def __getitem__(self, index):
file = self.videos[index]
video = Video(file)
align = Align(
os.path.join(self.align_path, video.speaker, 'align',
video.name + ".align"))
return {
'video_input':
video.get_frames(0, video.n_frames, self.video_padding_length),
'video_length':
self.video_padding_length,
'targets_input':
align.sentence(),
'targets_length':
align.sentence_length,
}
from sequence import Sequence, loadFasta
from score import PSSM
from align import Align
pssm = PSSM("WW domain")
for seq in loadFasta(r"../resources/fasta/msaresults-MUSCLE.fasta"):
pssm.add(seq)
pssm.setGapPenalty(4)
al = Align(pssm)
for toalign in loadFasta(r"../resources/fasta/test.fasta"):
for aligned in al.multiAlign(toalign):
print(aligned)
def main():
# m = examples["diff"]
# print PermanentSampler(0.5,0.1).estimate_permanent(m)
# r_perm = ryser(m)
# #n_perm = naive_permanent(m)
# print r_perm
# #print n_perm
from align import Align
from align import Alignment
import sys
mode = model_names[sys.argv[3]]
align = Align.from_files("data/eng-fr.full.fr", "data/eng-fr.full.en")
e_types, f_types = align.types()
prob_model = dist.ProbModel(f_types, e_types)
test_align = Align.from_files("data/eng-fr.dev.fr", "data/eng-fr.dev.en",
True)
gold_align = Alignment.read_alignments("data/eng-fr.dev.align")
if sys.argv[1] == "train":
dists = prob_model.rand_dists_sparse()
instances = align.instances()
test_instances = test_align.instances()
for r in range(20):
if mode == MANYTOONE:
score, _ = dev_assess(test_instances, gold_align, dists,
prob_model, viterbi_align_manyone)
else:
score, _ = dev_assess(test_instances, gold_align, dists,
prob_model, viterbi_align_oneone)
dists = em(instances, dists, mode, prob_model)
#print "Dist 'le'"
#for i, lscore in enumerate(dists['le']):
# if lscore > 1e-4:
# print i, align.eng_to_ind(i), lscore
print "Score is:", score
final_dist = dists
pickle.dump(final_dist, open(sys.argv[2], 'wb'))
elif sys.argv[1] == "test":
dists = pickle.load(open(sys.argv[2], 'rb'))
instances = test_align.instances()
if mode == MANYTOONE:
score, alignments = dev_assess(instances, gold_align, dists,
prob_model, viterbi_align_manyone)
else:
score, alignments = dev_assess(instances, gold_align, dists,
prob_model, viterbi_align_oneone)
print score
f_out = open("out.f", 'w')
e_out = open("out.e", 'w')
a_out = open("out.a", 'w')
gold_out = open("out.gold.a", 'w')
for ins, align in zip(instances, alignments):
print >> f_out, " ".join(ins.f)
print >> e_out, " ".join(ins.e)
print >> a_out, " ".join([str(e) + "-" + str(f) for e, f in align])
print >> gold_out, " ".join(
[str(e) + "-" + str(f) for e, f in gold_align[ins.num]])
from align import Align
from mongoConnect import tasksDB
import os
import time
Align.centerAlign("Welcome to the Manager")
print("\n")
Align.centerAlign("Tasks Remaining")
print("\n")
time.sleep(1)
tasksDB.showTask()
print("want to add new task?(y/n)")
ans = input()
while (ans == 'y'):
print("Enter new task")
newtask = input()
tasksDB.addTask(newtask)
print("\nwant to add more new task?(y/n)")
ans = input()
print("press any key to exit!")
input()
os.system('exit')
def FaceRecognition(imagePath):
W,W1,FisherMatrix,mean,trainEncodings,labels,label1 = loadModel()
threshold = 3300
img = cv2.imread(imagePath)
try:
t = face_recognition.face_locations(img, number_of_times_to_upsample=1)
except:
t = face_recognition.face_locations(img, number_of_times_to_upsample=1, model="cnn")
# loop over detected faces
if(len(t)==0):
return True
#timestamp check
else:
now = datetime.datetime.now()
time9am = now.replace(hour=9, minute=0, second=0, microsecond=0)
time6pm = now.replace(hour=18, minute=0, second=0, microsecond=0)
# if now < time9am or now > time6pm :
# print("Activate alert alarm") # add the thread related alert system
# return False
try:
im = PIL.Image.open(imagePath)
# im = im.resize((80,80))
im = im.convert('RGB')
img = np.array(im)
# print("ff")
align = Align(predictor_model)
img=align.align(80,img)
test_encoding = face_recognition.face_encodings(img, known_face_locations=[[0, 80,80, 0]] )[0]
results = face_recognition.compare_faces(trainEncodings,test_encoding,0.4)
for i in range(len(results)):
if(results[i]==True):
print(labels[i])
return True
return False
except:
img = cv2.imread(imagePath)
for face in t:
x = t[0][3]
y = t[0][0]
w = t[0][1] - x
h = t[0][2] - y
if(x<0):
x=0
if(y<0):
y=0
img = img[y:y+h, x:x+w]
# print(img)
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# print(img.shape)
img = cv2.resize(img,(80,80))
# print(img)
img = img.flatten()
mindist=1e19
deviationFromMean = (img-mean)
eigTestimg = np.dot(img,W)
FisherTestImg = np.dot(eigTestimg,W1)
for i in range(0,FisherMatrix.shape[0]):
dist = np.linalg.norm(FisherMatrix[i]-FisherTestImg)
# print(dist,mindist,dataTrainY[i])
if(dist<mindist):
# label = dataTrainY[i]
mindist = dist
ind = i
if(mindist>threshold):
return False
else:
print(label1[ind])
return True
from align import Align
from mongoConnect import usersDB
import os,time
Align.centerAlign("Register To the Manager")
print("Enter Email ID:")
email = input()
print("Enter Password:"******"cls")
os.system("python login.py")
def main():
# m = examples["diff"]
# print PermanentSampler(0.5,0.1).estimate_permanent(m)
# r_perm = ryser(m)
# #n_perm = naive_permanent(m)
# print r_perm
# #print n_perm
from align import Align
from align import Alignment
import sys
mode = model_names[sys.argv[3]]
align = Align.from_files("data/eng-fr.full.fr","data/eng-fr.full.en")
e_types, f_types = align.types()
prob_model = dist.ProbModel(f_types, e_types)
test_align = Align.from_files("data/eng-fr.dev.fr","data/eng-fr.dev.en", True)
gold_align= Alignment.read_alignments("data/eng-fr.dev.align")
if sys.argv[1] == "train":
dists = prob_model.rand_dists_sparse()
instances = align.instances()
test_instances = test_align.instances()
for r in range(20):
if mode == MANYTOONE:
score,_ = dev_assess(test_instances, gold_align, dists, prob_model, viterbi_align_manyone)
else:
score,_ = dev_assess(test_instances, gold_align, dists, prob_model, viterbi_align_oneone)
dists = em(instances, dists, mode, prob_model)
#print "Dist 'le'"
#for i, lscore in enumerate(dists['le']):
# if lscore > 1e-4:
# print i, align.eng_to_ind(i), lscore
print "Score is:", score
final_dist = dists
pickle.dump(final_dist, open(sys.argv[2], 'wb'))
elif sys.argv[1] == "test":
dists = pickle.load(open(sys.argv[2], 'rb'))
instances = test_align.instances()
if mode == MANYTOONE:
score,alignments = dev_assess(instances, gold_align, dists, prob_model, viterbi_align_manyone)
else:
score,alignments = dev_assess(instances, gold_align, dists, prob_model, viterbi_align_oneone)
print score
f_out = open("out.f", 'w')
e_out = open("out.e", 'w')
a_out = open("out.a", 'w')
gold_out = open("out.gold.a", 'w')
for ins,align in zip(instances,alignments):
print >>f_out, " ".join( ins.f)
print >>e_out, " ".join(ins.e)
print >>a_out, " ".join([str(e)+"-"+str(f) for e, f in align])
print >>gold_out, " ".join([str(e)+"-"+str(f) for e, f in gold_align[ins.num]])
import os, time
from align import Align
Align.centerAlign("Python Daily Manager")
print("do you want to login(L) or register(R)")
ans = input()
if (ans == 'L'):
os.system('cls')
os.system('python login.py')
elif (ans == 'R'):
os.system('cls')
os.system('python register.py')
else:
print("Wrong Choice Entered")
import mongoConnect
import os, time
from mongoConnect import verify, names
from align import Align
# print(mongoConnect.colNames())
# print(names.colNames())
# print(verify.checkUser(email))
Align.centerAlign("Login To the Manager")
print("Enter Email ID:")
email = input()
print("Enter Password:"******"Wrong Password")
else:
print("User Does'nt Exist!!")
print("Do you want to register yourself(y/n)")
ans = input()
if (ans == 'y'):
os.system('cls')
os.system('python register.py')