Anansi

Dependencies

Python, Pip, Poetry.

To install poetry, on osx, linux or bashonwindows terminals, type it:

curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | python -

Alternatively, poetry could be installed by pip (supposing you have python and pip already installed):

pip install poetry

Consuming on Jupyter notebook

That is only a suggestion, you could run anansi on any python terminal. Only tested on linux.

Perform the commands:

poetry install
poetry run python -m ipykernel install --user --name=$(basename $(pwd))
poetry run jupyter notebook > jupyterlog 2>&1 &

Straight to the point: Running Default Back Testing Operation

Importing Dependencies

from anansi.tradingbot.models import *
from anansi.tradingbot import traders
from anansi.tradingbot.views import create_user, create_default_operation

Add a new user

my_user_first_name = "John"

create_user(first_name=my_user_first_name,
                   last_name="Doe",
                   email = "{}@email.com".format(my_user_first_name.lower()))

Creating a default operation

my_user = User[1]
create_default_operation(user=my_user)

Instantiating a trader

my_op = Operation.get(id=1)
my_trader = traders.DefaultTrader(operation=my_op)

Run the trader

my_trader.run()

Playing with the database models

Getting all users

users = select(user for user in User)
users.show()

id|first_name|last_name|login_displayed_name|email         
--+----------+---------+--------------------+--------------
1 |John      |Doe      |                    |john@email.com

my_user.first_name

'John'

Some operation attribute

my_op.stop_loss.name

'StopTrailing3T'

Some trader attribute

my_trader.Classifier.parameters.time_frame

'6h'

Updating some attributes

before_update = my_trader.operation.position.side, my_trader.operation.position.exit_reference_price

my_trader.operation.position.update(side="Long", exit_reference_price=1020.94)

after_update = my_trader.operation.position.side, my_trader.operation.position.exit_reference_price

before_update, after_update

(('Zeroed', None), ('Long', 1020.94))

Requesting klines

Klines treated and ready for use, including market indicators methods

The example below uses the 'KlinesFromBroker' class from the 'handlers' module ('marketdata' package), which works as an abstraction over the data brokers, not only serializing requests (in order to respect brokers' limits), but also conforming the klines like a pandas dataframe, extended with market indicator methods.

from anansi.marketdata.handlers import KlinesFromBroker

BinanceKlines = KlinesFromBroker(
  broker_name="binance", ticker_symbol="BTCUSDT", time_frame="1h")

newest_klines = BinanceKlines.newest(2167)

newest_klines

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

Applying simple moving average indicators

indicator = newest_klines.apply_indicator.trend.simple_moving_average(number_of_candles=35)

indicator.name, indicator.last(), indicator.serie

('sma_ohlc4_35',
 10669.49407142858,
 0                NaN
 1                NaN
 2                NaN
 3                NaN
 4                NaN
             ...     
 2162    10619.190500
 2163    10632.213571
 2164    10644.682643
 2165    10657.128857
 2166    10669.494071
 Length: 2167, dtype: float64)

newest_klines

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

Same as above, but showing indicator column

indicator = newest_klines.apply_indicator.trend.simple_moving_average(
  number_of_candles=35, indicator_column="SMA_OHLC4_n35")

newest_klines

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

Raw klines, using the low level abstraction module "data_brokers"

DISCLAIMER: Requests here are not queued! There is a risk of banning the IP or even blocking API keys if some limits are exceeded. Use with caution.

from anansi.marketdata import data_brokers

BinanceBroker = data_brokers.BinanceDataBroker()

my_klines = BinanceBroker.get_klines(ticker_symbol="BTCUSDT", time_frame="1m")

my_klines

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

Same as above, but returning all information get from the data broker

my_klines = BinanceBroker.get_klines(ticker_symbol="BTCUSDT", time_frame="1m", show_only_desired_info=False)

my_klines

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}