def process(self, inputs):
"""Reads the light curve of a particular Kepler ID."""
kep_id = inputs["kepler_id"]
all_time = None
all_flux = None
filenames = kepler_io.kepler_filenames(
base_dir=self.kepler_data_dir,
kep_id=kep_id,
long_cadence=self.long_cadence,
quarters=self.quarters,
injected_group=self.injected_group)
if filenames:
try:
all_time, all_flux = kepler_io.read_kepler_light_curve(
filenames,
light_curve_extension=self.extension,
scramble_type=self.scramble_type,
invert=self.invert_light_curves)
except (IOError, ValueError) as e:
raise ValueError("Kepler ID: {}, {}".format(kep_id, e))
else:
Metrics.counter(self.__class__.__name__,
"no-fits-%s" % kep_id).inc()
raw_lc = light_curve_pb2.RawLightCurve()
for time, flux in zip(all_time, all_flux):
raw_lc.segments.add(time=time, flux=flux)
inputs["raw_light_curve"] = raw_lc
yield inputs
def process(self, inputs):
"""Reads the light curve of a particular Kepler ID."""
kep_id = inputs["kepler_id"]
Metrics.counter(self.__class__.__name__, "inputs").inc()
# Get light curve filenames.
filenames = kepler_io.kepler_filenames(
base_dir=self.kepler_data_dir,
kep_id=kep_id,
injected_group=self.injected_group)
if not filenames:
Metrics.counter(self.__class__.__name__,
"no-fits-{}".format(kep_id)).inc()
return
cadence_no, time, flux = self._read_kepler_light_curve(filenames)
# Additional normalization.
stddev = 1.4826 * np.median(np.abs(flux)) # 1.4826 * MAD (median is zero).
if self.upward_outlier_clipping:
# Clip values greater than n stddev from the median (which is zero).
flux = np.minimum(flux, self.upward_outlier_clipping * stddev)
if self.downward_outlier_clipping:
# pylint: disable=invalid-unary-operand-type
flux = np.maximum(flux, -self.downward_outlier_clipping * stddev)
if self.clip_lowest_n_values:
nth_flux = _nth_smallest(flux, self.clip_lowest_n_values)
flux = np.maximum(flux, nth_flux)
if self.normalize_stddev:
flux /= stddev
cadence_no, time, flux, mask = util.uniform_cadence_light_curve(
cadence_no, time, flux)
ex = tf.train.Example()
example_util.set_int64_feature(ex, "kepler_id", [kep_id])
example_util.set_bytes_feature(ex, "flux_column", [self.flux_column])
example_util.set_bytes_feature(ex, "injected_group", [self.injected_group])
example_util.set_bytes_feature(ex, "scramble_type", [self.scramble_type])
example_util.set_float_feature(ex, "time", time)
example_util.set_float_feature(ex, "flux", flux)
example_util.set_int64_feature(ex, "cadence_no", cadence_no)
example_util.set_int64_feature(ex, "mask", mask)
inputs["example"] = ex
Metrics.counter(self.__class__.__name__, "outputs").inc()
yield inputs
def read_light_curve(kepid, kepler_data_dir):
"""Reads a Kepler light curve.
Args:
kepid: Kepler id of the target star.
kepler_data_dir: Base directory containing Kepler data. See
kepler_io.kepler_filenames().
Returns:
all_time: A list of numpy arrays; the time values of the raw light curve.
all_flux: A list of numpy arrays corresponding to the time arrays in
all_time.
Raises:
IOError: If the light curve files for this Kepler ID cannot be found.
"""
# Read the Kepler light curve.
file_names = kepler_io.kepler_filenames(kepler_data_dir, kepid)
if not file_names:
raise IOError(
"Failed to find .fits files in {} for Kepler ID {}".format(
kepler_data_dir, kepid))
return kepler_io.read_kepler_light_curve(file_names)
def testKeplerFilenames(self):
# All quarters.
filenames = kepler_io.kepler_filenames("/my/dir/",
1234567,
check_existence=False)
self.assertCountEqual([
"/my/dir/0012/001234567/kplr001234567-2009131105131_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2009166043257_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2009259160929_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2009350155506_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2010078095331_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2010009091648_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2010174085026_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2010265121752_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2010355172524_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2011073133259_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2011177032512_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2011271113734_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2012004120508_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2012088054726_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2012179063303_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2012277125453_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2013011073258_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2013098041711_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2013131215648_llc.fits"
], filenames)
# Subset of quarters.
filenames = kepler_io.kepler_filenames("/my/dir/",
1234567,
quarters=[3, 4],
check_existence=False)
self.assertCountEqual([
"/my/dir/0012/001234567/kplr001234567-2009350155506_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2010078095331_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2010009091648_llc.fits"
], filenames)
# Injected group.
filenames = kepler_io.kepler_filenames("/my/dir/",
1234567,
quarters=[3, 4],
injected_group="inj1",
check_existence=False)
# pylint:disable=line-too-long
self.assertCountEqual([
"/my/dir/0012/001234567/kplr001234567-2009350155506_INJECTED-inj1_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2010078095331_INJECTED-inj1_llc.fits",
"/my/dir/0012/001234567/kplr001234567-2010009091648_INJECTED-inj1_llc.fits"
], filenames)
# pylint:enable=line-too-long
# Short cadence.
filenames = kepler_io.kepler_filenames("/my/dir/",
1234567,
long_cadence=False,
quarters=[0, 1],
check_existence=False)
self.assertCountEqual([
"/my/dir/0012/001234567/kplr001234567-2009131110544_slc.fits",
"/my/dir/0012/001234567/kplr001234567-2009166044711_slc.fits"
], filenames)
# Check existence.
filenames = kepler_io.kepler_filenames(self.data_dir,
11442793,
check_existence=True)
expected_filenames = [
os.path.join(self.data_dir,
"0114/011442793/kplr011442793-{}_llc.fits".format(q))
for q in ["2009350155506", "2010009091648", "2010174085026"]
]
self.assertCountEqual(expected_filenames, filenames)