def findAcc(path):
img = cv.imread(getFilePath(ORIGINAL, path))
low_green = np.array([0, 80, 0])
high_green = np.array([150, 255, 100])
hsv_img = cv.cvtColor(img, cv.COLOR_BGR2HSV)
mask = cv.inRange(hsv_img, low_green, high_green)
res = cv.bitwise_and(img, img, mask=mask)
# cv.imwrite(nam + " Mask.png", mask)
bw = cv.cvtColor(res, cv.COLOR_BGR2GRAY)
bw = cv.GaussianBlur(bw, (3, 3), 0)
a, thresh = cv.threshold(bw, 0, 255, cv.THRESH_BINARY + cv.THRESH_OTSU)
kernel = np.ones((5, 5), np.uint8)
canny = cv.Canny(bw, a * 0.5, a)
# cv.imwrite(nam + " Canny.png", canny)
canny = cv.dilate(canny, kernel, iterations=1)
# cv.imwrite(nam + " Dilated.png", canny)
# cv.imshow("ss", canny)
area, tot = cv.countNonZero(canny), img.shape[0] * img.shape[1]
date = datetime.datetime.now()
cv.putText(img, 'Forest Cover ' + str(area / tot * 100), (56, 80),
cv.FONT_HERSHEY_COMPLEX, 0.7, (2, 255, 2), 2)
cv.putText(img,
str(date.day) + '/' + str(date.month) + '/' + str(date.year),
(600, 500), cv.FONT_HERSHEY_COMPLEX, 0.7, (2, 255, 2), 2)
cv.imwrite(getFilePath(MODIFIED, path), img)
return area / tot
def home():
imagePaths = []
for imgdir in os.scandir(baseDirectory):
imagePathObject = {
'original': getFilePath(ORIGINAL, imgdir),
'modified': getFilePath(MODIFIED, imgdir),
'title': open(getFilePath(DATA, imgdir), 'r').read()
}
imagePaths.append(imagePathObject)
# TO-DO: Sort imagePaths w.r.t last modifed
return render_template('index.html', imagePaths=imagePaths)
def on_epoch_end(self, epoch, logs={}):
# Append the logs, losses and accuracies to the lists
self.logs.append(logs)
self.losses.append(logs.get('loss'))
self.acc.append(logs.get('acc'))
self.val_losses.append(logs.get('val_loss'))
self.val_acc.append(logs.get('val_acc'))
# Before plotting ensure at least 2 epochs have passed
if len(self.losses) > 1:
N = np.arange(0, len(self.losses))
plt.figure()
plt.plot(N, self.losses, label="train_loss")
plt.plot(N, self.acc, label="train_acc")
plt.plot(N, self.val_losses, label="val_loss")
plt.plot(N, self.val_acc, label="val_acc")
plt.title("Training Loss and Accuracy [Epoch {}]".format(epoch))
plt.xlabel("Epoch #")
plt.ylabel("Loss/Accuracy")
plt.legend()
plt.savefig(utils.getFilePath(TRAINING_IMG,'Epoch-{}.png'\
.format(epoch)))
plt.close()
def startPipeline(self, pickled_filename):
self.__preprocess()
self.__embbed_dataset(
ut.getFilePath(['pickled', 'word2vec'],
pickled_filename,
create=True))
self.__merge()
def train_model(getSantizeStanceData):
x_train, y_train, x_test, y_test = getSantizeStanceData.getTrainTestData()
model = baseline_model(x_train)()
print("[==] Model Summary")
print(model.summary())
print("[+] Training model")
estimator = model.fit(x_train,
y_train,
validation_data=(x_test, y_test),
batch_size=batch_size,
callbacks=[plot_losses],
epochs=nb_epoch,
verbose=2)
print("[==] Saving model")
model.save(utils.getFilePath('pickled', visual.MODEL_NAME))
print("--- %s seconds ---" % (time.time() - start_time))
print("[+] Training model")
estimator = model.fit(x_train,
y_train,
validation_data=(x_test, y_test),
batch_size=batch_size,
callbacks=[plot_losses],
epochs=nb_epoch,
verbose=2)
print("[==] Saving model")
model.save(utils.getFilePath('pickled', visual.MODEL_NAME))
print("--- %s seconds ---" % (time.time() - start_time))
if __name__ == "__main__":
#supply this to generate all the required dataset files
#embedding type specifies the vector formation whether tokenizer or tf_idf
pipline = pre.StanceDataPipeline(embedding_type='tokenizer')
#specify the name of of the pickle file
pipline.startPipeline(pickled_filename="word2vec_filename.pkl")
getStanceData = pre.GetStanceData(path_to_vec_pkl=utils.getFilePath(
['pickled', 'word2vec'], filename='word2vec_filename.pkl'),
embedding_type='tokenizer')
getStanceData.getTrainTestData(shuffle=True)
train_model(getSantizeStanceData=getStanceData)
import matplotlib
matplotlib.use('agg')
import matplotlib.pyplot as plt
import keras
import numpy as np
import utils
MODEL_NAME = "LSTM_CNN"
FOLDER_NAME = "training"
TRAINING_IMG = utils.getFilePath(foldername=[FOLDER_NAME, MODEL_NAME],
create=True)
class TrainingPlot(keras.callbacks.Callback):
# This function is called when the training begins
def on_train_begin(self, logs={}):
# Initialize the lists for holding the logs, losses and accuracies
self.losses = []
self.acc = []
self.val_losses = []
self.val_acc = []
self.logs = []
# This function is called at the end of each epoch
def on_epoch_end(self, epoch, logs={}):
# Append the logs, losses and accuracies to the lists
self.logs.append(logs)
from mwapi import MWapi
from sc import SC
from utils import getFilePath
getFilePath(__file__)
class VOC:
def __init__(self, temp_dir, temp_url):
self.sc = SC()
self.mwapi = MWapi(temp_dir, temp_url)
def lookup(self, word):
res = {"sc": None, "api": None}
iscorrect, res_sc = self.sc.check(word)
if not iscorrect:
# res["candi"] = res_sc
res = {"sc": res_sc}
return res
res_api = self.mwapi.lookup(word, mode="json")
res["voc"] = res_api
return res
import matplotlib
matplotlib.use('agg')
import matplotlib.pyplot as plt
import keras
import numpy as np
import utils
MODEL_NAME = "LSTM_CNN"
FOLDER_NAME = "training"
<<<<<<< HEAD
TRAINING_IMG = utils.getFilePath(
foldername=[FOLDER_NAME,MODEL_NAME],
create=True
)
=======
MODEL_NAME = "default"
BASE_DIR = os.path.dirname(os.path.abspath(__file__))
TRAINING_IMG = os.path.join(BASE_DIR,'training')
TRAINING_IMG = os.path.join(TRAINING_IMG,MODEL_NAME)
>>>>>>> bf2b4ef115818a186ebc39a0b859ee08b2b49175
class TrainingPlot(keras.callbacks.Callback):
# This function is called when the training begins
def on_train_begin(self, logs={}):
# Initialize the lists for holding the logs, losses and accuracies
self.losses = []
self.acc = []
self.val_losses = []
import utils
BODY_KEY = "body"
STANCE_KEY = "stance"
HEADLINE_KEY = 'headline'
DATASET_FOLDER = utils.getFilePath('dataset',create=True)
PICKLED_FOLDER = 'pickled'
TRAIN_DATASET_FILES = {
BODY_KEY: utils.getFilePath('dataset','train_bodies.csv'),
STANCE_KEY:utils.getFilePath('dataset','train_stances.csv')
}
utils.checkRequiredFiles(TRAIN_DATASET_FILES)
TRAIN_DATASET_NUMPY = {
STANCE_KEY:utils.getFilePath(PICKLED_FOLDER,'train_stance_scipy.npy'),
BODY_KEY:utils.getFilePath(PICKLED_FOLDER,'train_body_scipy.npy'),
HEADLINE_KEY:utils.getFilePath(PICKLED_FOLDER,'train_headline_scipy.npy'),
}
TEST_DATASET_FILES = {
BODY_KEY: utils.getFilePath('dataset','test_stances.csv'),
STANCE_KEY:utils.getFilePath('dataset','competition_test_stances_label.csv'),
}
utils.checkRequiredFiles(TEST_DATASET_FILES)
TEST_DATASET_NUMPY = {
STANCE_KEY:utils.getFilePath(PICKLED_FOLDER,'test_stance_scipy.npy'),