def _get_eng_params(hdf, align_param=None):
'''
Get eng parameters from hdf, and return the minimum.
:param hdf: hdf_file object.
:type hdf: hdfaccess.file.hdf_file
:returns: Minimum of normalised split parameters along with its frequency.
Will return None, None if no split parameters are available.
:rtype: (None, None) or (np.ma.masked_array, float)
'''
params = []
eng_params= ('Eng (1) N1', 'Eng (2) N1', 'Eng (3) N1', 'Eng (4) N1',
'Eng (1) N2', 'Eng (2) N2', 'Eng (3) N2', 'Eng (4) N2',
'Eng (1) Np', 'Eng (2) Np', 'Eng (3) Np', 'Eng (4) Np')
for param_name in eng_params:
try:
param = hdf[param_name]
except KeyError:
continue
if align_param:
# Align all other parameters to provided param or first available.
param.array = align(param, align_param)
else:
align_param = param
params.append(param)
if not len(params):
return None, None
# If there is at least one split parameter available.
stacked_params = vstack_params(*params)
split_params_avg = np.ma.average(stacked_params, axis=0)
return split_params_avg, align_param.frequency
def get_params(self, hdf_path, _slice, phase_name):
import shutil
import tempfile
from hdfaccess.file import hdf_file
with tempfile.NamedTemporaryFile() as temp_file:
shutil.copy(hdf_path, temp_file.name)
with hdf_file(hdf_path) as hdf:
pitch_capt = hdf.get('Pitch (Capt)')
pitch_fo = hdf.get('Pitch (FO)')
roll_capt = hdf.get('Roll (Capt)')
roll_fo = hdf.get('Roll (FO)')
cc_capt = hdf.get('Control Column Force (Capt)')
cc_fo = hdf.get('Control Column Force (FO)')
for par in pitch_capt, pitch_fo, roll_capt, roll_fo, cc_capt, \
cc_fo:
if par is not None:
ref_par = par
break
phase = S(name=phase_name, frequency=1)
phase.create_section(_slice)
phase = phase.get_aligned(ref_par)[0]
# Align the arrays, usually done in the Nodes
for par in pitch_fo, roll_capt, roll_fo, cc_capt, cc_fo:
if par is None:
continue
par.array = align(par, ref_par)
par.hz = ref_par.hz
return pitch_capt, pitch_fo, roll_capt, roll_fo, cc_capt, cc_fo, phase
def _get_normalised_split_params(hdf):
'''
Get split parameters (currently engine power and Groundspeed) from hdf, normalise
them on a scale from 0-1.0 and return the minimum.
:param hdf: hdf_file object.
:type hdf: hdfaccess.file.hdf_file
:returns: Minimum of normalised split parameters along with its frequency. Will return None, None if no split parameters are available.
:rtype: (None, None) or (np.ma.masked_array, float)
'''
params = []
first_split_param = None
for param_name in ('Eng (1) N1', 'Eng (2) N1', 'Eng (3) N1', 'Eng (4) N1',
'Eng (1) N2', 'Eng (2) N2', 'Eng (3) N2', 'Eng (4) N2',
'Eng (1) Np', 'Eng (2) Np', 'Eng (3) Np', 'Eng (4) Np',
'Groundspeed', 'Groundspeed (1)', 'Groundspeed (2)'):
try:
param = hdf[param_name]
except KeyError:
continue
if first_split_param:
# Align all other parameters to first available. #Q: Why not force to 1Hz?
param.array = align(param, first_split_param)
else:
first_split_param = param
params.append(param)
if not first_split_param:
return None, None
# If there is at least one split parameter available.
# normalise the parameters we'll use for splitting the data
stacked_params = vstack_params(*params)
normalised_params = normalise(stacked_params, scale_max=100)
split_params_min = np.ma.average(normalised_params, axis=0)
return split_params_min, first_split_param.frequency
def _generate_graph(self):
postvars = self._parse_post()
data = []
with hdf_file(postvars['file_path'][0]) as hdf:
params = hdf.get_params(postvars['parameters[]']).values()
align_param = params[0]
arrays = [align_param[0]]
for param in align_param:
arrays.append(align(param, align_param))
# TODO: Get parameter data and return it as AJAX.
for param in params.values():
data.append(zip(range(len(param.array)), param.array.data.tolist()))
return self._respond_with_json({'data': data})
def _get_normalised_split_params(hdf):
'''
Get split parameters (currently engine power and Groundspeed) from hdf,
normalise them on a scale from 0-1.0 and return the minimum.
:param hdf: hdf_file object.
:type hdf: hdfaccess.file.hdf_file
:returns: Minimum of normalised split parameters along with its frequency.
Will return None, None if no split parameters are available.
:rtype: (None, None) or (np.ma.masked_array, float)
'''
params = []
first_split_param = None
split_params = (
'Eng (1) N1', 'Eng (2) N1', 'Eng (3) N1', 'Eng (4) N1',
'Eng (1) N2', 'Eng (2) N2', 'Eng (3) N2', 'Eng (4) N2',
'Eng (1) Np', 'Eng (2) Np', 'Eng (3) Np', 'Eng (4) Np',
'Eng (1) Fuel Flow', 'Eng (2) Fuel Flow', 'Eng (3) Fuel Flow', 'Eng (4) Fuel Flow',
'Groundspeed', 'Groundspeed (1)', 'Groundspeed (2)'
)
for param_name in split_params:
try:
param = hdf[param_name]
except KeyError:
continue
if first_split_param:
# Align all other parameters to first available. #Q: Why not force
# to 1Hz?
param.array = align(param, first_split_param)
else:
first_split_param = param
params.append(param)
if not first_split_param:
return None, None
# If there is at least one split parameter available.
# normalise the parameters we'll use for splitting the data
stacked_params = vstack_params(*params)
# We normalise each in turn to the range 0-1 so they have equal weight
normalised_params = [normalise(i) for i in stacked_params]
# Using a true minimum leads to bias to a zero value. We take the average
# to allow each parameter equal weight, then (later) seek the minimum.
split_params_min = np.ma.average(normalised_params, axis=0)
return split_params_min, first_split_param.frequency
def _get_normalised_split_params(hdf):
'''
Get split parameters (currently engine power and Groundspeed) from hdf,
normalise them on a scale from 0-1.0 and return the minimum.
:param hdf: hdf_file object.
:type hdf: hdfaccess.file.hdf_file
:returns: Minimum of normalised split parameters along with its frequency.
Will return None, None if no split parameters are available.
:rtype: (None, None) or (np.ma.masked_array, float)
'''
params = []
first_split_param = None
split_params = (
'Eng (1) N1', 'Eng (2) N1', 'Eng (3) N1', 'Eng (4) N1',
'Eng (1) N2', 'Eng (2) N2', 'Eng (3) N2', 'Eng (4) N2',
'Eng (1) Np', 'Eng (2) Np', 'Eng (3) Np', 'Eng (4) Np',
'Eng (1) Fuel Flow', 'Eng (2) Fuel Flow', 'Eng (3) Fuel Flow', 'Eng (4) Fuel Flow',
'Groundspeed', 'Groundspeed (1)', 'Groundspeed (2)'
)
for param_name in split_params:
try:
param = hdf[param_name]
except KeyError:
continue
if first_split_param:
# Align all other parameters to first available. #Q: Why not force
# to 1Hz?
param.array = align(param, first_split_param)
else:
first_split_param = param
params.append(param)
if not first_split_param:
return None, None
# If there is at least one split parameter available.
# normalise the parameters we'll use for splitting the data
stacked_params = vstack_params(*params)
# We normalise each in turn to the range 0-1 so they have equal weight
normalised_params = [normalise(i) for i in stacked_params]
# Using a true minimum leads to bias to a zero value. We take the average
# to allow each parameter equal weight, then (later) seek the minimum.
split_params_min = np.ma.average(normalised_params, axis=0)
return split_params_min, first_split_param.frequency
def get_dt_arrays(hdf, fallback_dt, validation_dt):
now = datetime.utcnow().replace(tzinfo=pytz.utc)
if fallback_dt:
fallback_dts = []
for secs in range(0, int(hdf.duration)):
fallback_dts.append(fallback_dt + timedelta(seconds=secs))
onehz = P(frequency=1)
dt_arrays = []
precise = True
for name in ('Year', 'Month', 'Day', 'Hour', 'Minute', 'Second'):
param = hdf.get(name)
if param:
if name == 'Year':
year = getattr(validation_dt, 'year', None) or now.year
param.array = _mask_invalid_years(param.array, year)
# do not interpolate date/time parameters to avoid rollover issues
array = align(param, onehz, interpolate=False)
if len(array) == 0 or np.ma.count(array) == 0:
logger.warning("No valid values returned for %s", name)
elif (np.ma.all(array == 0)) and not (
name == 'Hour' and len(array) < 3600
): # Hour can be 0 for up to 1 hour (after midnight)
# Other than the year 2000 or possibly 2100, no date values
# can be all 0's
logger.warning("Only zero values returned for %s", name)
else:
# values returned, continue
dt_arrays.append(array)
continue
if fallback_dt:
precise = False
array = [getattr(dt, name.lower()) for dt in fallback_dts]
logger.warning(
"%s not available, using range from %d to %d from fallback_dt %s",
name, array[0], array[-1], fallback_dt)
dt_arrays.append(array)
continue
else:
raise TimebaseError("Required parameter '%s' not available" % name)
return dt_arrays, precise
def get_dt_arrays(hdf, fallback_dt, validation_dt):
now = datetime.utcnow().replace(tzinfo=pytz.utc)
if fallback_dt:
fallback_dts = []
for secs in xrange(0, int(hdf.duration)):
fallback_dts.append(fallback_dt + timedelta(seconds=secs))
onehz = P(frequency=1)
dt_arrays = []
for name in ('Year', 'Month', 'Day', 'Hour', 'Minute', 'Second'):
param = hdf.get(name)
if param:
if name == 'Year':
year = getattr(validation_dt, 'year', None) or now.year
param.array = _mask_invalid_years(param.array, year)
# do not interpolate date/time parameters to avoid rollover issues
array = align(param, onehz, interpolate=False)
if len(array) == 0 or np.ma.count(array) == 0:
logger.warning("No valid values returned for %s", name)
elif (np.ma.all(array == 0)) and not (name == 'Hour' and len(array) < 3600): # Hour can be 0 for up to 1 hour (after midnight)
# Other than the year 2000 or possibly 2100, no date values
# can be all 0's
logger.warning("Only zero values returned for %s", name)
else:
# values returned, continue
dt_arrays.append(array)
continue
if fallback_dt:
array = [getattr(dt, name.lower()) for dt in fallback_dts]
logger.warning("%s not available, using range from %d to %d from fallback_dt %s",
name, array[0], array[-1], fallback_dt)
dt_arrays.append(array)
continue
else:
raise TimebaseError("Required parameter '%s' not available" % name)
return dt_arrays
def _get_eng_params(hdf, align_param=None):
'''
Get eng parameters from hdf, and return the minimum.
:param hdf: hdf_file object.
:type hdf: hdfaccess.file.hdf_file
:returns: Minimum of normalised split parameters along with its frequency.
Will return None, None if no split parameters are available.
:rtype: (None, None) or (np.ma.masked_array, float)
'''
params = []
eng_params = (
'Eng (1) N1', 'Eng (2) N1', 'Eng (3) N1', 'Eng (4) N1',
'Eng (1) N2', 'Eng (2) N2', 'Eng (3) N2', 'Eng (4) N2',
'Eng (1) Np', 'Eng (2) Np', 'Eng (3) Np', 'Eng (4) Np',
'Eng (1) Fuel Flow', 'Eng (2) Fuel Flow', 'Eng (3) Fuel Flow', 'Eng (4) Fuel Flow'
)
for param_name in eng_params:
try:
param = hdf[param_name]
except KeyError:
continue
if align_param:
# Align all other parameters to provided param or first available.
param.array = align(param, align_param)
else:
align_param = param
params.append(param)
if not len(params):
return None, None
# If there is at least one split parameter available.
stacked_params = vstack_params(*params)
split_params_avg = np.ma.average(stacked_params, axis=0)
return split_params_avg, align_param.frequency
def _get_normalised_split_params(hdf):
'''
Get split parameters (currently engine power and Groundspeed) from hdf,
normalise them on a scale from 0-1.0 and return the minimum.
:param hdf: hdf_file object.
:type hdf: hdfaccess.file.hdf_file
:returns: Minimum of normalised split parameters along with its frequency.
Will return None, None if no split parameters are available.
:rtype: (None, None) or (np.ma.masked_array, float)
'''
params = []
first_split_param = None
for param_name in ('Eng (1) N1', 'Eng (2) N1', 'Eng (3) N1', 'Eng (4) N1',
'Eng (1) N2', 'Eng (2) N2', 'Eng (3) N2', 'Eng (4) N2',
'Eng (1) Np', 'Eng (2) Np', 'Eng (3) Np', 'Eng (4) Np',
'Groundspeed', 'Groundspeed (1)', 'Groundspeed (2)'):
try:
param = hdf[param_name]
except KeyError:
continue
if first_split_param:
# Align all other parameters to first available. #Q: Why not force
# to 1Hz?
param.array = align(param, first_split_param)
else:
first_split_param = param
params.append(param)
if not first_split_param:
return None, None
# If there is at least one split parameter available.
# normalise the parameters we'll use for splitting the data
stacked_params = vstack_params(*params)
normalised_params = normalise(stacked_params, scale_max=100)
split_params_min = np.ma.average(normalised_params, axis=0)
return split_params_min, first_split_param.frequency
def plot_parameters(params, axes, title=''):
'''
Plot resulting parameters.
'''
print 'Plotting parameters.'
max_freq = 0
min_freq = float('inf')
for name, param in params.iteritems():
max_freq = max(max_freq, param.frequency)
min_freq = min(min_freq, param.frequency)
for param_name, param in params.iteritems():
if max_freq == param.frequency:
param_max_freq = param
if param.frequency == min_freq:
param_min_freq_len = len(param.array)
# Truncate parameter arrays to successfully align them since the file
# has not been through split sections.
for param_name, param in params.iteritems():
array_len = param_min_freq_len * (param.frequency / min_freq)
if array_len != len(param.array):
print 'Truncated %s from %d to %d for display purposes' % (
param_name, len(param.array), array_len)
param.array = param.array[:array_len]
#==========================================================================
# Plot Preparation
#==========================================================================
# Invariant parameters are identified here. They could be inserted into
# the plot configuration file, but this is more straightforward.
plt.rc('axes', grid=True)
plt.rc('grid', color='0.75', linestyle='-', linewidth=0.5)
# These items are altered during the plot, so not suited to plt.rc setup
prop = fm.FontProperties(size=10)
legendprops = dict(shadow=True, fancybox=True, markerscale=0.5, prop=prop)
# Start by making a big clean canvas
fig = plt.figure(facecolor='white', figsize=(8, 6))
fig.canvas.set_window_title("%s %s" % (
title, datetime.now().strftime('%A, %d %B %Y at %X')))
# Add the "reference" altitude plot, and title this
# (If we title the empty plot, it acquires default 0-1 scales)
param_name = axes[1][0]
param = params[param_name]
array = align(param, param_max_freq)
first_axis = fig.add_subplot(len(axes), 1, 1)
first_axis.plot(array, label=param_name)
####plt.title("Processed on %s" %
#### datetime.now().strftime('%A, %d %B %Y at %X'))
setp(first_axis.get_xticklabels(), visible=False)
# Now plot the additional data from the AXIS_N lists at the top of the lfl
for index, param_names in axes.iteritems():
if index == 1:
continue
axis = fig.add_subplot(len(axes), 1, index, sharex=first_axis)
# Avoid iterating over string
if isinstance(param_names, basestring):
param_names = [param_names]
for param_name in param_names:
param = params[param_name]
# Data is aligned in time but the samples are not interpolated so
# that scaling issues can be easily addressed
label_text = param.name
args = []
if np.ma.all(param.array.mask):
args.append([])
label_text += ' <ALL MASKED>'
elif param.data_type == 'ASCII' or param.array.dtype.char == 'S':
print "Warning: ASCII not supported. Param '%s'" % param
args.append([])
label_text += ' <ASCII NOT DRAWN>'
elif param.hz != max_freq:
# Data is aligned in time but the samples are not
# interpolated so that scaling issues can be easily addressed
args.append(np.arange(len(param.array)) * (max_freq / param.hz))
args.append(param.array)
else:
args.append(param.array)
if param.units is None:
label_text += " [No units]"
else:
label_text += " : " + param.units
values_mapping = getattr(param.array, 'values_mapping', None)
if values_mapping:
label_text += '\n%s' % values_mapping
axis.plot(*args, label=label_text)
axis.legend(loc='upper right', **legendprops)
if index < len(axes):
setp(axis.get_xticklabels(), visible=False)
plt.legend(prop={'size': 10})
plt.show()
def _calculate_start_datetime(hdf, fallback_dt=None):
"""
Calculate start datetime.
:param hdf: Flight data HDF file
:type hdf: hdf_access object
:param fallback_dt: Used to replace elements of datetimes which are not available in the hdf file (e.g. YEAR not being recorded)
:type fallback_dt: datetime
HDF params used:
:Year: Optional (defaults to 1970)
:Month: Optional (defaults to 1)
:Day: Optional (defaults to 1)
:Hour: Required
:Minute: Required
:Second: Required
If required parameters are not available and fallback_dt is not provided,
a TimebaseError is raised
"""
now = datetime.now()
if fallback_dt is not None:
assert fallback_dt < now, \
("Fallback time '%s' in the future is not allowed. Current time "
"is '%s'." % (fallback_dt, now))
# align required parameters to 1Hz
onehz = P(frequency = 1)
dt_arrays = []
for name in ('Year', 'Month', 'Day', 'Hour', 'Minute', 'Second'):
param = hdf.get(name)
if param:
if name == 'Year':
year = getattr(fallback_dt, 'year', None) or now.year
param.array = _mask_invalid_years(param.array, year)
# do not interpolate date/time parameters to avoid rollover issues
array = align(param, onehz, interpolate=False)
if len(array) == 0 or np.ma.count(array) == 0 or np.ma.all(array == 0):
# Other than the year 2000 or possibly 2100, no date values
# can be all 0's
logger.warning("No valid values returned for %s", name)
else:
# values returned, continue
dt_arrays.append(array)
continue
if fallback_dt:
array = [getattr(fallback_dt, name.lower())]
logger.warning("%s not available, using %d from fallback_dt %s",
name, array[0], fallback_dt)
dt_arrays.append(array)
continue
else:
raise TimebaseError("Required parameter '%s' not available" % name)
length = max([len(array) for array in dt_arrays])
if length > 1:
# ensure all arrays are the same length
for n, arr in enumerate(dt_arrays):
if len(arr) == 1:
# repeat to the correct size
arr = np.repeat(arr, length)
dt_arrays[n] = arr
elif len(arr) != length:
raise ValueError("After align, all arrays should be the same "
"length")
else:
pass
# establish timebase for start of data
try:
timebase = calculate_timebase(*dt_arrays)
except (KeyError, ValueError) as err:
raise TimebaseError("Error with timestamp values: %s" % err)
if timebase > now:
# Flight Data Analysis in the future is a challenge, lets see if we
# can correct this first...
if 'Day' not in hdf:
# unlikely to have year, month or day.
# Scenario: that fallback_dt is of the current day but recorded
# time is in the future of the fallback time, therefore resulting
# in a futuristic date.
a_day_before = timebase - relativedelta(days=1)
if a_day_before < now:
logger.info("Timebase was in the future, using a day before satisfies requirements")
return a_day_before
# continue to take away a Year
if 'Year' not in hdf:
# remove a year from the timebase
a_year_before = timebase - relativedelta(years=1)
if a_year_before < now:
logger.info("Timebase was in the future, using a day before satisfies requirements")
return a_year_before
raise TimebaseError("Timebase '%s' is in the future.", timebase)
if settings.MAX_TIMEBASE_AGE and \
timebase < (now - timedelta(days=settings.MAX_TIMEBASE_AGE)):
# Only allow recent timebases.
error_msg = "Timebase '%s' older than the allowed '%d' days." % (
timebase, settings.MAX_TIMEBASE_AGE)
raise TimebaseError(error_msg)
logger.info("Valid timebase identified as %s", timebase)
return timebase
def derive(self,
gl=M('Gear (L) On Ground'),
gr=M('Gear (R) On Ground'),
vert_spd=P('Vertical Speed'),
torque=P('Eng (*) Torque Avg'),
ac_series=A('Series'),
collective=P('Collective')):
if gl and gr:
delta = abs((gl.offset - gr.offset) * gl.frequency)
if 0.75 < delta or delta < 0.25:
# If the samples of the left and right gear are close together,
# the best representation is to map them onto a single
# parameter in which we accept that either wheel on the ground
# equates to gear on ground.
self.array = np.ma.logical_or(gl.array, gr.array)
self.frequency = gl.frequency
self.offset = gl.offset
return
else:
# If the paramters are not co-located, then
# merge_two_parameters creates the best combination possible.
self.array, self.frequency, self.offset = merge_two_parameters(
gl, gr)
return
elif gl or gr:
gear = gl or gr
self.array = gear.array
self.frequency = gear.frequency
self.offset = gear.offset
elif vert_spd and torque:
vert_spd_limit = 100.0
torque_limit = 30.0
if ac_series and ac_series.value == 'Columbia 234':
vert_spd_limit = 125.0
torque_limit = 22.0
collective_limit = 15.0
vert_spd_array = align(
vert_spd,
torque) if vert_spd.hz != torque.hz else vert_spd.array
collective_array = align(
collective,
torque) if collective.hz != torque.hz else collective.array
vert_spd_array = moving_average(vert_spd_array)
torque_array = moving_average(torque.array)
collective_array = moving_average(collective_array)
roo_vs_array = runs_of_ones(
abs(vert_spd_array) < vert_spd_limit, min_samples=1)
roo_torque_array = runs_of_ones(torque_array < torque_limit,
min_samples=1)
roo_collective_array = runs_of_ones(
collective_array < collective_limit, min_samples=1)
vs_and_torque = slices_and(roo_vs_array, roo_torque_array)
grounded = slices_and(vs_and_torque, roo_collective_array)
array = np_ma_zeros_like(vert_spd_array)
for _slice in slices_remove_small_slices(grounded, count=2):
array[_slice] = 1
array.mask = vert_spd_array.mask | torque_array.mask
array.mask = array.mask | collective_array.mask
self.array = nearest_neighbour_mask_repair(array)
self.frequency = torque.frequency
self.offset = torque.offset
else:
vert_spd_array = align(
vert_spd,
torque) if vert_spd.hz != torque.hz else vert_spd.array
# Introducted for S76 and Bell 212 which do not have Gear On Ground available
vert_spd_array = moving_average(vert_spd_array)
torque_array = moving_average(torque.array)
grounded = slices_and(
runs_of_ones(abs(vert_spd_array) < vert_spd_limit,
min_samples=1),
runs_of_ones(torque_array < torque_limit, min_samples=1))
array = np_ma_zeros_like(vert_spd_array)
for _slice in slices_remove_small_slices(grounded, count=2):
array[_slice] = 1
array.mask = vert_spd_array.mask | torque_array.mask
self.array = nearest_neighbour_mask_repair(array)
self.frequency = torque.frequency
self.offset = torque.offset
else:
# should not get here if can_operate is correct
raise NotImplementedError()
def derive(self,
gl=M('Gear (L) On Ground'),
gr=M('Gear (R) On Ground'),
vert_spd=P('Vertical Speed'),
torque=P('Eng (*) Torque Avg'),
ac_series=A('Series'),
collective=P('Collective')):
if gl and gr:
delta = abs((gl.offset - gr.offset) * gl.frequency)
if 0.75 < delta or delta < 0.25:
# If the samples of the left and right gear are close together,
# the best representation is to map them onto a single
# parameter in which we accept that either wheel on the ground
# equates to gear on ground.
self.array = np.ma.logical_or(gl.array, gr.array)
self.frequency = gl.frequency
self.offset = gl.offset
return
else:
# If the paramters are not co-located, then
# merge_two_parameters creates the best combination possible.
self.array, self.frequency, self.offset = merge_two_parameters(gl, gr)
return
elif gl or gr:
gear = gl or gr
self.array = gear.array
self.frequency = gear.frequency
self.offset = gear.offset
elif vert_spd and torque:
vert_spd_limit = 100.0
torque_limit = 30.0
if ac_series and ac_series.value == 'Columbia 234':
vert_spd_limit = 125.0
torque_limit = 22.0
collective_limit = 15.0
vert_spd_array = align(vert_spd, torque) if vert_spd.hz != torque.hz else vert_spd.array
collective_array = align(collective, torque) if collective.hz != torque.hz else collective.array
vert_spd_array = moving_average(vert_spd_array)
torque_array = moving_average(torque.array)
collective_array = moving_average(collective_array)
roo_vs_array = runs_of_ones(abs(vert_spd_array) < vert_spd_limit, min_samples=1)
roo_torque_array = runs_of_ones(torque_array < torque_limit, min_samples=1)
roo_collective_array = runs_of_ones(collective_array < collective_limit, min_samples=1)
vs_and_torque = slices_and(roo_vs_array, roo_torque_array)
grounded = slices_and(vs_and_torque, roo_collective_array)
array = np_ma_zeros_like(vert_spd_array)
for _slice in slices_remove_small_slices(grounded, count=2):
array[_slice] = 1
array.mask = vert_spd_array.mask | torque_array.mask
array.mask = array.mask | collective_array.mask
self.array = nearest_neighbour_mask_repair(array)
self.frequency = torque.frequency
self.offset = torque.offset
else:
vert_spd_array = align(vert_spd, torque) if vert_spd.hz != torque.hz else vert_spd.array
# Introducted for S76 and Bell 212 which do not have Gear On Ground available
vert_spd_array = moving_average(vert_spd_array)
torque_array = moving_average(torque.array)
grounded = slices_and(runs_of_ones(abs(vert_spd_array) < vert_spd_limit, min_samples=1),
runs_of_ones(torque_array < torque_limit, min_samples=1))
array = np_ma_zeros_like(vert_spd_array)
for _slice in slices_remove_small_slices(grounded, count=2):
array[_slice] = 1
array.mask = vert_spd_array.mask | torque_array.mask
self.array = nearest_neighbour_mask_repair(array)
self.frequency = torque.frequency
self.offset = torque.offset
else:
# should not get here if can_operate is correct
raise NotImplementedError()
def derive(self,
vert_spd=P('Vertical Speed'),
torque=P('Eng (*) Torque Avg'),
ac_series=A('Series'),
collective=P('Collective')):
vert_spd_limit = 100.0
torque_limit = 30.0
if ac_series and ac_series.value == 'Columbia 234':
vert_spd_limit = 125.0
torque_limit = 22.0
collective_limit = 15.0
vert_spd_array = align(
vert_spd,
torque) if vert_spd.hz != torque.hz else vert_spd.array
collective_array = align(
collective,
torque) if collective.hz != torque.hz else collective.array
vert_spd_array = moving_average(vert_spd_array)
torque_array = moving_average(torque.array)
collective_array = moving_average(collective_array)
roo_vs_array = runs_of_ones(abs(vert_spd_array) < vert_spd_limit,
min_samples=1)
roo_torque_array = runs_of_ones(torque_array < torque_limit,
min_samples=1)
roo_collective_array = runs_of_ones(
collective_array < collective_limit, min_samples=1)
vs_and_torque = slices_and(roo_vs_array, roo_torque_array)
grounded = slices_and(vs_and_torque, roo_collective_array)
array = np_ma_zeros_like(vert_spd_array)
for _slice in slices_remove_small_slices(grounded, count=2):
array[_slice] = 1
array.mask = vert_spd_array.mask | torque_array.mask
array.mask = array.mask | collective_array.mask
self.array = nearest_neighbour_mask_repair(array)
self.frequency = torque.frequency
self.offset = torque.offset
else:
vert_spd_array = align(
vert_spd,
torque) if vert_spd.hz != torque.hz else vert_spd.array
# Introducted for S76 and Bell 212 which do not have Gear On Ground available
vert_spd_array = moving_average(vert_spd_array)
torque_array = moving_average(torque.array)
grounded = slices_and(
runs_of_ones(abs(vert_spd_array) < vert_spd_limit,
min_samples=1),
runs_of_ones(torque_array < torque_limit, min_samples=1))
array = np_ma_zeros_like(vert_spd_array)
for _slice in slices_remove_small_slices(grounded, count=2):
array[_slice] = 1
array.mask = vert_spd_array.mask | torque_array.mask
self.array = nearest_neighbour_mask_repair(array)
self.frequency = torque.frequency
self.offset = torque.offset
def _calculate_start_datetime(hdf, fallback_dt=None):
"""
Calculate start datetime.
:param hdf: Flight data HDF file
:type hdf: hdf_access object
:param fallback_dt: Used to replace elements of datetimes which are not
available in the hdf file (e.g. YEAR not being recorded)
:type fallback_dt: datetime
HDF params used:
:Year: Optional (defaults to 1970)
:Month: Optional (defaults to 1)
:Day: Optional (defaults to 1)
:Hour: Required
:Minute: Required
:Second: Required
If required parameters are not available and fallback_dt is not provided,
a TimebaseError is raised
"""
now = datetime.utcnow().replace(tzinfo=pytz.utc)
if fallback_dt is not None:
if (fallback_dt.tzinfo is None or
fallback_dt.tzinfo.utcoffset(fallback_dt) is None):
# Assume fallback_dt is UTC.
fallback_dt = fallback_dt.replace(tzinfo=pytz.utc)
assert fallback_dt < now, (
"Fallback time '%s' in the future is not allowed. Current time "
"is '%s'." % (fallback_dt, now))
# align required parameters to 1Hz
onehz = P(frequency=1)
dt_arrays = []
for name in ('Year', 'Month', 'Day', 'Hour', 'Minute', 'Second'):
param = hdf.get(name)
if param:
if name == 'Year':
year = getattr(fallback_dt, 'year', None) or now.year
param.array = _mask_invalid_years(param.array, year)
# do not interpolate date/time parameters to avoid rollover issues
array = align(param, onehz, interpolate=False)
if len(array) == 0 or np.ma.count(array) == 0 \
or np.ma.all(array == 0):
# Other than the year 2000 or possibly 2100, no date values
# can be all 0's
logger.warning("No valid values returned for %s", name)
else:
# values returned, continue
dt_arrays.append(array)
continue
if fallback_dt:
array = [getattr(fallback_dt, name.lower())]
logger.warning("%s not available, using %d from fallback_dt %s",
name, array[0], fallback_dt)
dt_arrays.append(array)
continue
else:
raise TimebaseError("Required parameter '%s' not available" % name)
length = max([len(a) for a in dt_arrays])
if length > 1:
# ensure all arrays are the same length
for n, arr in enumerate(dt_arrays):
if len(arr) == 1:
# repeat to the correct size
arr = np.repeat(arr, length)
dt_arrays[n] = arr
elif len(arr) != length:
raise ValueError("After align, all arrays should be the same "
"length")
else:
pass
# establish timebase for start of data
try:
timebase = calculate_timebase(*dt_arrays)
except (KeyError, ValueError) as err:
raise TimebaseError("Error with timestamp values: %s" % err)
if timebase > now:
# Flight Data Analysis in the future is a challenge, lets see if we
# can correct this first...
if 'Day' not in hdf:
# unlikely to have year, month or day.
# Scenario: that fallback_dt is of the current day but recorded
# time is in the future of the fallback time, therefore resulting
# in a futuristic date.
a_day_before = timebase - relativedelta(days=1)
if a_day_before < now:
logger.info(
"Timebase was in the future, using a DAY before "
"satisfies requirements: %s", a_day_before)
return a_day_before
# continue to take away a Year
if 'Year' not in hdf:
# remove a year from the timebase
a_year_before = timebase - relativedelta(years=1)
if a_year_before < now:
logger.info("Timebase was in the future, using a YEAR before "
"satisfies requirements: %s", a_year_before)
return a_year_before
raise TimebaseError("Timebase '%s' is in the future.", timebase)
if settings.MAX_TIMEBASE_AGE and \
timebase < (now - timedelta(days=settings.MAX_TIMEBASE_AGE)):
# Only allow recent timebases.
error_msg = "Timebase '%s' older than the allowed '%d' days." % (
timebase, settings.MAX_TIMEBASE_AGE)
raise TimebaseError(error_msg)
logger.info("Valid timebase identified as %s", timebase)
return timebase