/AutoTrade

Quantitative Analysis Tool

Primary LanguagePython

AutoTrade

Quantitative Analysis Toolkit.

Pre-requisites

  • Python 3.7+

    • requests: HTTP requests
      • requests-futures: Asynchronous HTTP requests
    • selenium: Powerful raw crawling
      • beautifulsoup4: HTML parser
    • psycopg2: PostgreSQL Python binder
    • asyncpg: Asynchronous PostgreSQL Python binder
    • pony: Excellent ORM for database
    • ccxt: Cryptocurrency exchanges binder

  • PostgreSQL 10+

    • pgadmin 4 (additional): GUI for PostgreSQL

Abstraction

  • Market: (base, quote, exchange)

    This structure is used to introduce markets.

    • base: The symbol of base stock or currency. ex) USDT
    • quote: The symbol of target stock or currency. ex) BTC
    • exchange: The exchange of that trading pair. ex) Binance

  • PriceTick: (market, timestamp, price, volume)

    This structure is used to store market price tick data.

    • market: Market.
    • timestamp: The timestamp of transaction.
    • price: The price of transaction.
    • volume: The volume of transaction.

  • OHLCV: (market, interval, timestamp, open, close, high, low, volume)

    This structure is used to store market price OHLCV data.

    • market: Market.
    • interval: The length of period, stored using datetime.timedelta.
    • timestamp: The timestamp of starting time of period unit.
    • open, close, high, low, volume: OHLCV data, stored using Decimal(24,8).

Database

Using PostgreSQL DB from localhost or AWS.

  • Table structure:

    Constructed one table per market due to the optimization.

    Table name = PriceData_(exchange)_ (base)_ (quote)_ (minuteInterval) mins (ex: PriceData_Bitstamp_USD_BTC_1mins)

    timestamp open high low close volume
    2017-12-24 13:00:00+00 13019.82 13019.82 13019.82 13019.82 0.05
    2017-12-24 13:01:00+00 13098.09 13098.09 13098.09 13098.09 0.96350729
    2017-12-24 13:02:00+00 13148.11 13148.11 13148.11 13148.11 0.07149971
    2017-12-24 13:03:00+00 13140 13140 13140 13140 0.26503305

    Note that all columns has constraint NOT NULL.

    • timestamp: The timestamp of starting time of the period, using TIMESTAMPTZ. This column is the primary key, and has additional constraint CHECK(timestamp <= NOW()).
    • open: Opening price of the period, using NUMERIC(24, 8).
    • high: Highest price of the period, using same type as open.
    • low: Lowest price of the period, using same type as open.
    • close: Closing price of the period, using same type as open.
    • volume: Trading volume of the period, using same type as open. This column has additional constraint CHECK(volume > 0). This means there is no data for periods with no trading volume.
  • Example function
def smallf(timestamp: datetime, i: int): return round((timestamp - i * timedelta(minutes = 1)).timestamp())
if len(price_data) < 5:
   return None, False, False
else:
   recentAverage = statistics.mean(price_data[smallf(timestamp, i)][criteria] for i in range(50)
                                   if smallf(timestamp, i) in price_data)
   if smallf(timestamp, 0) not in price_data: return recentAverage, False, False
   nowPrice = price_data[smallf(timestamp, 0)][criteria]
   return recentAverage, recentAverage > nowPrice, recentAverage < nowPrice