Market Data Form Finder Toolkit (MDFFT). The project goal is to make working with a raw market data easier.
The project goal is to make working with a raw market data easier. Provide a tool that allow to identify recurrent market patterns. Any types of market can processed by this.
The library allow to carry out following tasks:
pip install mdfft
The namespace mdfft contains of whole list of lib's classes. The example of import is below:
from mdfft import Candles, SimpleParser, Styler, painter as p, RawCandle, Trader
Classes:
The collection of candles is created by parsing quotes or making an array of RawCandle objects. The array of RawCandles is created by the programmer himself.
raw_candles = [
RawCandle(
dt=datetime(2020, 1, 1), # Candle datetime
o=2, # Open price
h=4, # High price
l=1, # Low price
c=3 # Close price
),
RawCandle(
dt=datetime(2020, 1, 2), # dt
o=12, # ohlc prices
h=14,
l=11,
c=13
)
]
candles = Candles.from_raw_candles(raw_candles, timeframe="1D")
print(candles)
Next I will use my SimpleParser and my rather large data set of OHLC quotes.
parser = SimpleParser(file="./data/EURUSD_H1.txt")
candles = Candles.from_raw_candles(parser.parse(), timeframe="1H")
print(candles.candles_count())
print(len(candles))
start_candle = candles.head()
print("Start candle: " + str(start_candle))
end_candle = candles.tail()
print("End candle: " + str(end_candle))
my_candle = candles.candle_at(1985)
same_candle = candles[1985]
Navigation is performed by calling next() and prev() methods of a candle object.
next_candle = my_candle.next()
prev_candle = my_candle.prev()
next20_candle = my_candle.next(20)
prev30_candle = my_candle.prev(30)
found_candle = candles.find_candle(datetime(1999, 1, 4, 5))
if not found_candle:
print("Candle not found at 1999-01-04 05:00")
return print("Found candle: " + str(found_candle))
# take 10 skip 5 candles
print(candles.take(10, skip=5))
# or print(candles[5:15])
# skip 5 candles from collection
print(candles.skip(5))
# or print(candles[5:])
from datetime import datetime
# New candles collection from 2000-01-01 to 2000-01-10
# Note Only if candles are exists
candles_range = candles.range_dt( datetime(2000, 1, 3, 1 ), datetime(2000, 1, 10, 1))
if not candles_range:
print("No candles in range")
return 0
print("Count candles: " + str(candles_range.candles_count()))
print(candles_range)
# Iterate candles
for candle in candles:
print("Candle: " + str(candle))
# Get the high candle and the low candle
candle_h, candle_l = candles.high_and_low_candles()
print("High candle: " + str(candle_h))
print("Low candle: " + str(candle_l))
A candle collection may be converted from a current timeframe to another timeframe. A new timeframe must be bigger than the source timeframe.
Example: the timeframe of 1 hour may be converted to 2 hours, 8 hours, 2 days, or another one, but bigger than source one. The 1 hour timeframe must not convert to 45 minutes, 30 minutes, or another one less.
A timeframe label consists of two parts: a number of time units and a time unit label.
Example of timeframes: 24M - 24 minutes timeframe, 1H - 1 hour timeframe, 11H - 11 hours timeframe, 1D - 1 day timeframe, 11D - 11 days timeframe
Possible time unit labels:
print(candles.timeframe)
candles_2h = candles.to_tf("2H")
candles_3h = candles_2h.to_tf("3H")
candles_4h = candles_3h.to_tf("4H")
candles_13h = candles_4h.to_tf("13H")
candles_1d = candles_4h.to_tf("1D")
Convert a candles collection to a pandas dataframe
from mdfft import PandasCandles
df = PandasCandles.candles_to_pandas_df(candles)
print(df)
You may create a dataset in which a row contains several candles prices.
from mdfft import PandasCandles
# 3 candles in a single row
df_windows = PandasCandles.candles_windows_to_pandas_df(candles, 3)
print(df_windows)
# Candle at 1985 index
candle = candles.candle_at(1985)
# Next candle
next_candle = candle.next()
# 30 candles to forward
next30_candle = candle.next(30)
# Prev candle
prve_candle = candle.prev()
# Prev 100 candle
prev100_candle = candle.prev(100)
# bull/bear/doji see constants in Candle.DIRECT_*
direct = candle.direct # or candle.d or candle['d'] print(direct)
# All prices different methods
price_open = candle.open # or candle.o or candle['o']
price_high = candle.high # or candle.h or candle['h']
price_low = candle.low # or candle.l or candle['l']
price_close = candle.close # or candle.c or candle['c']
price_mid = candle.mid # or candle.m or candle['m']
# Returns datetime object
candle_dt = candle.datetime # or candle.dt or candle['dt']
candle_date = candle.date
candle_time = candle.time
# Returns int
candle_year = candle.year
candle_month = candle.month
candle_day = candle.day
candle_hour = candle.hour
candle_minute = candle.minute
# Date time format
dt_str = candle.dt_as_str()
print(dt_str)
dt_str_fmt = candle.dt_as_str('%d, %b %Y')
print(dt_str_fmt)
# Timestamp
ts = candle.timestamp()
print(ts)
ts_utc = candle.timestamp_utc()
print(ts_utc)
# Candle index from start
ci = candle.index # or candle.i or candle['i']
# Candle index from back
cib = candle.index_back # or candle.ib or candle['ib']
To start painting you must import bellow classes
from mdfft import painter as p, Styler
Matplotlib is used for drawing. That means you may use colors, marker, visual elements, etc. from Matplotlib.
p.start_paint()
p.title="Candles"
p.paint(candles_collection)
p.paint_done()
p.start_paint()
p.title = "Save paint"
p.paint(candles_to_paint)
p.save_paint('candles.jpg')
The appearance is performed through the object of class Styler that included in object painter. Possible settings is an enum in class Styler.
# The appearance of the chart is configured through the object Styler
# Each candle has own the color of body, shadow and border
s = Styler()
p.title="Rainbow candles"
s.color_bear_body = [plt.cm.get_cmap('Pastel1', candles_cnt)(c) for c in range(candles_cnt)]
s.color_bear_border = [plt.cm.get_cmap('Set1_r', candles_cnt)(c) for c in range(candles_cnt)]
s.color_bear_shadow = "blue"
s.color_bull_body = [plt.cm.get_cmap('turbo', candles_cnt)(c) for c in range(candles_cnt)]
s.color_bull_border = [plt.cm.get_cmap('tab20b', candles_cnt)(c) for c in range(candles_cnt)]
s.color_bull_shadow = "green"
# Apply style
p.styler = s
# Draw candles
p.start_paint()
p.paint(candles)
p.paint_done()
Possible markers and lines
p.start_paint()
# Paint candles
p.paint(candles_to_paint)
# Paint markers for each 5 candles
for c in candles_to_paint:
if (c.index % 5) == 0:
p.paint_marker(c, "l", marker_size=4, marker_color='red', marker='v')
p.paint_marker(c, "h", marker_size=3, marker_color='black', marker='^')
p.paint_marker(c, c.mid, marker='_')
# Paint line. From hightest price to lowest price
hc, lc = candles_to_paint.high_and_low_candles()
p.paint_line( hc, hc.high, lc, lc.low, line_style="dotted", line_color="green")
# Show picture
p.paint_done()
For the success or failure of a position, the Trader class is used. The class contains a single method that returns the result of a position at a certain point.
pip = 1 / 10000
start_candle = candles.candle_at(12345)
trader = Trader()
profit, distance, action_price = trader.place_order(
candle=start_candle, # candle for order, not datetime
order_price='o', # open price - possible 'o' 'h' 'l' 'c' OR float
place_type=Trader.PLACE_TYPE_BUY, # PLACE_TYPE_BUY or PLACE_TYPE_SELL
sl=start_candle.low - pip, # stop loss
tp=start_candle.high + pip * 100, # take profit
tl=1000 # candles before force close. time to live position
)
# My profit. Positive or negative
print(f"Profit pips: {profit / pip}")
# Candles from start before close order
print(f"Candles to close order: {distance}")
# Price that close order
print(f"Price that close order: {action_price}")
FAQs
Market Data Form Finder Toolkit (MDFFT). The project goal is to make working with a raw market data easier.
We found that mdfft demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
In this segment of the Risky Business podcast, Feross Aboukhadijeh and Patrick Gray discuss the challenges of tracking malware discovered in open source softare.
Research
Security News
A threat actor's playbook for exploiting the npm ecosystem was exposed on the dark web, detailing how to build a blockchain-powered botnet.
Security News
NVD’s backlog surpasses 20,000 CVEs as analysis slows and NIST announces new system updates to address ongoing delays.