(datetime(2014, 2, 19, 15).timetuple())
wind = WindDataset.open_latest()
elevation = ElevationDataset()
stages = models.standard_profile(5.0, 30000, 5.0, wind, elevation)
rise, fall = solver.solve(t0, lat0, lng0, alt0, stages)
assert rise[-1] == fall[0]
with open("test_prediction_data.js", "w") as f:
f.write("var data = ")
json.dump([(lat, lon) for _, lat, lon, _ in rise + fall], f, indent=4)
f.write(";\n")
markers = [{
'name': 'launch',
'description': 'TODO',
'point': rise[0]
}, {
'name': 'landing',
def run_prediction(req):
"""
Run the prediction.
"""
# Response dict
resp = {
"request": req,
"prediction": [],
}
warningcounts = WarningCounts()
# Find wind data location
ds_dir = app.config.get('WIND_DATASET_DIR', WindDataset.DEFAULT_DIRECTORY)
# Dataset
try:
if req['dataset'] == LATEST_DATASET_KEYWORD:
tawhiri_ds = WindDataset.open_latest(persistent=True, directory=ds_dir)
else:
tawhiri_ds = WindDataset(datetime.fromtimestamp(req['dataset']), directory=ds_dir)
except IOError:
raise InvalidDatasetException("No matching dataset found.")
except ValueError as e:
raise InvalidDatasetException(*e.args)
# Note that hours and minutes are set to 00 as Tawhiri uses hourly datasets
resp['request']['dataset'] = \
tawhiri_ds.ds_time.strftime("%Y-%m-%dT%H:00:00Z")
# Stages
if req['profile'] == PROFILE_STANDARD:
stages = models.standard_profile(req['ascent_rate'],
req['burst_altitude'],
req['descent_rate'],
tawhiri_ds,
ruaumoko_ds(),
warningcounts)
elif req['profile'] == PROFILE_FLOAT:
stages = models.float_profile(req['ascent_rate'],
req['float_altitude'],
req['stop_datetime'],
tawhiri_ds,
warningcounts)
else:
raise InternalException("No implementation for known profile.")
# Run solver
try:
result = solver.solve(req['launch_datetime'], req['launch_latitude'],
req['launch_longitude'], req['launch_altitude'],
stages)
except Exception as e:
raise PredictionException("Prediction did not complete: '%s'." %
str(e))
# Format trajectory
if req['profile'] == PROFILE_STANDARD:
resp['prediction'] = _parse_stages(["ascent", "descent"], result)
elif req['profile'] == PROFILE_FLOAT:
resp['prediction'] = _parse_stages(["ascent", "float"], result)
else:
raise InternalException("No implementation for known profile.")
# Convert request UNIX timestamps to RFC3339 timestamps
for key in resp['request']:
if "datetime" in key:
resp['request'][key] = _timestamp_to_rfc3339(resp['request'][key])
resp["warnings"] = warningcounts.to_dict()
return resp
import time
from datetime import timedelta
import calendar
from tawhiri import solver, models
from tawhiri.dataset import Dataset as WindDataset
from tawhiri.warnings import WarningCounts
from ruaumoko import Dataset as ElevationDataset
elevation = ElevationDataset()
warningcounts = WarningCounts()
lat0 = 52.0
lng0 = 0.0
alt0 = 0.0
n_repeats = 100
start_time = time.time()
for i in range(n_repeats):
wind = WindDataset.open_latest(persistent=True)
t0 = wind.ds_time + timedelta(hours=12)
t0 = calendar.timegm(t0.timetuple())
stages = models.standard_profile(5.0, 30000, 5.0, wind, elevation, warningcounts)
result = solver.solve(t0, lat0, lng0, alt0, stages)
end_time = time.time()
print("Averaged {:.1f}ms per prediction".format(
((end_time - start_time)/n_repeats)*1000.0))
import json
import calendar
from tawhiri import solver, models, kml
from tawhiri.dataset import Dataset as WindDataset
from ruaumoko import Dataset as ElevationDataset
lat0 = 52.5563
lng0 = 360 - 3.1970
alt0 = 0.0
t0 = calendar.timegm(datetime(2014, 2, 19, 15).timetuple())
wind = WindDataset.open_latest()
elevation = ElevationDataset()
stages = models.standard_profile(5.0, 30000, 5.0, wind, elevation)
rise, fall = solver.solve(t0, lat0, lng0, alt0, stages)
assert rise[-1] == fall[0]
with open("test_prediction_data.js", "w") as f:
f.write("var data = ")
json.dump([(lat, lon) for _, lat, lon, _ in rise + fall], f, indent=4)
f.write(";\n")
markers = [
{'name': 'launch', 'description': 'TODO', 'point': rise[0]},
{'name': 'landing', 'description': 'TODO', 'point': fall[-1]},
{'name': 'burst', 'description': 'TODO', 'point': fall[0]}
]