def Update(Bodies, Steps=100, Snaps=10, L=3, njobs=1, Plots=False, direct=''):
'''
This function updates the system.
In Args:
- Bodies: Array containing the bodies to be updated.
- *Steps: (Optional) The number of update steps. If not given, it will update 100 steps.
- *Snaps: (Optional) The number of snapshots to take. This will only be useful when Plots = True.
- *L = 3 The scale of the system. Default is 3 AU.
- *Plots (Optional) If true, will save certain stapshots on the given directory.
- *direct (Needed if Plots = True) A string with the Path where you want to save the outputs.
- *njobs: (Optional) The number of cores to be used during the computation. By default is 1
i.e. the program will run sequentially.
Out Args:
- Bodies: Array containing the updated bodies.
'''
j = 0 # This is a counter for the number of snapshots.
for k in tqdm(range(Steps)):
time.sleep(0.01)
if Plots == True:
N_steps_snapshot = int(Steps / Snaps)
if k % N_steps_snapshot == 0: # This will plot only the necesary steps.
Visualization.Plot(
Bodies, k, j, L, direct
) # This will make a plot of the actual step of the simulation.
j += 1
B = Bodies # This is a "temporal array" that will serve to compute the acceleration in parallel
# without updating the positions of the bodies before we compute all the accelerations.
Parallel(n_jobs=njobs)(delayed(Compute_Accel)(i, B) for i in Bodies)
for i in range(len(Bodies)):
Bodies[i].velocity = Bodies[i].Compute_velocity()
Bodies[i].position = Bodies[i].Compute_position()
# Once the main parameters have been updated we define the acceleration (i) equal to the
# computed acceleration (i+1).
Bodies[i].acceleration_i = Bodies[i].acceleration_i_1
return Bodies
import PreProcessing
import pickle
import numpy as np
import Visualization
model = pickle.load(open("Model7", 'rb'))
score = model.evaluate(PreProcessing.testX, PreProcessing.testY, verbose=0)
print('Test loss:', score[0])
Original_data = PreProcessing.scaler.inverse_transform(PreProcessing.features)
train = PreProcessing.scaler.inverse_transform(PreProcessing.train)
validation = PreProcessing.scaler.inverse_transform(PreProcessing.val)
testPredict = PreProcessing.scaler.inverse_transform(
model.predict(PreProcessing.testX))
validationPlot = np.empty_like(PreProcessing.features)
validationPlot[:, :] = np.nan
validationPlot[len(train):len(PreProcessing.features) -
len(PreProcessing.test), :] = validation
testPredictPlot = np.empty_like(PreProcessing.features)
testPredictPlot[:, :] = np.nan
testPredictPlot[len(PreProcessing.features) - len(PreProcessing.test) +
1:len(PreProcessing.features) - 1, :] = testPredict
Visualization.Plot(Original_data, train, validationPlot, testPredictPlot)