A python library for computing technical analysis indicators on streaming data.
pip install streaming-indicators
There are many other technical analysis python packages, most notably ta-lib, then why another library?
All other libraries work on static data, you can not add values to any indicator. But in real-time trading system, price values (ticks/candles) keeps streaming, and indicators should update on real-time. This library is for that purpose.
Each indicator is a class, and is statefull. It will have 3 main functions:
- Constructor: initialise all parameters such as period.
- update: To add new data point in the indicator computation. Returns the new value of the indicator.
- compute: Compute indicator value with a new data point, but don't update it's state. This is useful in some cases, for example, compute indictor on ltp, but don't update it.
- Simple Moving Average (SMA)
import streaming_indicators as si
period = 14
SMA = si.SMA(period)
for idx, candle in candles.iterrows():
sma = SMA.update(candle['close'])
print(sma)
- Exponential Moving Average (EMA)
period = 14
EMA = si.EMA(period)
for idx, candle in candles.iterrows():
ema = EMA.update(candle['close'])
print(ema)
- Weighted Moving Average (WMA)
- Smoothed Moving Average (SMMA)
- Relative Strength Index (RSI)
period = 14
RSI = si.RSI(period)
for idx, candle in candles.iterrows():
rsi = RSI.update(candle['close'])
print(rsi)
- True Range (TRANGE)
- Average True Range (ATR)
atr_period = 20
ATR = si.ATR(atr_period)
for idx, candle in candles.iterrows():
atr = ATR.update(candle) # Assumes candle to have 'open',high','low','close' - TODO: give multiple inputs to update.
print(atr)
- SuperTrend (SuperTrend)
st_atr_length = 10
st_factor = 3
ST = si.SuperTrend(st_atr_length, st_factor)
for idx, candle in candles.iterrows():
st = ST.update(candle)
print(st) # (st_direction:1/-1, band_value)
- Heikin Ashi Candlesticks (HeikinAshi)
HA = si.HeikinAshi()
for idx, candle in candles.iterrows():
ha_candle = HA.update(candle)
print(ha_candle) # {'close': float, 'open': float, 'high': float, 'low': float}
- Renko Bricks (Renko)
# For fixed brick size
brick_size = 20
Renko = si.Renko()
for idx, candle in candles.iterrows():
bricks = Renko.update(candle['close'], brick_size)
print(bricks) # [{'direction': 1/-1, 'brick_num': int, 'wick_size': float, 'brick_size': float, 'brick_end_price': float, 'price': float}, {}]: list of bricks formed after this candle
# For brick size using ATR
atr_period = 20
ATR = si.ATR(atr_period)
Renko = si.Renko()
for idx, candle in candles.iterrows():
atr = ATR.update(candle)
print(atr)
bricks = Renko.update(candle['close'], atr)
print(bricks)
- Order Checking (IsOrder) Checks if the running sequence is in a given order, eg increasing, decreasing, exponential, etc. Useful when checking if consecutive n candles/ltps were increasing.
period = 10
all_increasing = IsOrder('>', period)
for idx, candle in candles.iterrows():
is_increasing = all_increasing.update(candle['close'])
print(is_increasing) # True/False
- Not all indicators currently support compute method.
- Add documentation.
- SuperTrend,HeikinAshi,ATR depends on key names, eg ('open','close'). Should be independent, i.e. given in input.
- Implement more indicators.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent