cydal

pd.DataFrame({
    "t_2": df.head(11)["energy"].shift(2), 
    "t_1": df.head(11)["energy"].shift(), 
    "energy": df.head(11)["energy"]
})

cydal

df.head(11).diff()

cydal

from tsextract.feature_extraction.extract import build_features, build_features_forecast
from tsextract.domain.statistics import mean, median, std

features_request = {
    "window": [48],
    "window_statistic": [48, mean], 
    "difference_statistic": [48, 1, std], 
}

build_df = build_features(df.energy, features_request, target_lag=48, include_tzero=True)

cydal

scaler_features = StandardScaler().fit(build_df[build_df.columns.values[:-1]])
scaled_features = scaler_features.transform(build_df[build_df.columns.values[:-1]])

scaler_label = StandardScaler().fit(np.array(build_df[build_df.columns.values[-1]]).reshape(-1, 1))
scaled_label = scaler_label.transform(np.array(build_df[build_df.columns.values[-1]]).reshape(-1, 1))


### Split data using train proportion of 0.7
train_size = int(scaled_features[:, :-1].shape[0] * 0.7)

cydal

from keras.models import Sequential
from keras.layers import Dense, Dropout


model = Sequential()
model.add(Dense(128, input_dim=X_train.shape[1], kernel_initializer='normal', activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(64, kernel_initializer='normal', activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(1, kernel_initializer='normal'))

cydal

! pip install --upgrade git+https://github.com/SheffieldSolar/PV_Live-API

from pvlive_api import PVLive
from datetime import datetime
import pytz

pvl = PVLive()

years = [2014, 2015, 2016, 2017, 2018, 2019]

cydal

## Build Features for forecasting

build_forecast_df = build_features_forecast(df.energy, features_request, include_tzero=True)
scaled_features_forecast = scaler_features.transform(build_forecast_df.tail(48))
pred = model.predict(scaled_features_forecast[:, :-1])
pred = scaler_label.inverse_transform(pred)

# Range for next 24 hours
range = pd.date_range(start=build_df.index[-1]+timedelta(minutes=30), 
              end=build_df.index[-1]+timedelta(hours=24), 

cydal

from tsextract.plots.eval import actualPred, get_lag_corr, scatter

actualPred(y_test.reshape(-1, ), test_pred.reshape(-1))
scatter(y_test.reshape(-1, ), test_pred.reshape(-1))
get_lag_corr(y_test.reshape(-1, ), test_pred.reshape(-1), num_lags=20)

cydal

from PIL import Image
import numpy as np
from scipy import ndimage
from skimage import filters
from skimage.filters import laplace, sobel_h, sobel_v
from skimage.feature import canny
from skimage.color import rgb2hsv

img = np.array(Image.open(img_path))
img_hsv = rgb2hsv(img)

cydal

from plantcv import plantcv as pcv
from PIL import Image
import numpy as np
import cv2

img = cv2.imread(img_path)

a = pcv.rgb2gray_lab(img, channel='a')
img_binary = pcv.threshold.binary(gray_img=a, threshold=120, max_value=255, object_type='dark')
greenPlants = pcv.fill(bin_img=img_binary, size=200)