def get_compressed(year,month,day,hour):
prmsl=twcr.load('prmsl',datetime.datetime(year,month,day,hour),
version='2c')
prmsl=to_analysis_grid(prmsl.extract(iris.Constraint(member=1)))
t2m=twcr.load('air.2m',datetime.datetime(year,month,day,hour),
version='2c')
t2m=to_analysis_grid(t2m.extract(iris.Constraint(member=1)))
u10m=twcr.load('uwnd.10m',datetime.datetime(year,month,day,hour),
version='2c')
u10m=to_analysis_grid(u10m.extract(iris.Constraint(member=1)))
v10m=twcr.load('vwnd.10m',datetime.datetime(year,month,day,hour),
version='2c')
v10m=to_analysis_grid(v10m.extract(iris.Constraint(member=1)))
insol=to_analysis_grid(load_insolation(year,month,day,hour))
# Convert the validation data into tensor format
t2m_t = tf.convert_to_tensor(normalise_t2m(t2m.data),numpy.float32)
t2m_t = tf.reshape(t2m_t,[79,159,1])
prmsl_t = tf.convert_to_tensor(normalise_prmsl(prmsl.data),numpy.float32)
prmsl_t = tf.reshape(prmsl_t,[79,159,1])
u10m_t = tf.convert_to_tensor(normalise_wind(u10m.data),numpy.float32)
u10m_t = tf.reshape(u10m_t,[79,159,1])
v10m_t = tf.convert_to_tensor(normalise_wind(v10m.data),numpy.float32)
v10m_t = tf.reshape(v10m_t,[79,159,1])
insol_t = tf.convert_to_tensor(normalise_insolation(insol.data),numpy.float32)
insol_t = tf.reshape(insol_t,[79,159,1])
ict = tf.concat([t2m_t,prmsl_t,u10m_t,v10m_t,insol_t],2) # Now [79,159,5]
ict = tf.reshape(ict,[1,79,159,5])
result = autoencoder.predict_on_batch(ict)
result = tf.reshape(result,[79,159,5])
ls = encoder.predict_on_batch(ict)
# Convert the encoded fields back to unnormalised cubes
t2m_r=t2m.copy()
t2m_r.data = tf.reshape(result.numpy()[:,:,0],[79,159]).numpy()
t2m_r.data = unnormalise_t2m(t2m_r.data)
prmsl_r=prmsl.copy()
prmsl_r.data = tf.reshape(result.numpy()[:,:,1],[79,159]).numpy()
prmsl_r.data = unnormalise_prmsl(prmsl_r.data)
u10m_r=u10m.copy()
u10m_r.data = tf.reshape(result.numpy()[:,:,2],[79,159]).numpy()
u10m_r.data = unnormalise_wind(u10m_r.data)
v10m_r=v10m.copy()
v10m_r.data = tf.reshape(result.numpy()[:,:,3],[79,159]).numpy()
v10m_r.data = unnormalise_wind(v10m_r.data)
return {'t2m':t2m_r,'prmsl':prmsl_r,'u10m':u10m_r,'v10m':v10m_r,'ls':ls}
def random_state():
ls = tf.convert_to_tensor(numpy.random.normal(size=100), numpy.float32)
ls = tf.reshape(ls, [1, 100])
result = generator.predict_on_batch(ls)
result = tf.reshape(result, [79, 159, 5])
t2m = dummy_cube()
t2m.data = tf.reshape(result.numpy()[:, :, 0], [79, 159]).numpy()
t2m.data = unnormalise_t2m(t2m.data)
prmsl = dummy_cube()
prmsl.data = tf.reshape(result.numpy()[:, :, 1], [79, 159]).numpy()
prmsl.data = unnormalise_prmsl(prmsl.data)
u10m = dummy_cube()
u10m.data = tf.reshape(result.numpy()[:, :, 2], [79, 159]).numpy()
u10m.data = unnormalise_wind(u10m.data)
v10m = dummy_cube()
v10m.data = tf.reshape(result.numpy()[:, :, 3], [79, 159]).numpy()
v10m.data = unnormalise_wind(v10m.data)
return (t2m, prmsl, u10m, v10m)
def ls_load_at_timepoint(year, month, day, hour):
pfile = ("%s/ML_GCM/GCM_mucdf/" + "%04d-%02d-%02d:%02d.pkl") % (
os.getenv('SCRATCH'), year, month, day, hour)
res = pickle.load(open(pfile, 'rb'))
ls = res['latent_s']
t2m = dummy_cube()
t2m.data = res['state_v'][0, :, :, 0]
t2m.data = unnormalise_t2m(t2m.data)
prmsl = dummy_cube()
prmsl.data = res['state_v'][0, :, :, 1]
prmsl.data = unnormalise_prmsl(prmsl.data)
u10m = dummy_cube()
u10m.data = res['state_v'][0, :, :, 2]
u10m.data = unnormalise_wind(u10m.data)
v10m = dummy_cube()
v10m.data = res['state_v'][0, :, :, 3]
v10m.data = unnormalise_wind(v10m.data)
return (ls, t2m, prmsl, u10m, v10m)
insol_t = tf.convert_to_tensor(normalise_insolation(insol.data), numpy.float32)
insol_t = tf.reshape(insol_t, [79, 159, 1])
# Get autoencoded versions of the validation data
model_save_file = ("%s/ML_GCM/autoencoder.tst/" + "Epoch_%04d/autoencoder") % (
os.getenv('SCRATCH'), args.epoch)
autoencoder = tf.keras.models.load_model(model_save_file, compile=False)
ict = tf.concat([t2m_t, prmsl_t, u10m_t, v10m_t, insol_t], 2) # Now [79,159,5]
ict = tf.reshape(ict, [1, 79, 159, 5])
result = autoencoder.predict_on_batch(ict)
result = tf.reshape(result, [79, 159, 5])
# Convert the encoded fields back to unnormalised cubes
t2m_r = t2m.copy()
t2m_r.data = tf.reshape(result.numpy()[:, :, 0], [79, 159]).numpy()
t2m_r.data = unnormalise_t2m(t2m_r.data)
prmsl_r = prmsl.copy()
prmsl_r.data = tf.reshape(result.numpy()[:, :, 1], [79, 159]).numpy()
prmsl_r.data = unnormalise_prmsl(prmsl_r.data)
u10m_r = u10m.copy()
u10m_r.data = tf.reshape(result.numpy()[:, :, 2], [79, 159]).numpy()
u10m_r.data = unnormalise_wind(u10m_r.data)
v10m_r = v10m.copy()
v10m_r.data = tf.reshape(result.numpy()[:, :, 3], [79, 159]).numpy()
v10m_r.data = unnormalise_wind(v10m_r.data)
# Plot the two fields and a scatterplot for each variable
fig = Figure(figsize=(9.6 * 1.2, 10.8),
dpi=100,
facecolor=(0.88, 0.88, 0.88, 1),
edgecolor=None,
return (dummy_cube)
# Load the latent-space representation, and convert it back into normal space
model_save_file = ("%s/ML_GCM/predictor.delta/" +
"Epoch_%04d/generator") % (os.getenv('SCRATCH'), args.epoch)
generator = tf.keras.models.load_model(model_save_file, compile=False)
# Random latent state
ls = tf.convert_to_tensor(numpy.random.normal(size=100), numpy.float32)
ls = tf.reshape(ls, [1, 100])
result = generator.predict_on_batch(ls)
result = tf.reshape(result, [79, 159, 5])
t2m = dummy_cube()
t2m.data = tf.reshape(result.numpy()[:, :, 0], [79, 159]).numpy()
t2m.data = unnormalise_t2m(t2m.data)
prmsl = dummy_cube()
prmsl.data = tf.reshape(result.numpy()[:, :, 1], [79, 159]).numpy()
prmsl.data = unnormalise_prmsl(prmsl.data)
u10m = dummy_cube()
u10m.data = tf.reshape(result.numpy()[:, :, 2], [79, 159]).numpy()
u10m.data = unnormalise_wind(u10m.data)
v10m = dummy_cube()
v10m.data = tf.reshape(result.numpy()[:, :, 3], [79, 159]).numpy()
v10m.data = unnormalise_wind(v10m.data)
mask = iris.load_cube("%s/fixed_fields/land_mask/opfc_global_2019.nc" %
os.getenv('SCRATCH'))
# Define the figure (page size, background color, resolution, ...
fig = Figure(