def contact_handler(bot, update):
upd = update.to_dict()
phone_no = str('+') + upd.get('message').get('contact').get('phone_number')
content['contact'] = phone_no
if check_skin('./temp/' + str(update.message.chat_id) + '.jpg'):
preds_dict = predict_class('./temp/' + str(update.message.chat_id) + '.jpg')
dict_dis = sorted(preds_dict.items(), key=lambda x: x[1], reverse=True)
print(dict_dis)
dict_dis = dict(sorted(preds_dict.items(), key=lambda x: x[1], reverse=True)[:3])
max_val = max(dict_dis, key=dict_dis.get)
if dict_dis[max_val] <= 39:
bot.send_message(chat_id=update.message.chat_id, text='Healthy Skin Detected')
else:
plt.bar(range(len(dict_dis)), list(dict_dis.values()), align='center')
x1,x2,y1,y2 = plt.axis()
plt.axis((x1,x2,0,100))
plt.xticks(range(len(dict_dis)), range(len(dict_dis)))
plt.savefig('./temp/plots/tempfig.png')
bot.send_message(chat_id=update.message.chat_id, text="Thank you!\n \nYou will recieve your report Soon")
bot.send_photo(chat_id=update.message.chat_id, photo=open('./temp/plots/tempfig.png', 'rb'))
os.remove('./temp/plots/tempfig.png')
plt.cla()
keys = list(dict_dis.keys())
message_text = '0 ->' + str(keys[0]) + '\n1 ->' + str(keys[1]) + '\n2 ->' + str(keys[2])
bot.send_message(chat_id=update.message.chat_id, text=message_text)
else:
bot.send_message(chat_id=update.message.chat_id, text='Please Upload Infected Skin Area for Diagnosis')
def predictor(img_url, filename):
model = model_load(filename)
img = get_image(img_url)
labels, names = get_labels(filename)
predicted_id, predicted_class, confidence = predict_class(
img, model, labels, names)
return predicted_id, predicted_class, confidence
def home():
if request.method == 'POST':
img = request.form['canvasData']
predicted_class = predict_class(img)
return render_template('index.html', predicted_class=predicted_class)
else:
return render_template('index.html')
def get_target_image_url():
"""
@param url: post : The target image url.
@returns: 201 : a label (cat / dog) as a flask/response object.
@raise 400: if unable to download the image.
"""
if not request.get_json():
abort(400)
else:
data = request.get_json(force=True)
if not data.get('image_url'):
abort(400)
else:
url = data.get('image_url')
r = requests.get(url)
image_dir = os.path.join('/media', str(url).split('/')[-1])
if not os.path.exists('/media'):
os.mkdir('/media')
if not os.path.exists(image_dir):
with open(image_dir, 'wb') as f:
f.write(r.content)
status = 200
else:
status = 201
class_name = predict_class(os.path.join("..", image_dir))
return jsonify({'class': class_name,
'status': status})
def main(argv):
try:
opts, args = getopt.getopt(argv, "htp:", ["train", "predict="])
except getopt.GetoptError:
logging.info(
'Incorrect argument... \nmain.py -t or --train to train the model || main.py -p or --predict <Text> to predict on model output'
)
sys.exit(2)
for opt, arg in opts:
if opt == '-h':
logging.info(
'main.py -t or --train to train the model || main.py -p or --predict <Text> to predict on model output'
)
sys.exit()
elif opt in ("-t", "--train"):
logging.info("training model started...")
get_data = LoadTrainingData(CONFIG_PATH)
df = get_data.load_data_from_tables()
logging.info(f"training data loaded... {df.shape}")
train_model(df)
logging.info("training model finished!")
elif opt in ("-p", "--predict"):
text = arg
if len(text) > 10:
main_prouct = predict_class(text)
logging.info("Prediction done!")
else:
logging.info("too small sentence to predict")
def predict():
data = request.get_json()
data = json.dumps(data)
pred, pred_proba = predict_class(model, lecancer, lerelationships, history_class_codes, ethnicity_codes, data, df)
print(pred, pred_proba[0].tolist())
return_obj = {
'pred': pred,
'predProba': pred_proba[0].tolist()
}
return json.dumps(return_obj)
def classify_audio():
#sends username and youtube url
json_data = request.json
url = json_data['url']
username = json_data['username']
label = predict_class(url)
if label == None:
return json.dumps('error')
emot = label_to_emot[label]
addUrlToDb(username, emot, url)
return json.dumps(emot)
def process_img(img):
# time.sleep(2)
preds = predict.predict_class(img)
with lock:
global last_object_class
global last_object_count
global last_object_ts
global detected_object_class
last_object_ts = time.time()
detected_object_class = ""
if last_object_class == preds and (last_object_ts - time.time()) < 1:
last_object_count += 1
if last_object_count == last_object_target_count:
detected_object_class = last_object_class
last_object_count = 0
print(detected_object_class)
else:
last_object_class = copy.deepcopy(preds)
last_object_count = 0
return True
from glob import glob
import cv2
import numpy as np
# from skin_detection import get_hist, check_skin
# img_list = sorted(glob('./imgs/*.jpg'))
# for img in img_list:
# hist = get_hist(img)
# check = check_skin(hist)
# if check:
# print(img, ':Skin')
# else:
# print(img, ':Not Skin')
from predict import predict_class, classes
imgs_list = ['03DermatitisArm.jpg']
# '03eczema091205.jpg',
# '04acidBurn89-GP3.jpg',
# 'chapped-fissured-feet-34.jpg']
for img in imgs_list:
dict_dis = predict_class('./imgs/' + img)
dict_dis = sorted(dict_dis.items(), key=lambda x: x[1], reverse=True)
print(dict_dis[:3])
def smartatm(phrase):
result = predict.predict_class(phrase)
apilog.info("Predict Request : " + phrase)
apilog.info("Predicted as : " + result)
return result