html.Div([
html.P([html.Strong('Test for dialog')]),
sd_material_ui.Dialog([
html.H3('Sample Dialog'),
html.Div(html.Button('Close Dialog'), id='closer')
], id='output2'),
html.Div(id='input2', children=[
html.Button(children='Open the dialog')
]),
]),
spacer,
html.Div([
html.P([html.Strong('Sample FontIcon')]),
sd_material_ui.FontIcon(id='fonticon', iconName='insert_emoticon'),
]),
spacer,
html.Div([
html.P([html.Strong('Test for toggle switch')]),
sd_material_ui.Toggle(
id='toggle-input',
label='Johnny?',
toggled=False,
),
html.P(id='toggle-output', children=['Flame off']),
]),
spacer,
zDepth=5,
circle=True,
style=dict(
height=100,
width=100,
margin=20,
textAlign='center',
display='inline-block',
),
),
spacer,
sd_material_ui.Paper(
children=[
sd_material_ui.FontIcon(
className='material-icons',
iconName='help',
color='green',
)
],
zDepth=4,
rounded=False,
style=dict(
height=100,
width=100,
margin=20,
textAlign='center',
display='inline-block',
),
),
spacer,
def predict_stocks_decision_daily(json_data, figure):
"""
Callback to get stock data from the hidden div and perform RL to predict the optimal decision to make.
"""
# Parse
json_res = json.loads(json_data)
series = json_res["Time Series (Daily)"]
days = [day for day in series]
closing = [float(series[day]["4. close"]) for day in days]
# Get the last 100 days stock prices
model_input = closing[:100]
model_input = [float(numeric_string) for numeric_string in model_input]
model_input = np.reshape(model_input, (1, 1, 100))
# Load model
json_file = open('models/dqn_model.json', 'r')
loaded_model_json = json_file.read()
json_file.close()
loaded_model = model_from_json(loaded_model_json)
# load weights into new model
loaded_model.load_weights("models/dqn_weights.h5")
# Run model
decisions = loaded_model.predict(model_input)
# Reset for future model predictions
K.clear_session()
# Render output
if np.argmax(decisions) == 0: # Stay
return sd_material_ui.Paper(
children=[
sd_material_ui.FontIcon(className='material-icons',
iconName='trending_flat',
hoverColor="#1125ff"),
html.P('The Deep Q Network advices to ignore this stock')
],
zDepth=2,
rounded=True,
style=dict(height=150,
width=150,
margin=20,
textAlign='center',
display='inline-block'),
),
elif (np.argmax(decisions) == 1): # Buy
return sd_material_ui.Paper(
children=[
sd_material_ui.FontIcon(className='material-icons',
iconName='trending_up',
hoverColor="#56fc0a"),
html.P('The Deep Q Network advices to invest in this stock')
],
zDepth=2,
rounded=True,
style=dict(height=150,
width=150,
margin=20,
textAlign='center',
display='inline-block'),
),
else: # Sell
return sd_material_ui.Paper(
children=[
sd_material_ui.FontIcon(className='material-icons',
iconName='trending_down',
hoverColor="#fc000c"),
html.P('The Deep Q Network advices to sell this stock')
],
zDepth=2,
rounded=True,
style=dict(height=150,
width=150,
margin=20,
textAlign='center',
display='inline-block'),
),