def setup_for_order(r, datafile, training_epochs, validation_epochs):
# load up training and validation data for order r & set up results
training_data = wobble.Data(datafile, orders=[r],
epochs=training_epochs, min_snr=0.)
training_results = wobble.Results(training_data)
validation_data = wobble.Data(datafile, orders=training_data.orders,
epochs=validation_epochs, min_snr=0.)
validation_results = wobble.Results(validation_data)
return training_data, training_results, validation_data, validation_results
star_reg_file = '../wobble/regularization/{0}_star_K{1}.hdf5'.format(starname, K_star)
tellurics_reg_file = '../wobble/regularization/{0}_t_K{1}.hdf5'.format(starname, K_t)
plot_dir = '../results/plots_{0}_Kstar{1}_Kt{2}/'.format(starname, K_star, K_t)
print("running wobble on star {0} with K_star = {1}, K_t = {2}".format(starname, K_star, K_t))
start_time = time()
orders = np.arange(72)
data = wobble.Data(filename='../data/'+starname+'_e2ds.hdf5', orders=orders)
if True: # reload data and remove all post-upgrade spectra
upgrade = 2457174.5 # June 2015
e = data.epochs[data.dates < upgrade]
data = wobble.Data(filename='../data/'+starname+'_e2ds.hdf5', orders=orders, epochs=e)
data.drop_bad_orders()
data.drop_bad_epochs()
orders = np.copy(data.orders)
results = wobble.Results(data=data)
print("data loaded")
print("time elapsed: {0:.2f} min".format((time() - start_time)/60.0))
elapsed_time = time() - start_time
if plots:
print("plots will be saved under directory: {0}".format(plot_dir))
if not os.path.exists(plot_dir):
os.makedirs(plot_dir)
star_learning_rate = 0.1
telluric_learning_rate = 0.01
for r,o in enumerate(orders):
model = wobble.Model(data, results, r)
model.add_star('star', variable_bases=K_star,
f['orders'] = np.append(orders, g['orders'][()])
#check which orders are already there because wobble might drop some
# TODO clean way is probably to make some neat ordered list of orders in either file first
key_list = list(f.keys())
for r in range(n_orders):
if 'order{0}'.format(
r
) not in key_list: #find first order index not already present
for r_chunk in range(g['R'][()]):
r_tot = r + r_chunk
#f.create_group('order{0}'.format(r_tot))
g.copy('order{0}'.format(r_chunk),
f,
name='order{0}'.format(r_tot))
break
# combine orders even though thiscurrently yields garbage RVs, still makes issues for chunks (different n_Epochs?, maybe it was just median breaking on chunksize 1?)
print('n_orders = R =', f['R'][()])
results = wobble.Results(filename=results_file)
results.combine_orders('star')
print("final RVs calculated.")
print("time elapsed: {0:.2f} minutes".format((time() - start_time) / 60.0))
results.write(results_file)
print("all scripts executed")
print("time elapsed total: {0:.2f} min".format((time() - start_time) / 60.0))
def improve_order_regularization(r,
o,
star_filename,
tellurics_filename,
training_data,
training_results,
validation_data,
validation_results,
verbose=True,
plot=False,
basename='',
K_star=0,
K_t=0,
L1=True,
L2=True,
tellurics_template_fixed=False):
"""
Use a validation scheme to determine the best regularization parameters for
all model components in a given order r.
Update files at star_filename, tellurics_filename with the best parameters.
"""
training_model = wobble.Model(training_data, training_results, r)
training_model.add_star('star', variable_bases=K_star)
if tellurics_template_fixed: # hackity hack hack
results_51peg = wobble.Results(
filename=
'/Users/mbedell/python/wobble/results/results_51peg_Kstar0_Kt0.hdf5'
)
template_xs = np.copy(results_51peg.tellurics_template_xs[o])
template_ys = np.copy(results_51peg.tellurics_template_ys[o])
training_model.add_telluric('tellurics',
rvs_fixed=True,
template_fixed=True,
variable_bases=K_t,
template_xs=template_xs,
template_ys=template_ys)
else:
training_model.add_telluric('tellurics',
rvs_fixed=True,
variable_bases=K_t)
training_model.setup()
training_model.optimize(niter=0, verbose=verbose, rv_uncertainties=False)
if plot:
n = 0 # epoch to plot
title = 'Initialization'
filename = '{0}_init'.format(basename)
plot_fit(r,
n,
training_data,
training_results,
title=title,
basename=filename)
validation_model = wobble.Model(validation_data, validation_results, r)
validation_model.add_star('star',
variable_bases=K_star,
template_xs=training_results.star_template_xs[r]
) # ensure templates are same size
if tellurics_template_fixed: # hackity hack hack
validation_model.add_telluric(
'tellurics',
rvs_fixed=True,
template_fixed=True,
variable_bases=K_t,
template_xs=training_results.tellurics_template_xs[r],
template_ys=training_results.tellurics_template_ys[r])
else:
validation_model.add_telluric(
'tellurics',
rvs_fixed=True,
variable_bases=K_t,
template_xs=training_results.tellurics_template_xs[r])
validation_model.setup()
# the order in which these are defined will determine the order in which they are optimized:
tensors_to_tune = [
training_model.components[1].L2_template_tensor,
training_model.components[0].L2_template_tensor,
training_model.components[1].L1_template_tensor,
training_model.components[0].L1_template_tensor
]
tensor_names = [
'L2_template', 'L2_template', 'L1_template', 'L1_template'
] # this isonly needed bc TF appends garbage to the end of the tensor name
tensor_components = ['tellurics', 'star', 'tellurics', 'star'] # ^ same
if K_star > 0:
tensors_to_tune = np.append(tensors_to_tune, [
training_model.components[0].L2_basis_vectors_tensor,
training_model.components[0].L1_basis_vectors_tensor
])
tensor_names = np.append(tensor_names,
['L2_basis_vectors', 'L1_basis_vectors'])
tensor_components = np.append(tensor_components, ['star', 'star'])
if K_t > 0:
tensors_to_tune = np.append(tensors_to_tune, [
training_model.components[1].L2_basis_vectors_tensor,
training_model.components[1].L1_basis_vectors_tensor
])
tensor_names = np.append(tensor_names,
['L2_basis_vectors', 'L1_basis_vectors'])
tensor_components = np.append(tensor_components,
['tellurics', 'tellurics'])
regularization_dict = {}
#o_init = max(0, o-1) # initialize from previous order, or if o=0 use defaults
o_init = o # always initialize from starting guess (TODO: decide which init is better)
for i, tensor in enumerate(tensors_to_tune):
if tensor_components[i] == 'star':
filename = star_filename
elif tensor_components[i] == 'tellurics':
filename = tellurics_filename
else:
print("something has gone wrong.")
assert False
with h5py.File(filename, 'r') as f:
regularization_dict[tensor] = np.copy(f[tensor_names[i]][o_init])
i = 0 # track order in which parameters are improved
for component, (tensor, name) in zip(tensor_components,
zip(tensors_to_tune, tensor_names)):
if (name[0:2] == "L1" and L1) or (name[0:2] == "L2" and L2):
i += 1
regularization_dict[tensor] = improve_parameter(
tensor,
training_model,
validation_model,
regularization_dict,
validation_data,
validation_results,
verbose=verbose,
plot=plot,
basename=basename + '_par{0}'.format(i))
if component == 'star':
filename = star_filename
elif component == 'tellurics':
filename = tellurics_filename
else:
print("something has gone wrong.")
assert False
with h5py.File(filename, 'r+') as f:
f[name][o] = np.copy(regularization_dict[tensor])
if plot:
test_regularization_value(tensor,
regularization_dict[tensor],
training_model,
validation_model,
regularization_dict,
validation_data,
validation_results,
plot=False,
verbose=False) # hack to update results
title = 'Final'
filename = '{0}_final'.format(basename)
plot_fit(r,
n,
validation_data,
validation_results,
title=title,
basename=filename)
fig = plt.figure()
ax = fig.add_subplot(111)
val_rvs = validation_results.star_rvs[r] + validation_results.bervs
train_rvs = training_results.star_rvs[r] + training_results.bervs
ax.plot(validation_results.dates, val_rvs - np.mean(val_rvs), 'r.')
ax.plot(training_results.dates,
train_rvs - np.mean(train_rvs),
'k.',
alpha=0.5)
ax.set_ylabel('RV (m/s)')
ax.set_xlabel('JD')
fig.tight_layout()
plt.savefig(basename + '_final_rvs.png')
plt.close(fig)
else: # HACK for HD 189733
e = np.asarray([
0, 1, 6, 7, 9, 17, 18, 19, 21, 23, 24, 26, 30, 33, 34, 35, 36, 37,
38, 40, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 53, 55, 56, 61, 66,
69, 70, 72, 73, 75
]) # night of August 28, 2007
validation_epochs = np.random.choice(e, len(e) // 8, replace=False)
training_epochs = np.delete(e, validation_epochs)
training_data = wobble.Data(starname + '_e2ds.hdf5',
filepath='../data/',
orders=orders,
epochs=training_epochs,
min_snr=3)
training_results = wobble.Results(training_data)
validation_data = wobble.Data(starname + '_e2ds.hdf5',
filepath='../data/',
orders=training_data.orders,
epochs=validation_epochs,
min_snr=1) # HACK
validation_results = wobble.Results(validation_data)
assert len(training_data.orders) == len(
validation_data.orders
), "Number of orders used is not the same between training and validation data."
orders = training_data.orders
# improve each order's regularization:
for r, o in enumerate(
orders
): # r is an index into the (cleaned) data. o is an index into the 72 orders (and the file tracking them).
print("running wobble on star {0} with K_star = {1}, K_t = {2}".format(
starname, K_star, K_t))
start_time = time()
orders = np.arange(72)
'''
e = [ 0, 1, 6, 7, 9, 17, 18, 19, 21, 23, 24, 26, 30, 33, 34, 35, 36,
37, 38, 40, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 53, 55, 56, 61,
66, 69, 70, 72, 73, 75] # night of August 28, 2007
data = wobble.Data(starname+'_e2ds.hdf5', filepath='../data/', orders=orders, epochs=e)
'''
data = wobble.Data(starname + '_e2ds.hdf5',
filepath='../data/',
orders=orders)
orders = np.copy(data.orders)
results = wobble.Results(data=data)
results_51peg = wobble.Results(
filename=
'/Users/mbedell/python/wobble/results/results_51peg_Kstar0_Kt0.hdf5')
print("data loaded")
print("time elapsed: {0:.2f} min".format((time() - start_time) / 60.0))
elapsed_time = time() - start_time
if plots:
print("plots will be saved under directory: {0}".format(plot_dir))
if not os.path.exists(plot_dir):
os.makedirs(plot_dir)
star_learning_rate = 0.1
telluric_learning_rate = 0.01
if vis == True:
orders = np.arange(11,53)
lowest_optimized_order = 11
data = wobble.Data(starname+'_vis'+'_e2ds.hdf5', filepath= data_directory, orders=orders, min_flux=10**-5, min_snr=0)
else:
orders = np.arange(0,56)
lowest_optimized_order = 0
data = wobble.Data(starname+'_nir_split'+'_e2ds.hdf5', filepath= data_directory, orders=orders, min_flux=10**-5, min_snr=0)
######
#plot_dir = plot_dir + results_name + "/"#+ "_bad_ep/"
plot_dir = plot_dir + results_name + "_bad_ord(49)/"
#plot_dir = plot_dir + "GJ1148_o49_snr" + "/"
os.makedirs(plot_dir, exist_ok = True)
results = wobble.Results(filename = results_directory + results_name +".hdf5")
telluric_mask = np.genfromtxt("/data/cmatthe/python/wobble_aux/telluric_masks/" +
"telluric_mask_carm_short.dat"
)
#epochs = [ 5, 24, 65, 67, 74, 88, 93] #bad GJ436
#epochs = [10, 13, 32, 48, 66] # bad GJ1148
#epochs = [8, 20, 23, 24, 70, 97,107, 123, 141, 154] #bad Wolf294 (20 as "good standart")
epochs_results = results.epochs
#epochs = epochs_results[10:50:5]
#epochs = [0,9,11,10,12,60]
#epochs = list(set(epochs) & set(results.epochs)) #build intersectionn to make sure epoch
epochs = epochs_results
def improve_order_regularization(r,
o,
star_filename,
tellurics_filename,
training_data,
training_results,
validation_data,
validation_results,
verbose=True,
plot=False,
basename='',
K_star=0,
K_t=0,
L1=True,
L2=True,
tellurics_template_fixed=False):
"""
Use a validation scheme to determine the best regularization parameters for
all model components in a given order r.
Update files at star_filename, tellurics_filename with the best parameters.
By default, this tunes in the following order:
tellurics L2, star L2, tellurics L1, star L1.
Parameters
----------
r : int
Index into `training_data` and `validation_data` to retrieve desired order.
o : int
Index into `star_filename` and `telluric_filename` to retrieve desired order.
star_filename : str
Filename containing regularization amplitudes for the star.
tellurics_filename : str
Filename containing regularization amplitudes for the tellurics.
training_data : wobble.Data object
Data to train template on (should be the majority of available data).
training_results : wobble.Results object
Results object corresponding to `training_data`.
validation_data : wobble.Data object
Data to use in assessing goodness-of-fit for template
(should be a representative minority of the available data).
validation_results : wobble.Results object
Results object corresponding to `validation_data`.
verbose : bool (default `True`)
Toggle print statements and progress bars.
plot : bool (default `False`)
Generate and save plots of fits to validation data.
basename : str (default ``)
String to append to the beginning of saved plots (file path and base).
K_star : int (default `0`)
Number of variable basis vectors for the star.
K_t : int (default `0`)
Number of variable basis vectors for the tellurics.
L1 : bool (default `True`)
Whether to tune L1 amplitudes.
L2 : bool (default `True`)
Whether to tune L2 amplitudes.
tellurics_template_fixed : bool (default `False`)
(currently hard-coded to work with MB's laptop, don't use this!)
Whether to keep tellurics template fixed to values from 51 Peg fit.
"""
training_model = wobble.Model(training_data, training_results, r)
training_model.add_star('star', variable_bases=K_star)
if tellurics_template_fixed: # hackity hack hack
results_51peg = wobble.Results(
filename=
'/Users/mbedell/python/wobble/results/results_51peg_Kstar0_Kt0.hdf5'
)
template_xs = np.copy(results_51peg.tellurics_template_xs[o])
template_ys = np.copy(results_51peg.tellurics_template_ys[o])
training_model.add_telluric('tellurics',
rvs_fixed=True,
template_fixed=True,
variable_bases=K_t,
template_xs=template_xs,
template_ys=template_ys)
else:
training_model.add_telluric('tellurics',
rvs_fixed=True,
variable_bases=K_t)
training_model.setup()
training_model.optimize(niter=0, verbose=verbose, rv_uncertainties=False)
if plot:
n = 0 # epoch to plot
title = 'Initialization'
filename = '{0}_init'.format(basename)
plot_fit(r,
n,
training_data,
training_results,
title=title,
basename=filename)
validation_model = wobble.Model(validation_data, validation_results, r)
validation_model.add_star('star',
variable_bases=K_star,
template_xs=training_results.star_template_xs[r]
) # ensure templates are same size
if tellurics_template_fixed: # hackity hack hack
validation_model.add_telluric(
'tellurics',
rvs_fixed=True,
template_fixed=True,
variable_bases=K_t,
template_xs=training_results.tellurics_template_xs[r],
template_ys=training_results.tellurics_template_ys[r])
else:
validation_model.add_telluric(
'tellurics',
rvs_fixed=True,
variable_bases=K_t,
template_xs=training_results.tellurics_template_xs[r])
validation_model.setup()
# the order in which these are defined will determine the order in which they are optimized:
tensors_to_tune = [
training_model.components[1].L2_template_tensor,
training_model.components[0].L2_template_tensor,
training_model.components[1].L1_template_tensor,
training_model.components[0].L1_template_tensor
]
tensor_names = [
'L2_template', 'L2_template', 'L1_template', 'L1_template'
] # this isonly needed bc TF appends garbage to the end of the tensor name
tensor_components = ['tellurics', 'star', 'tellurics', 'star'] # ^ same
if K_star > 0:
tensors_to_tune = np.append(tensors_to_tune, [
training_model.components[0].L2_basis_vectors_tensor,
training_model.components[0].L1_basis_vectors_tensor
])
tensor_names = np.append(tensor_names,
['L2_basis_vectors', 'L1_basis_vectors'])
tensor_components = np.append(tensor_components, ['star', 'star'])
if K_t > 0:
tensors_to_tune = np.append(tensors_to_tune, [
training_model.components[1].L2_basis_vectors_tensor,
training_model.components[1].L1_basis_vectors_tensor
])
tensor_names = np.append(tensor_names,
['L2_basis_vectors', 'L1_basis_vectors'])
tensor_components = np.append(tensor_components,
['tellurics', 'tellurics'])
regularization_dict = {}
#o_init = max(0, o-1) # initialize from previous order, or if o=0 use defaults
o_init = o # always initialize from starting guess (TODO: decide which init is better)
for i, tensor in enumerate(tensors_to_tune):
if tensor_components[i] == 'star':
filename = star_filename
elif tensor_components[i] == 'tellurics':
filename = tellurics_filename
else:
print("something has gone wrong.")
assert False
with h5py.File(filename, 'r') as f:
regularization_dict[tensor] = np.copy(f[tensor_names[i]][o_init])
i = 0 # track order in which parameters are improved
for component, (tensor, name) in zip(tensor_components,
zip(tensors_to_tune, tensor_names)):
if (name[0:2] == "L1" and L1) or (name[0:2] == "L2" and L2):
i += 1
regularization_dict[tensor] = improve_parameter(
tensor,
training_model,
validation_model,
regularization_dict,
validation_data,
validation_results,
verbose=verbose,
plot=plot,
basename=basename + '_par{0}'.format(i))
if component == 'star':
filename = star_filename
elif component == 'tellurics':
filename = tellurics_filename
else:
print("something has gone wrong.")
assert False
with h5py.File(filename, 'r+') as f:
f[name][o] = np.copy(regularization_dict[tensor])
if plot:
test_regularization_value(tensor,
regularization_dict[tensor],
training_model,
validation_model,
regularization_dict,
validation_data,
validation_results,
plot=False,
verbose=False) # hack to update results
title = 'Final'
filename = '{0}_final'.format(basename)
plot_fit(r,
n,
validation_data,
validation_results,
title=title,
basename=filename)
fig = plt.figure()
ax = fig.add_subplot(111)
val_rvs = validation_results.star_rvs[r] + validation_results.bervs
train_rvs = training_results.star_rvs[r] + training_results.bervs
ax.plot(validation_results.dates, val_rvs - np.mean(val_rvs), 'r.')
ax.plot(training_results.dates,
train_rvs - np.mean(train_rvs),
'k.',
alpha=0.5)
ax.set_ylabel('RV (m/s)')
ax.set_xlabel('JD')
fig.tight_layout()
plt.savefig(basename + '_final_rvs.png')
plt.close(fig)
print("running wobble on star {0} with K_star = {1}, K_t = {2}".format(
starname, K_star, K_t))
start_time = time()
orders = np.arange(72)
'''
e = [ 0, 1, 6, 7, 9, 17, 18, 19, 21, 23, 24, 26, 30, 33, 34, 35, 36,
37, 38, 40, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 53, 55, 56, 61,
66, 69, 70, 72, 73, 75] # night of August 28, 2007
data = wobble.Data(starname+'_e2ds.hdf5', filepath='../data/', orders=orders, epochs=e)
'''
data = wobble.Data(starname + '_e2ds.hdf5',
filepath='../data/',
orders=orders)
orders = np.copy(data.orders)
results = wobble.Results(data=data)
results_51peg = wobble.Results(
filename='../results/results_51peg_Kstar0_Kt0.hdf5')
print("data loaded")
print("time elapsed: {0:.2f} min".format((time() - start_time) / 60.0))
elapsed_time = time() - start_time
if plots:
print("plots will be saved under directory: {0}".format(plot_dir))
if not os.path.exists(plot_dir):
os.makedirs(plot_dir)
star_learning_rate = 0.1
telluric_learning_rate = 0.01
for r, o in enumerate(orders):
def run_wobble(parameters):
p = parameters
results_name = 'results_{0}_Kstar{1}_Kt{2}_'.format(
p.starname, p.K_star, p.K_t, p.niter) + p.output_suffix
results_file_base = p.results_dir + results_name
results_file = results_file_base + '.hdf5'
data_file = p.data_file = p.data_dir + p.starname + p.data_suffix + '_e2ds.hdf5'
temp_dir = p.temp_dir = p.results_dir + '/temp_' + results_name + '/'
plot_dir = p.plot_dir = p.results_dir + '/plots_' + results_name + '/'
#make (output) directory
#TODO these data and results dirs should be handlesd somewhere else
os.makedirs(p.results_dir, exist_ok=True)
os.makedirs(p.data_dir, exist_ok=True)
os.makedirs(temp_dir, exist_ok=True)
os.makedirs(plot_dir, exist_ok=True)
start_time = p.start_time = time()
#generate epoch list
# if parameters has been passed a global epochs list use this going forward. This is primarily used by regularization.py
try: #skipping the except will make major issues unless dropped epochs and orders are already handled
epochs_list = p.epochs_list = p.global_epochs_list
p.drop_orders = data.drop_orders # this check here only works if data object is not yet initialized with empty drop order list.
except (AttributeError, UnboundLocalError) as e:
print(e)
print("Loading data. May take a few minutes")
try:
data = wobble.Data(data_file,
orders=np.arange(p.start, p.end),
min_flux=10**-5,
min_snr=p.min_snr,
parameters=p)
epochs_list = p.epochs_list = data.epochs.tolist()
except wobble.data.AllDataDropped:
p.min_snr = 5
print("restarting with min_snr = {}".format(p.min_snr))
data = wobble.Data(data_file,
orders=np.arange(p.start, p.end),
min_flux=10**-5,
min_snr=p.min_snr,
parameters=p)
epochs_list = p.epochs_list = data.epochs.tolist()
#orders_list = p.orders_list = data.orders.tolist() #too agressive in visible # TODO implement alternate nir and vis handling
try:
p.drop_orders = data.drop_orders
print("data.drop_orders", data.drop_orders)
except AttributeError:
print("data.drop_orders is not defined")
base_orders_list = np.arange(p.start, p.end).tolist()
orders_list = p.orders_list = [
x for x in base_orders_list if x not in p.drop_orders
]
print("orders_list", orders_list)
chunks = p.chunks = chunk_list(p.start, p.end, p.chunk_size, p.orders_list)
print("Chunks: ", chunks)
#Loop over chunks
for i in range(len(chunks)):
#pass parameters object to chunk script
p.i = i
with open(
os.path.dirname(os.path.abspath(__file__)) +
"/carmenes_aux_files/chunk_parameters.pkl", "wb") as f:
dill.dump(p, f)
#start chunk script
os.system("python3 {0}/chunk.py".format(
os.path.dirname(os.path.abspath(__file__))))
'''HACK to be removed
#import parameters back after possible changes in chunk
with open(os.path.dirname(os.path.abspath(__file__)) + "/" + "carmenes_aux_files/chunk_parameters.pkl", "rb") as f:
p = dill.load(f)
'''
print("all chunks optimized: writing combined file")
results_file_stitch(p.chunks, results_file, temp_dir)
#Combine orders
results = wobble.Results(filename=results_file)
results.combine_orders('star')
print("final RVs calculated.")
print("time elapsed: {0:.2f} minutes".format((time() - start_time) / 60.0))
results.write(results_file)
append_dates_utc(
results_file, data_file
) # cannot be done before results.write, as .write will remove dates_utc
append_parameters(p, results_file)
'''
# test
loaded_parameters = read_parameters_from_results(results_file)
attrs = dir(loaded_parameters)
print(attrs)
for attr in attrs:
#print(attr, getattr(loaded_parameters, attr))
q = getattr(p, attr) == getattr(loaded_parameters, attr)
if q == True:
print(q)
else:
print(attr, q)
'''
print("results saved as: {0}".format(results_file))
print("time elapsed: {0:.2f} minutes".format((time() - start_time) / 60.0))
#delete temp_dir which at this point only contains duplicates
shutil.rmtree(temp_dir)
print("deleted: {0}".format(temp_dir))