main.py

# MIT License
#
# Copyright (c) 2017 Luca Angioloni
#
# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in all
# copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.

import numpy as np
from time import time
from keras import optimizers, callbacks
from timeit import default_timer as timer
from dataset import get_dataset_reshaped, split_dataset, get_resphaped_dataset_paper, get_cb513, is_filtered
import model

import pickle

filtered = is_filtered()

do_log = True
stop_early = False
show_plots = True

start_time = timer()

print("Collecting Dataset...")

if filtered:
    # Split the dataset in 0.8 train, 0.1 test, 0.1 validation with shuffle (optionally seed)
    X_train, X_val, X_test, Y_train, Y_val, Y_test = get_dataset_reshaped(seed=100)
else:
    # Slit the dataset with the same indexes used in the paper (Only for CullPDB6133 not filtered)
    X_train, X_val, X_test, Y_train, Y_val, Y_test = get_resphaped_dataset_paper()

end_time = timer()
print("\n\nTime elapsed getting Dataset: " + "{0:.2f}".format((end_time - start_time)) + " s")

if filtered:
    print("Using CullPDB Filtered dataset")

net = model.CNN_model()

start_time = timer()

history = None

call_b = [model.checkpoint]

if filtered:
    logDir = "logs/CullPDB_Filtered/{}".format(time())
else:
    logDir = "logs/CullPDB/{}".format(time())

if do_log:
    call_b.append(callbacks.TensorBoard(log_dir=logDir, histogram_freq=0, write_graph=True))

if stop_early:
    call_b.append(model.early_stop)

history = net.fit(X_train, Y_train, epochs=model.nn_epochs, batch_size=model.batch_dim, shuffle=True,
                        validation_data=(X_val, Y_val), callbacks=call_b)

end_time = timer()
print("\n\nTime elapsed: " + "{0:.2f}".format((end_time - start_time)) + " s")

scores = net.evaluate(X_test, Y_test)
print("Loss: " + str(scores[0]) + ", Accuracy: " + str(scores[1]) + ", MAE: " + str(scores[2]))
#print(scores)

CB_x, CB_y = get_cb513()

cb_scores = net.evaluate(CB_x, CB_y)
print("CB513 -- Loss: " + str(cb_scores[0]) + ", Accuracy: " + str(cb_scores[1]) + ", MAE: " + str(cb_scores[2]))

pickle_out = open("lasthistory.pickle","wb")
pickle.dump(history, pickle_out)
pickle_out.close()

if show_plots:
    from plot_history import plot_history
    plot_history(history)