tgraph.py
#!/usr/bin/env python3
"""
docstring
"""
from datetime import datetime, timedelta
from math import pi
import pandas as pd
# bokeh libraries
from bokeh.plotting import figure
from bokeh.io import export_png
from bokeh.layouts import layout
COLOR_THEME = {
"normal":
{"up": "#98FB98", "down": "#FF0000", "sell_order": "#FF0000", "buy_order": "#98FB98"},
"colorblind":
{"up": "#00ee00", "down": "#ff00ff", "sell_order": "#ff00ff", "buy_order": "#00ee00"},
"monochrome":
{"up": "white", "down": "black", "sell_order": "black", "buy_order": "black"}
}
class Tgraph:
def __init__(self, candles_data, active_orders, orders_data, symbol, **kwargs):
self.colors = self.set_colors(**kwargs)
self.symbol = symbol
self.candles_data = candles_data
self.active_orders = active_orders
self.orders_data = orders_data
candles_df = self.build_dataframe()
self.candle_width = 30 * 60 * 1000 # half an hour in ms
self.x_min = candles_df['date'].min() - timedelta(hours=1)
self.x_max = candles_df['date'].max() + timedelta(hours=1)
cdl_graph = self.build_candles_graph(candles_df)
ao_graph = self.build_active_orders_graph()
vol_graph = self.build_volume_graph(candles_df)
rsi_graph = self.build_rsi_graph(candles_df)
self.graphs_layout = layout(
children=[
[cdl_graph, ao_graph],
[vol_graph],
[rsi_graph]
]
)
def build_dataframe(self):
candles_df = pd.DataFrame(
self.candles_data,
columns=['date', 'open', 'close', 'high', 'low', 'volume']
)
def get_date(unixtime):
time_stamp = unixtime / 1000
formated_date = datetime.utcfromtimestamp(time_stamp).strftime('%Y-%m-%d %H:%M:%S')
return formated_date
candles_df['date'] = candles_df['date'].apply(get_date)
candles_df['volume'] = candles_df['volume'].astype(int)
candles_df["date"] = pd.to_datetime(candles_df["date"])
# Window length for moving average
window_length = 14
candles_df['seq'] = candles_df['date']
close = candles_df['close'][::-1]
# Get the difference in price from previous step
delta = close.diff()
# Make the positive gains (up) and negative gains (down) Series
up_gains, down_gains = delta.copy(), delta.copy()
up_gains[up_gains < 0] = 0
down_gains[down_gains > 0] = 0
# Calculate the EWMA
roll_up1 = up_gains.ewm(
com=window_length,
min_periods=0,
adjust=True,
ignore_na=False
).mean()
roll_down1 = down_gains.abs().ewm(
com=window_length,
min_periods=0,
adjust=True,
ignore_na=False
).mean()
# Calculate the RSI based on EWMA
rs1 = roll_up1 / roll_down1
rsi1 = 100.0 - (100.0 / (1.0 + rs1))
candles_df['rsi_ewma'] = rsi1
candles_df = candles_df[::-1][16:]
return candles_df
def build_candles_graph(self, candles_df):
x_text = candles_df['date'].min()
candles_graph = figure(
x_axis_type="datetime",
toolbar_location=None,
x_range=(self.x_min, self.x_max),
plot_width=1000,
title=self.symbol)
candles_graph.title.text_font_size = "25pt"
candles_graph.yaxis.major_label_text_font_size = "15pt"
candles_graph.yaxis[0].ticker.desired_num_ticks = 20
candles_graph.xaxis.major_label_text_font_size = "12pt"
candles_graph.xaxis[0].ticker.desired_num_ticks = 30
candles_graph.xaxis.major_label_orientation = pi / 2
candles_graph.grid.grid_line_alpha = 0.3
candles_graph.segment(
candles_df.date,
candles_df.high,
candles_df.date,
candles_df.low,
color="black"
)
inc = candles_df.close > candles_df.open
candles_graph.vbar(
candles_df.date[inc],
self.candle_width,
candles_df.open[inc],
candles_df.close[inc],
fill_color=self.colors['up'],
line_color="black"
)
dec = candles_df.open > candles_df.close
candles_graph.vbar(
candles_df.date[dec],
self.candle_width,
candles_df.open[dec],
candles_df.close[dec],
fill_color=self.colors['down'],
line_color="black"
)
sells = []
buys = []
max_price = float(candles_df['high'].max())
min_price = float(candles_df['low'].min())
for order in self.active_orders:
price = float(order['price'])
if price > max_price or price < min_price:
continue
line = [price] * candles_df.shape[0]
if order['side'] == 'sell':
sign = "-"
sells.append(line)
candles_graph.rect(
x=x_text + timedelta(hours=3),
y=price + 0.002,
width=timedelta(hours=7),
height=0.005,
fill_color="white"
)
else:
buys.append(line)
sign = "+"
candles_graph.rect(
x=x_text + timedelta(hours=3),
y=price + 0.002,
width=timedelta(hours=7),
height=0.005,
fill_color="white"
)
candles_graph.text(
x=x_text,
y=price,
text=[f"{sign}{order['remaining_amount']}"],
text_font_size='8pt',
text_font_style='bold'
)
# add current price
current_price = float(candles_df['close'][0])
new_line = [current_price] * candles_df.shape[0]
if candles_df['open'][0] > candles_df['close'][0]:
sells.append(new_line)
candles_graph.rect(
x=x_text + timedelta(hours=3),
y=current_price + 0.002,
width=timedelta(hours=7),
height=0.005,
fill_color="white"
)
else:
buys.append(new_line)
candles_graph.rect(
x=x_text + timedelta(hours=3),
y=current_price + 0.002,
width=timedelta(hours=7),
height=0.005,
fill_color="white"
)
candles_graph.text(
x=x_text,
y=current_price,
text=[f"{current_price}"],
text_font_size='8pt',
text_font_style='bold'
)
# add sell lines
candles_graph.multi_line(
xs=[candles_df.seq] * len(sells),
ys=sells,
line_color=self.colors['sell_order'],
line_dash="dashed",
line_width=1
)
# add buy lines
candles_graph.multi_line(
xs=[candles_df.seq] * len(buys),
ys=buys,
line_color=self.colors['buy_order'],
line_width=1
)
return candles_graph
def build_rsi_graph(self, candles_df):
rsi_graph = figure(
plot_width=1000,
plot_height=100,
x_range=(self.x_min, self.x_max),
toolbar_location=None,
y_range=(0, 100),
title="Relative Strength Index"
)
rsi_graph.xaxis.visible = False
rsi_graph.multi_line(
xs=[candles_df.seq] * 4,
ys=[
candles_df.rsi_ewma,
[20] * candles_df.shape[0],
[50] * candles_df.shape[0],
[80] * candles_df.shape[0]
],
line_color=['red', 'black', 'gray', 'black'],
line_width=1
)
rsi_graph.yaxis.major_label_text_font_size = "11pt"
return rsi_graph
def build_volume_graph(self, candles_df):
# Historic Volume GRAPH
volume_max = int(candles_df['volume'].max())
volume_graph = figure(
plot_width=1000,
plot_height=100,
x_range=(self.x_min, self.x_max),
toolbar_location=None,
y_range=(0, volume_max),
title="Volume"
)
volume_graph.xaxis.visible = False
volume_graph.left[0].formatter.use_scientific = False
volume_graph.vbar(
candles_df.date,
self.candle_width,
candles_df.volume * 0,
candles_df.volume,
fill_color="blue",
line_color="black"
)
return volume_graph
def build_active_orders_graph(self):
# Volume in active orders GRAPH
orderbook_df = pd.DataFrame.from_dict(self.orders_data, orient='columns')
orders_vol_graph = figure(
plot_width=200,
toolbar_location=None,
y_range=orderbook_df.price,
title="Orderbook"
)
orders_vol_graph.below[0].formatter.use_scientific = False
orders_vol_graph.yaxis.visible = False
orders_vol_graph.xaxis.major_label_text_font_size = "8pt"
orders_vol_graph.xaxis[0].ticker.desired_num_ticks = 5
orders_vol_graph.xaxis.major_label_orientation = pi / 2
orders_vol_graph.hbar(
y=orderbook_df.price,
left=orderbook_df.amount * 0,
right=orderbook_df.amount,
height=0.1
)
return orders_vol_graph
def set_colors(self, **kwargs):
colors_dict = {}
colors_dict['up'] = COLOR_THEME['normal']['up']
colors_dict['down'] = COLOR_THEME['normal']['down']
colors_dict['sell_order'] = COLOR_THEME['normal']['sell_order']
colors_dict['buy_order'] = COLOR_THEME['normal']['buy_order']
for key, value in kwargs.items():
if key == 'graphtheme':
colors_dict['up'] = COLOR_THEME[value]['up']
colors_dict['down'] = COLOR_THEME[value]['down']
colors_dict['sell_order'] = COLOR_THEME[value]['sell_order']
colors_dict['buy_order'] = COLOR_THEME[value]['buy_order']
return colors_dict
def save_picture(self):
# export the graph
export_png(self.graphs_layout, filename="graph.png")
