def BCAFunction(df, tol, list_treat, list_score, verbose, split):
print('----------------------')
print('')
print('BCA Method')
print('')
print('----------------------')
start = datetime.now()
step = 0
if verbose:
print('step : ', step)
# list_treat = list(df.columns.values)
list_delete = list(set(list(df.columns.values)) - set(list_treat))
df = df.drop(list_delete, axis=1)
features = list(df.columns.values)
features.remove('Label')
df_X = df[features]
N = df_X.shape[1]
df_X_init = df_X.copy()
df_Y = df['Label']
condition = True
X_opt = np.zeros(N)
y_opt = score(X_opt, df_X_init, df_Y, split)
while condition:
step += 1
for i in range(N):
X = X_opt.copy()
X[i] = int(not X_opt[i])
score_X = score(X, df_X_init, df_Y, split)
score_X_opt = score(X_opt, df_X_init, df_Y, split)
if score_X >= score_X_opt:
if score_X > list_score[-1]:
print('score : ', score(X, df_X_init, df_Y, split))
list_score.append(score_X)
else:
list_score.append(list_score[-1])
X_opt = X.copy()
else:
if score_X > list_score[-1]:
list_score.append(score_X)
else:
list_score.append(list_score[-1])
y = score(X_opt, df_X_init, df_Y, split)
if verbose:
print('step : {0}\ny_opt : {1}\ny :{2}'.format(step, y_opt, y))
if abs(y - y_opt) < tol or step > 10:
condition = False
y_opt = y
list_features = selec_feat(X_opt, df_X_init, False)
if verbose:
print('It took : ', datetime.now() - start)
return X_opt, y_opt, list_features, df_X_init, df_Y, list_score
def random_sampling(protein, configurations, fold_factor): protein_coor = get_coor_array(protein) len_protein = len(protein_coor) print len_protein protein_array = [] j = 0 while j < configurations: folds = random.randint(0, fold_factor*len_protein) print folds start_coor = copy.deepcopy(protein_coor) # print start_coor for i in range(folds): new_coor = folder_protein(start_coor, random.randint(2,len_protein), random.randint(1,3)) # print 'made fold' start_coor = copy.deepcopy(new_coor) new_protein = update_objects(protein, new_coor) # print 'made protein' grid = functions.protein_visual(new_protein) score = functions.score(new_protein, grid) protein_info = [score, copy.deepcopy(new_protein)] # print new_protein if protein_info not in protein_array: # print 'unique' protein_array.append(protein_info) j += 1 else: # print 'not unique' print protein_info print protein_array return protein_array
def evaluate_(self):
if self.EVALUATETeamui.comboBox.currentIndex() == 0:
error_dialog = QtWidgets.QErrorMessage()
error_dialog.showMessage('Oops! please select match.')
error_dialog.exec_()
elif self.EVALUATETeamui.comboBox_2.currentIndex() == 0:
error_dialog = QtWidgets.QErrorMessage()
error_dialog.showMessage('Oops! please select team.')
error_dialog.exec_()
else:
player_data.cur.execute(
"SELECT * FROM teams WHERE name=?;",
(self.EVALUATETeamui.comboBox_2.currentText(), ))
player_ = []
for temp in tuple(player_data.cur.fetchall()):
player_.append(temp[1])
score = 0
for temp in player_:
player_data.cur.execute(
"SELECT * FROM " +
self.EVALUATETeamui.comboBox.currentText() +
''' WHERE player="''' + temp + '''";''')
t = functions.score(player_data.cur.fetchone())
score += t
self.EVALUATETeamui.label_5.setText(str(score))
def ajax_search(request):
try:
query = request.POST[u'query']
except KeyError:
try:
query = request.GET[u'query']
except KeyError:
return HttpResponse(u'')
tokens = re.split(ur'(?u)[^-\w]', query)
while u'' in tokens:
tokens.remove(u'')
candidates = []
for candidate in models.Entity.objects.all():
candidates.append([candidate, 0])
for token in tokens:
new_candidates = []
for candidate in candidates:
if score(candidate[0], token) > 0:
candidate[1] += score(candidate[0], token)
new_candidates.append(candidate)
candidates = new_candidates
candidates.sort(lambda a, b: -cmp(a[1], b[1]))
export = []
for candidate in candidates:
if candidate[0].image_mimetype:
image = True
else:
image = False
export.append({
u'department': name,
u'description': candidate[0].description,
u'id': candidate[0].id,
u'image': image,
u'name': candidate[0].name,
u'title': candidate[0].title,
})
first_portion = export[:settings.INITIAL_RESULTS]
second_portion = export[settings.INITIAL_RESULTS:]
return render_to_response(u'search_internal.html', {
u'first_portion': first_portion,
u'second_portion': second_portion,
u'query': urllib.quote(query),
})
def find_best(protein_object, rep_num, high_score): # array with times and iterations result_array = [] start = timer() result_array.append(['start', start]) high_protein_object = copy.deepcopy(protein_object) high_coor_array = [] coor_array = get_coor_array(protein_object) count = 0 rep_count = 0 # folds random amino acid rep_num times # rep_count checks if protein changes and returns # best protein if protein stays the same after 30 folds while count < rep_num and rep_count < 50: start_ran = timer() - start fold_coor = folder_protein(coor_array, random.randint(2, len(coor_array) - 1)) protein_object_new = update_objects(protein_object, fold_coor) grid = functions.protein_visual(protein_object_new) score = functions.score(protein_object_new, grid) # checks if coordinates stay the same (no change) if fold_coor == coor_array: rep_count += 1 else: coor_array = copy.deepcopy(fold_coor) count += 1 # updates high score and object that belongs to high score if score < high_score: high_score = copy.copy(score) high_protein_object = copy.deepcopy(protein_object_new) high_coor_array = copy.deepcopy(coor_array) # adds end time and score after every fold set completion end_ran = timer() - start result_array.append(['best ran', start_ran, end_ran, count + rep_count, high_score]) end = timer() result_array.append(['end', end]) return [high_score, high_protein_object]
def change_(self):
self.EVALUATETeamui.listWidget_2.clear()
if self.EVALUATETeamui.comboBox.currentIndex() != 0:
player_data.cur.execute(
"SELECT * FROM teams WHERE name=?;",
(self.EVALUATETeamui.comboBox_2.currentText(), ))
player_ = []
for temp in tuple(player_data.cur.fetchall()):
player_.append(temp[1])
if self.EVALUATETeamui.comboBox_2.currentIndex() != 0:
for temp in player_:
player_data.cur.execute(
"SELECT * FROM " +
self.EVALUATETeamui.comboBox.currentText() +
''' WHERE player="''' + temp + '''";''')
self.EVALUATETeamui.listWidget_2.addItem(
str(functions.score(player_data.cur.fetchone())))
self.EVALUATETeamui.label_5.setText(' ##')
# COMMON PRIMER DESIGN
# Forward Common Primers
# Common forward primer for all reactions are send for optimization.
forwcom, forwcom_left, forwcom_right = justtrim(
nucsequence, nucsequence[common_fwd_pos[0]:common_fwd_pos[1]],
common_fwd_pos[0], common_fwd_pos[1], tm_target)
if forwcom_left:
for forward_common, pos in zip([forwcom, forwcom_left, forwcom_right],
['both', 'left', 'right']):
tm_common_forward = tmcalculator(forward_common)
primers[pos]['forward'] = [forward_common, None, tm_common_forward]
a, b, c, d = score(
forward_common, tm_target,
1) # 1 is a flag to score function telling no NNS in primer.
df.loc[len(df)] = [
0, 'Forward Common (' + pos + ')',
str(forward_common),
reversecomplement(forward_common, 0),
str(tm_common_forward),
gccontent(forward_common),
len(forward_common), a, b, c, np.nan, d
]
else:
tm_common_forward = tmcalculator(forwcom)
primers['both']['forward'] = \
primers['right']['forward'] = \
primers['left']['forward'] = [forwcom, None, tm_common_forward]
def anneal(protein_object, name, temperature): result_array = [] cur_score = 0 start = timer() result_array.append(['start', start]) # make starting point # cur_protein_array = folder_iter.find_best(protein_object, rep_num, 0) # cur_score = cur_protein_array[0] cur_protein = copy.deepcopy(protein_object) cur_coor = folder_iter.get_coor_array(cur_protein) start_temp = temperature #set high_score high_score = cur_score high_protein = copy.deepcopy(cur_protein) j = 0 # i = 0 counter = 0 while j < 15: start_an = timer() - start # functions.Visualizer2D(cur_protein, 'HHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH', cur_score, "test%s+%s" %(i,counter)) # i+=1 result_array.append(['Start Anneal', start_an, temperature]) while temperature > 0.01: # functions.Visualizer2D(cur_protein, 'HHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH', cur_score, "test%s+%s" %(i,counter)) # i+=1 # make next protein object next_coor = folder_iter.folder_protein(cur_coor, random.randint(2,len(cur_protein) - 1), random.randint(1,3)) next_protein = folder_iter.update_objects(cur_protein, next_coor) next_grid = functions.protein_visual(next_protein) next_score = functions.score(next_protein, next_grid) # calculate acceptation_chance acceptation_chance = calc_chance(next_score, cur_score, temperature) # print acceptation_chance if acceptation_chance < temperature: # print "CHANGE" cur_score = next_score cur_protein = copy.deepcopy(next_protein) cur_coor = copy.deepcopy(next_coor) # temperature is updated (not in function, because it is more effort, lines of code and time) temperature -= temperature * 0.005 # print "temperature:" + str(temperature) # print temperature if cur_score < high_score: # print "NEW HIGHSCORE" high_score = copy.copy(cur_score) high_protein = copy.deepcopy(cur_protein) result_array.append(['Score Iteration', high_score, counter, temperature]) counter += 1 # print cur_score, next_score, high_score, acceptation_chance cur_protein = copy.deepcopy(high_protein) cur_score = high_score cur_coor = folder_iter.get_coor_array(high_protein) start_temp = start_temp * 1.1 temperature = start_temp end = timer() - start result_array.append(['Result SA %s' %name, start_an, end, counter, high_score]) j+=1 # functions.Visualizer2D(cur_protein, main_SA.protein, cur_score, "test") return [high_score, high_protein, result_array]
corr.sort_values(key=lambda x: abs(x), ascending=False))
DO_LIN_FIT = 0
DO_NN_FIT = 1
DO_NNLIN_FIT = 0
if DO_LIN_FIT:
print("-----------------------------------------")
print(" ######## LIN MODEL ########")
print("-----------------------------------------")
### Linear Model
linModel, w = F.linRegression(x_train, y_train, x_test, y_test)
w.plot.barh()
plt.title("Weights")
linPred = linModel.predict(x_test)
scoreLin = F.score(linPred, y_test)
print("Score:", np.round(scoreLin, 3))
## linear shap
explainer = shap.LinearExplainer(linModel,
x_train,
feature_dependence="independent")
shap_valsLin = explainer.shap_values(x_test)
## summary plot
plt.figure()
shap.summary_plot(shap_valsLin, x_test.values, x_train.columns, alpha=0.5)
## one sample person
# i=1
# shap.force_plot(explainer.expected_value,shap_valsLin[i,:],x_test.values[i,:],x_test.columns,matplotlib=True)
shap.force_plot(explainer.expected_value, shap_valsLin, x_test.values,
def homepage(request):
id = u'null'
try:
query = request.REQUEST[u'query']
except KeyError:
query = u''
try:
id = request.REQUEST[u'id']
if id:
template = get_template(u'profile_internal.html')
entity = models.Entity.objects.filter(id=int(id))[0]
if entity.gps:
gps = entity.gps
elif entity.location and entity.location.gps:
gps = entity.location.gps
else:
gps = u''
if gps:
gps_url = u'http://maps.google.com/maps?f=q&hl=en&q=' + gps.replace(u' ', u'+') + "&iwloc=A&hl=en"
else:
gps_url = u''
profile = template.render(RequestContext(request, {
u'entities': models.Entity.objects.filter(is_invisible=False),
u'entity': entity,
u'first_statuses': models.Status.objects.filter(entity=id).order_by(u'-timestamp')[:settings.INITIAL_STATUSES],
u'gps': gps,
u'gps_url': gps_url,
u'id': int(id),
u'emails': models.Email.objects.filter(entity=entity, is_invisible=False),
u'phones': models.Phone.objects.filter(entity=entity, is_invisible=False),
u'query': urllib.quote(query),
u'second_statuses': models.Status.objects.filter(entity=id).order_by(u'-timestamp')[settings.INITIAL_STATUSES:],
u'time_zones': models.TIME_ZONE_CHOICES,
u'tags': models.Tag.objects.filter(entity=entity, is_invisible=False).order_by(u'name'),
u'urls': models.URL.objects.filter(entity=entity, is_invisible=False),
}))
except KeyError:
profile = u''
try:
if query:
template = get_template(u'search_internal.html')
tokens = re.split(ur'(?u)[^-\w]', query)
while u'' in tokens:
tokens.remove(u'')
candidates = []
for candidate in models.Entity.objects.all():
candidates.append([candidate, 0])
for token in tokens:
new_candidates = []
for candidate in candidates:
if score(candidate[0], token) > 0:
candidate[1] += score(candidate[0],token)
new_candidates.append(candidate)
candidates = new_candidates
candidates.sort(lambda a, b: -cmp(a[1], b[1]))
export = []
for candidate in candidates:
if candidate[0].image_mimetype:
image = True
else:
image = False
export.append({
u'department': candidate[0].department,
u'description': candidate[0].description,
u'id': candidate[0].id,
u'image': image,
u'name': candidate[0].name,
u'title': candidate[0].title,
})
first_portion = export[:settings.INITIAL_RESULTS]
second_portion = export[settings.INITIAL_RESULTS:]
template = get_template(u'search_internal.html')
search_results = template.render(RequestContext(request, {
u'first_portion': first_portion,
u'query': query,
u'second_portion': second_portion,
}))
else:
search_results = u''
except KeyError:
search_results = u''
return render_to_response(u'search.html', RequestContext(request, {
u'profile': profile,
u'query': urllib.quote(query),
u'search_results': search_results,
u'settings': settings,
}))
# main loop
run = True
while run:
clock.tick(30) # 30 fps
for event in pygame.event.get(): # checks if the player closed the screen
if event.type == pygame.QUIT:
run = False
keys = pygame.key.get_pressed() # read the keyboard inputs
# left player commands
if keys[pygame.K_w] and left_P.y != 0:
left_P.y -= left_P.vel
if keys[pygame.K_s] and left_P.y != (WIND_HEIGHT - left_P.height):
left_P.y += left_P.vel
# right player commands
if keys[pygame.K_UP] and right_P.y != 0:
right_P.y -= right_P.vel
if keys[pygame.K_DOWN] and right_P.y != (WIND_HEIGHT - right_P.height):
right_P.y += right_P.vel
f.movement(ball, left_P, right_P) # checks the ball movement
if f.score(ball, left_P, right_P): # checks score
ball = cb.ball(10)
drawScreen() # draw screen after inputs
pygame.quit() # finish the game
def find_best_p(protein_object, rep_num, best_score): result_array = [] start = timer() result_array.append(['start', start]) high_protein_object = copy.deepcopy(protein_object) # arrays with h, p and c amino acid chain numbers h_loc = [] p_loc = [] c_loc = [] # gets coordinates array coor_array = get_coor_array(protein_object) # fills H and P arrays for i in range(len(protein_object)): if protein_object[i].ac_type == 'H': h_loc.append(i) elif protein_object[i].ac_type == 'P': p_loc.append(i) elif protein_object[i].ac_type == 'C': c_loc.append(i) count = 0 rep_count = 0 # folds random P amino acid rep_num times # rep_count checks if protein changes and returns # best protein if protein stays the same after 30 folds while count < rep_num and rep_count < 30: start_p = timer() - start random_bend = 0 # gets random amino num form p array (not amino num 0 and 1) while random_bend == 0 or random_bend == 1: random_bend = random.choice(p_loc) # bool to chose between two folds or one rand_bool = random.randint(0, 1) # if two P's are next to each other the both get folded to form a U turn if random_bend + 1 in p_loc and rand_bool == 1: # folds amino num, amino num + 1 for i in range(2): fold_coor = folder_protein(coor_array, random_bend) protein_object_new = copy.deepcopy(update_objects(protein_object, fold_coor)) grid = functions.protein_visual(protein_object_new) score = functions.score(protein_object_new, grid) if fold_coor == coor_array: rep_count += 1 else: coor_array = copy.deepcopy(fold_coor) count += 1 random_bend += 1 # checks if current score is new high score and updates score + object if score < best_score: best_score = copy.copy(score) best_protein_object = copy.deepcopy(protein_object_new) # folds random amino num else: fold_coor = folder_protein(coor_array, random_bend) protein_object_new = copy.deepcopy(update_objects(protein_object, fold_coor)) grid = functions.protein_visual(protein_object_new) score = functions.score(protein_object_new, grid) protein_object = copy.deepcopy(protein_object_new) # only updates coor array if it differs from previous coor array if fold_coor == coor_array: rep_count += 1 else: coor_array = copy.deepcopy(fold_coor) count += 1 if score < best_score: best_score = copy.copy(score) best_protein_object = copy.deepcopy(protein_object_new) # adds end time and score after every fold set completion end_p = timer() - start result_array.append(['get best P', start_p, end_p, count + rep_count, best_score]) end = timer() result_array.append(['end', end]) return [best_score, best_protein_object]
def home():
# if request.method == 'POST':
# user = request.form['nm']
# return redirect(url_for('success',name = user))
# else:
# user = request.args.get('nm')
# return redirect(url_for('success',name = user))
# LOAD DATA
location = 'data/dataobat.xml'
documentNumber = func.docNumber(location)
documentHeadline = func.docHeadline(location)
documentText = func.docText(location)
documentTotal = len(documentNumber)
konten = []
for i in range(documentTotal):
konten.append(documentHeadline[i] + documentText[i])
# PREPROCESSING
text = func.removePunctuation(konten)
text = func.caseFolding(text)
text = func.tokenize(text)
text = func.stopwordRemove(text)
text = func.numberRemove(text)
text = func.stemming(text)
# GET ALL TERMS IN COLLECTION
terms = func.getAllTerms(text)
# INDEXING
# index = createIndex(text,documentNumber, terms)
index = func.createIndex(text, documentNumber)
# tambahan untuk query yang belum di ekspan
raw_query = ["100 mg of dacarbazine, "]
# raw_query = newQuery
query = func.removePunctuation(raw_query)
query = func.caseFolding(query)
query = func.tokenize(query)
query = func.stopwordRemove(query)
query = func.numberRemove(query)
query = func.stemming(query)
query = query[0]
# Check Query In Index
query = func.queryInIndex(query, index)
# RANKED RETRIEVAL
N = documentTotal
tfidf_list = []
docFrequency = func.df(query, index)
invDocFrequency = func.idf(docFrequency, N)
termFrequency = func.tf(query, index)
TFIDF = func.tfidf(termFrequency, invDocFrequency)
sc = func.score(TFIDF)
relevanceDocNumber = []
count = 0
print('Query: ', raw_query, '\n\n')
print('RESULTS: \n')
for i in range(len(sc)):
relevanceDocNumber.append(int(sc[i]))
a = documentNumber.index(sc[i])
print(documentHeadline[a] + 'Document Number: ', sc[i])
# print('\nContent:')
# print(documentText[a][0:400], '[read more]>>')
print('-------------------------------------------\n')
count = count + 1
if (count >= 5):
break
# akhir tambahan
return render_template('home.html',
title='Load_data',
raw_query=raw_query,
dh=documentHeadline)
