Development branch to testing new features. This develop version has a lot of improvements and fixes over master branch. The recommendation is that you use the code in this branch.
- It creates candle bar charts with MAs, RSI and MACD. These images can be sent as part of a Telegram notification or a Webhook call.
- It allows to include prices as part of the notification message.
- New configuration to easily add many coins. Check bellow for "all_pairs".
- New config var to use a custom "indicator_label" for each configured indicator and crossovers. Mainly useful for std_crossover.
- New indicator iiv (Increase In Volume) to try to identify a pump/dump.
- New indicator MA Ribbon
- New config values "hot_label" and "cold_label" for each indicator setup to set custom texts instead of the typical "hot" and "cold".
- New indicator ADX (Average Directional Index)
- New indicator Klinger Oscillator
- New indicator MACD Cross
- New indicator StochRSI Cross
- New option to customize ichimoku strategies and added chikou span
Because this is a development branch you need to build your custom Docker image. The commands listed below are intended to be run in a terminal.
Be sure you have git installed in your system.
-
Clone this repo
git clone https://github.com/w1ld3r/crypto-signal.git -
Enter to cripto-signal folder
cd crypto-signal -
Switch to develop branch
git checkout develop -
Create a config.yml file and put it into "app" folder.
-
Build your own Docker image, for example,
docker build -t dev/crypto-signals:latest . -
For testing and debugging run docker with "-t" option
docker run --rm -ti -v $PWD/app:/app dev/crypto-signals:latest -
For production run in daemon mode using "-d" option
docker run --rm -di -v $PWD/app:/app dev/crypto-signals:latest
All possible options for config.yml are almost the same for original CryptoSignal, so look here
However there are some aditional options to use.
If the indicator you are using dispatch notifications for hot/cold signals you can use the "mute_cold" option to silence (mute) cold messages. Example:
indicators:
rsi:
- enabled: true
alert_enabled: true
alert_frequency: once
signal:
- rsi
hot: 30
cold: 70
candle_period: 1d
period_count: 14
mute_cold: true
You can enable/disable charts creation.
settings:
log_level: INFO
update_interval: 600
enable_charts: true
market_pairs:
- XRP/USDT
- ETH/USDT
- BTC/USDT
....
It is very useful to process all symbols of a specific base market (BTC, ETH, USD, USDT). In this way you can avoid to write pair by pair in settings.market_pairs option.
The all_pairs option must be specified by each exchange and base market. Example:
exchanges:
binance:
required:
enabled: true
all_pairs:
- USDT
bittrex:
required:
enabled: false
all_pairs:
- ETH
- BTC
....
When using all_pairs we can exclude some pairs in particular. For example, in the following config, we are tracking all USDT pairs in Binance, except the other stable coins.
exchanges:
binance:
required:
enabled: true
all_pairs:
- USDT
exclude:
- USDC
- PAX
- BUSD
bittrex:
required:
enabled: false
all_pairs:
- ETH
....
If you want prices in your notification messages, you can use the "prices" variable or "price_value".
notifiers:
telegram:
required:
token: 791615820:AAGFgGSumWUrb-CyXtGxzAuYaabababababababa
chat_id: 687950000
optional:
parse_mode: html
template: "[{{analysis.config.candle_period}}] {{market}} {{values}} Prices: [{{prices}}]"
notifiers:
telegram:
required:
token: 791615820:AAGFgGSumWUrb-CyXtGxzAuYaabababababababa
chat_id: 687950000
optional:
parse_mode: html
template: "[{{analysis.config.candle_period}}] {{market}} {{values}} Price 15m low: [{{price_value['15m'].low}}]"
By the way, to have this feature you need to configure "ohlcv" informant for each candle period of your indicators.
informants:
....
bollinger_bands:
- enabled: false
ohlcv:
- enabled: true
signal:
- high
- low
- close
candle_period: 1h
period_count: 14
- enabled: true
signal:
- high
- low
- close
candle_period: 4h
period_count: 14
Uses TA-LIB for candlestick pattern recognition https://github.com/mrjbq7/ta-lib
- Set up
signalfor all candle patterns you want to check. - Set up
candle_checkfor how many candles you want to check. (default = 1 = checks last candle for pattern) For example your bot runs every 4 hours, but you want to check the candles of 1 hour, you can set it to trigger if the pattern happened on the last 4 candles - Set up
notificationfor what kind of notification you want when triggered, depending on what candle you check for. (default = hot) - Possible to enable coloring on candle chart: allows for 5 different candles, priority coloring from up to down
Candle_checkallows for checking multiple candles for a pattern. [1=last candle, 2=last 2 candles etc] For example, if the program runs every 4 hours but you setup a candle recognition for 1 hour you can get a notification if there was a candle found in the last 4 candles by settingcandle_check:4.
possible signals
'two_crows': talib.CDLTRISTAR
'three_black_crows': talib.CDL3BLACKCROWS
'three_inside_up_down': talib.CDL3INSIDE
'three_line_strike': talib.CDL3LINESTRIKE
'thee_stars_in_the_south': talib.CDL3OUTSIDE
'three_advancing_white_soldiers': talib.CDL3WHITESOLDIERS
'abandoned_baby': talib.CDLABANDONEDBABY
'advance_block': talib.CDLADVANCEBLOCK
'belt_hold': talib.CDLADVANCEBLOCK
'breakaway': talib.CDLBREAKAWAY
'closing_marubozu': talib.CDLCLOSINGMARUBOZU
'concealing_baby_swallow': talib.CDLCONCEALBABYSWALL
'counterattack': talib.CDLCOUNTERATTACK
'dark_cloud_cover': talib.CDLDARKCLOUDCOVER
'doji': talib.CDLDOJI
'doji_star': talib.CDLDOJISTAR
'dragonfly_doji': talib.CDLDRAGONFLYDOJI
'engulfing_pattern': talib.CDLENGULFING
'evening_doji_star': talib.CDLEVENINGDOJISTAR
'evening_star': talib.CDLEVENINGSTAR
'gap_sidesidewhite': talib.CDLGAPSIDESIDEWHITE
'gravestone_doji': talib.CDLGRAVESTONEDOJI
'hammer': talib.CDLHAMMER
'hanging_man': talib.CDLHANGINGMAN
'harami_pattern': talib.CDLHARAMI
'harami_cross_patern': talib.CDLHARAMICROSS
'high_wave_candle': talib.CDLHIGHWAVE
'modified_hikkake_pattern': talib.CDLHIKKAKEMOD
'homing_pigeon': talib.CDLHOMINGPIGEON
'identical_three_crows': talib.CDLIDENTICAL3CROWS
'in_neck_pattern': talib.CDLINNECK
'inverted_hammer': talib.CDLINVERTEDHAMMER
'kicking': talib.CDLKICKING
'kicking_bb': talib.CDLKICKINGBYLENGTH
'ladder_bottom': talib.CDLLADDERBOTTOM
'long_legged_doji': talib.CDLLONGLEGGEDDOJI
'long_line_candle': talib.CDLLONGLINE
'marubozu': talib.CDLMARUBOZU
'matching_low': talib.CDLMATCHINGLOW
'mat_hold': talib.CDLMATHOLD
'morning_doji_star': talib.CDLMORNINGDOJISTAR
'morning_star': talib.CDLMORNINGSTAR
'on_neck_pattern': talib.CDLONNECK
'piercing_pattern': talib.CDLPIERCING
'rickshaw_man': talib.CDLRICKSHAWMAN
'risfall_three_methods': talib.CDLRISEFALL3METHODS
'seperating_lines': talib.CDLSEPARATINGLINES
'shooting_star': talib.CDLSHOOTINGSTAR
'short_line_candle': talib.CDLSHORTLINE
'spinning_top': talib.CDLSPINNINGTOP
'stalled_pattern': talib.CDLSTALLEDPATTERN
'stick_sandwich': talib.CDLSTICKSANDWICH
'takuri': talib.CDLTAKURI
'tasuki_gap': talib.CDLTASUKIGAP
'thrusting_pattern': talib.CDLTHRUSTING
'tristar_pattern': talib.CDLTRISTAR
'unique_three_river': talib.CDLUNIQUE3RIVER
'upside_gap_two_crows': talib.CDLUPSIDEGAP2CROWS
'xside_gap_three_methods': talib.CDLXSIDEGAP3METHODS
indicators:
candle_recognition:
- enabled: true
alert_enabled: true
alert_frequency: always
signal:
- doji
candle_check: 1
notification: hot
candle_period: 1d
hot: 0
cold: 0
chart: true
The Aroon Oscillator is the difference between Aroon-Up and Aroon-Down. These two indicators are usually plotted together for easy comparison, but chartists can also view the difference between these two indicators with the Aroon Oscillator. This indicator fluctuates between -100 and +100 with zero as the middle line. An upward trend bias is present when the oscillator is positive, while a downward trend bias exists when the oscillator is negative.
The Aroon Oscillator is positive AND daily volume was above the (default)50-day moving average of volume = hot notification The Aroon Oscillator is negative AND daily volume was above the (default)50-day moving average of volume = cold notification
Periods in example are the default values
indicators:
aroon_oscillator:
- enabled: true
alert_enabled: true
alert_frequency: always
sma_vol_period = 50
period_count = 25
signal:
- aroon
candle_period: 1d
hot: 0
cold: 0
The Klinger oscillator (kvo) was developed by Stephen Klinger to determine the long-term trend of money flow while remaining sensitive enough to detect short-term fluctuations. The indicator compares the volume flowing through a security with the security's price movements and then converts the result into an oscillator. The Klinger oscillator shows the difference between two moving averages which are based on more than price. Traders watch for divergence on the indicator to signal potential price reversals. Like other oscillators, a signal line can be added to provide additional trade signals.
If klinger signal line (kvo_signal) > 0 AND mean price is moving up from last candle = hot notification If klinger signal line (kvo_signal) < 0 AND mean price is moving down from last candle = cold notification mean price = (high + low + close) / 3
Periods in example are the default values
indicators:
klinger_oscillator:
- enabled: true
alert_enabled: true
alert_frequency: always
ema_short_period = 34
ema_long_period = 55
signal_period = 13
signal:
- kvo
- kvo_signal
candle_period: 1d
The Average Directional Movement Index (ADX) is designed to quantify trend strength by measuring the amount of price movement in a single direction. ADX is non-directional; it registers trend strength whether price is trending up or down The ADX is part of the Directional Movement system published by J. Welles Wilder, and is the average resulting from the Directional Movement indicators. ADX calculations are based on a moving average of price range expansion over a given period of time. The default setting is 14 bars, although other time periods can be used. When the +DMI(pdi) is above the -DMI(ndi), prices are moving up, and ADX measures the strength of the uptrend. When the -DMI(ndi) is above the +DMI(pdi), prices are moving down, and ADX measures the strength of the downtrend.
0-25 absent or weak trend 25-50 strong trend 50-75 very strong trend 75-100 extremely strong trend
indicators:
adx:
- enabled: true
alert_enabled: true
alert_frequency: always
signal:
- adx
- pdi
- ndi
hot: 50
cold: 25
candle_period: 1d
period_count: 14
There is a new indicator called "iiv" enabled for default for 5m period. The hot value is 5 by default, you can adjust it as you want, may be 10 o 15. This value is a measure about how strong is the increase in volume.
The main idea is to try to identify a possible pump or dump.
indicators:
iiv:
- enabled: true
alert_enabled: true
alert_frequency: always
signal:
- iiv
hot: 5
cold: 0
candle_period: 5m
If you don't want to receive such notifications just disable the iiv indicator in your config file.
indicators:
iiv:
- enabled: false
candle_period: 5m
Of course, this indicator can be used in other candle periods, 15m, 1h.. etc.
First, read the following link to know about what it is: https://www.investopedia.com/terms/m/movingaverageribbon.asp
Second, the code used to implement this indicator was taken from another proyect called pyktrader2. It is not my own and therefore I don't know the details about how this indicator works, so what I comment below is what I think it is.
indicators:
ma_ribbon:
- enabled: true
alert_enabled: true
alert_frequency: always
signal:
- pval
- corr
hot: 10
cold: -10
hot_label: 'Uptrend is coming'
cold_label: 'Downtred is coming'
candle_period: 15m
pval_th: 20
ma_series: 5, 15, 25, 35, 45
The "corr" value changes from -100 to 100. When corr starts increasing from 0 to 100 we talk of an possible Uptrend, when corr goes from 0 to -100 we are entering in a Downtrend. The "hot" and "cold" values in config file are associated with such corr.
To confirm the change in trend, the "pval" is used. This value changes from 0 to 100, but, when this value approaches 0 it confirms the signal given by corr.
The "ma_series" config value is used to set the moving averages that will be used by the indicator.
It is a new indicator created with the idea of making the configuration of MA crossovers simpler and clearer. It is called "ma_crossover" and is configured into "indicators" section.
indicators:
ma_crossover:
- enabled: true
candle_period: 1h
alert_enabled: true
alert_frequency: once
exponential: true
ma_fast: 50
ma_slow: 100
signal:
- open
- close
hot_label: 'Uptrend is coming'
cold_label: 'Downtred is coming'
indicator_label: 'EMA 50/100 Cross'
The three basic values to configure are "exponential", "ma_slow" and "ma_fast". The "exponential" value is a true/false value indicating when to use Exponential Moving Average. If false, Simple Moving Average will be used.
"ma_fast" is the number of periods to use for the fast line, and "ma_slow" is the number of periods to use for the slow line.
"signal" isn't really a meaning value for this indicator, but we have to set something because the code base requires that parameter for all indicators. So we can simply set some ohlcv value to avoid an error config.
It is a new indicator created to easily get alerts when the close price crosses the Up/Low bands calculated for the Bollinger Bands. It is different of the Bollinger informant, and therefore it is configured in the "indicators" section.
indicators:
bollinger:
- enabled: true
candle_period: 1h
alert_enabled: true
alert_frequency: once
period_count: 20
std_dev: 2
signal:
- low_band
- close
- up_band
hot_label: 'Lower Band'
cold_label: 'Upper Band'
indicator_label: 'Bollinger Crossing'
mute_cold: false
The most important config is the number of standar deviations to use. Tipically values are 2 and 1.
The Percent B indicator reflects closing price as a percentage of the lower and upper Bollinger Bands. The idea is almost the same as Bollinger indicator, to get alerts when the close price crosses the Up/Low bands.
indicators:
bbp:
- enabled: true
candle_period: 1h
alert_enabled: true
alert_frequency: once
period_count: 20
hot: 0
cold: 0.8
std_dev: 2
signal:
- bbp
- mfi
hot_label: 'Lower Band'
cold_label: 'Upper Band'
indicator_label: 'Bollinger Crossing'
mute_cold: false
The existing "macd" indicator always sends hot signals when the macd > 0, and cold signals when macd < 0. I repeat, always. Perhaps this is not the desired behavior and you only want to receive notifications when macd crosses the signal line, then this new MACD Cross indicator is the appropriate.
indicators:
macd_cross:
- enabled: true
candle_period: 4h
alert_enabled: true
alert_frequency: always
signal:
- macd
- signal
hot_label: 'Uptrend is coming'
cold_label: 'Downtred is coming'
indicator_label: 'MACD Cross 4h'
mute_cold: false
This indicator is useful if you are interested in knowing when smooth K line crosses the smooth D line of the StochRSI. Unlike the StochRSI indicator, this new indicator allows you to set values for K and D.
indicators:
stochrsi_cross:
- enabled: true
candle_period: 4h
alert_enabled: true
alert_frequency: always
smooth_k: 10
smooth_d: 3
signal:
- stoch_rsi
- smooth_k
- smooth_d
hot_label: 'Uptrend is coming'
cold_label: 'Downtred is coming'
indicator_label: 'StochRSI Cross 4h'
mute_cold: false
The config for a webhook notifier is the same as original CryptoSignal, no changes here, BUT the data sent in the request is completely different.
notifiers:
telegram:
required:
token: 791615820:AAGFgGSumWUrb-CyXtGxzAuY7UPxxxxx
chat_id: 687957000
optional:
parse_mode: html
template: "[{{indicator_label}}] {{market}} {{values}}, Prices: [{{prices}}]"
webhook:
required:
url: http://somename.or.ip.here:8888/someuri
optional:
username: null
password: null
Then in your webhook you will have a parameter "messages" which is a list/array of the notification messages in form of string values. Apply a json decode to access all values. In python pseudo code:
msg = self.get_argument('messages')
print ( json.loads(msg) )
[{'values': {'rsi': '84.20'}, 'exchange': 'binance', 'market': 'VET/USDT', 'base_currency': 'VET', 'quote_currency': 'USDT', 'indicator': 'rsi', 'indicator_number': 0, 'analysis': {'config': {'enabled': True, 'alert_enabled': True, 'alert_frequency': 'always', 'signal': ['rsi'], 'hot': 40, 'cold': 60, 'candle_period': '15m', 'period_count': 13}, 'status': 'cold'}, 'status': 'cold', 'last_status': 'cold', 'creation_date': '2019-01-04 15:31:40', 'indicator_label': 'rsi 15m'}]
If you have enable_charts: true, you will have a parameter "chart" in the same way of a HTTP file upload. Reading this file depends 100% in your backend. Example python pseudo code to save the received chart:
fileinfo = self.request.files['chart'][0]
print ("filename is " , fileinfo['filename'], ', content_type: ', fileinfo['content_type'])
fname = fileinfo['filename']
extn = os.path.splitext(fname)[1]
cname = str(uuid.uuid4()) + extn
fh = open(__UPLOADS__ + cname, 'wb')
fh.write(fileinfo['body'])
The default ichimoku strategy is replaced with TK cross and future cross. You can either select one of them or both. If you choose both, you will receive a signal when a TK cross or future cross happens.
indicators:
ichimoku:
- enabled: true
alert_enabled: true
alert_frequency: once
signal:
- leading_span_a
- leading_span_b
- tenkansen
- kijunsen
hot: true
cold: true
hot_label: 'Uptrend is coming'
cold_label: 'Downtred is coming'
candle_period: 4h
indicator_label: 'Ichimoku'
If you want to add your own strategy, you could add your strategy to app/user_data/strategies folder (if it doesn't exist, create one). The name of your strategy should be the same name as the class name. Your custom strategy should use te following template:
class <name of your strategy>:
def analyze(ichimoku_values, dataframe):
"""Custom Ichimoku strategy
Args:
ichimoku_values (list, optional): historical list containg 'leading_span_a', 'leading_span_b', 'tenkansen', 'kijunsen' and 'chikou_span'
dataframe (float, optional): historical pandas dataframe containing 'timestamp', 'open', 'high', 'low', 'close' and 'volume'
Returns:
is_hot, is_cold: A dataframe containing the indicators and hot/cold values.
"""
return is_hot, is_cold
Then in your config.yaml, add the custom_strategy property, this will override the default behavior.
indicators:
ichimoku:
- enabled: true
custom_strategy: <name of your strategy>
alert_enabled: true
alert_frequency: once
signal:
- leading_span_a
- leading_span_b
- tenkansen
- kijunsen
- chikou_span
hot: true
cold: true
hot_label: 'Uptrend is coming'
cold_label: 'Downtred is coming'
candle_period: 4h
indicator_label: 'Ichimoku'
Example of a custom strategy:
class CustomStrategy:
def analyze(ichimoku_values, dataframe):
date = dataframe.index[-1]
leading_span_date = ichimoku_values.index[-1]
tk_cross_hot = ichimoku_values['tenkansen'][date] > ichimoku_values['kijunsen'][date]
cloud_breakout_hot = dataframe['close'][date] > ichimoku_values['leading_span_a'][date]
price_hot = dataframe['close'][date] > ichimoku_values['kijunsen'][date]
is_hot = tk_cross_hot and price_hot and cloud_breakout_hot
tk_cross_cold = ichimoku_values['tenkansen'][date] < ichimoku_values['kijunsen'][date]
cloud_breakout_cold = dataframe['close'][date] < ichimoku_values['leading_span_a'][date]
price_cold = dataframe['close'][date] < ichimoku_values['kijunsen'][date]
is_cold = tk_cross_cold and price_cold and cloud_breakout_cold
return is_hot, is_cold
Setting a custom text for the "hot" or "cold" signals allows to have a really cool notification message.
indicators:
rsi:
- enabled: true
alert_enabled: true
alert_frequency: always
signal:
- rsi
hot: 40
cold: 60
hot_label: We Are In Oversold!
cold_label: Attention! Overbought!
candle_period: 1h
period_count: 13
So, in the message template the "hot_cold_label" variable will have one of the two possible values.
template: "[{{indicator_label}}] **{{hot_cold_label}}** {{market}} Prices: [{{prices}}]"
To use this feature it is necessary to configure "ohlcv" informant for each candle period of your indicators. For example, this config is used to get High, Low and Close prices for 1h and 4h indicators.
informants:
....
bollinger_bands:
- enabled: false
ohlcv:
- enabled: true
signal:
- high
- low
- close
candle_period: 1h
period_count: 14
- enabled: true
signal:
- high
- low
- close
candle_period: 4h
period_count: 14
Then you can use the "price_value" variable to have the values of prices and be able to do some operations on them.
notifiers:
telegram:
required:
token: 580514307:AAETsNsxs4QCdyEZ59vVROLlBxxxxx
chat_id: 2073900000
optional:
parse_mode: html
template: "{{ market }}
BUY {{ price_value['1h'].close }}
SL: {{ decimal_format|format(price_value['4h'].low * 0.9) }}
TP: {{ decimal_format|format(price_value['1h'].close * 1.02) }} {{ decimal_format|format(price_value['4h'].close * 1.04) }} "
The code for "decimal_format" and "format" is necessary to obtain the prices formatted with the corresponding zeros.