Techan

TechAn is a technical analysis library for Go! It provides a suite of tools and frameworks to analyze financial data and make trading decisions.

NOTE: This is my fork of the original techan library found at https://github.com/sdcoffey/techan. My fork preserves the structure of the timeseries and inidicator calculation while completely overhauling the strategy and trade simulation to make it more functional for my real use.

Features

• Basic and advanced technical analysis indicators
• Strategy building
• Backtesting and account simulation

Installation

$ go get github.com/schmidthole/techan

Examples

Quickstart

series := techan.NewTimeSeries()

// fetch this from your preferred exchange
dataset := [][]string{
	// Timestamp, Open, Close, High, Low, volume
	{"1234567", "1", "2", "3", "5", "6"},
}

for _, datum := range dataset {
	start, _ := strconv.ParseInt(datum[0], 10, 64)
	period := techan.NewTimePeriod(time.Unix(start, 0), time.Hour*24)

	candle := techan.NewCandle(period)
	candle.OpenPrice = big.NewFromString(datum[1])
	candle.ClosePrice = big.NewFromString(datum[2])
	candle.MaxPrice = big.NewFromString(datum[3])
	candle.MinPrice = big.NewFromString(datum[4])

	series.AddCandle(candle)
}

closePrices := techan.NewClosePriceIndicator(series)
movingAverage := techan.NewEMAIndicator(closePrices, 10) // Create an exponential moving average with a window of 10

fmt.Println(movingAverage.Calculate(0).FormattedString(2))

Creating trading strategies

A Strategy in Techan is the application of a Rule against a particular security/asset. For ease of reference, the Strategy struct contains the original Timeseries, all Indicators used to calculate the Rule, and the rull itself.

// using the timeseries and indicators from above...

// we simply create a rule that is satisfied if the price is above the moving average.
rule := OverIndicatorRule {
    First: movingAverage,
    Second: closePrices
}

strategy := Strategy{
    Security: "TEST",
    Timeseries: series,
    Indicators: map[string]Indicator{
        "ema10": movingAverage,
    },
    Rule: rule,
}

Strategies against individual securities/assets can be combined into a more comprehensive strategy using an Allocator. An allocator is simply an interface that accepts a list of strategies and outputs a portfolio allocation. This can be as simple or as complex as needed.

// again using the strategy defined above...

// we create a naive allocator which allows a single position to be 50% of a portfolio and 100% of a portfoliio
// to be allocated.
allocator := NewNaiveAllocator(big.NewDecimal(0.5), big.NewDecimal(1.0))

allocations := allocator.Allocate(0, []Strategy{strategy})

Using the outputted portfolio Allocations, we can then create a TradePlan by combining information from an Account. The TradePlan will use the current security prices, open positions in the account, and the allocations provided to identify the orders which can be placed to achieve the desired allocation. This can be used to calculate orders to execute with a broker based on the strategy results.

account := NewAccount()
account.Deposit(big.NewDecimal(10000.00))

// create pricing data from timeseries candles above.
pricing := Pricing{
    "ONE": big.NewDecimal(2.0)
}

tradePlan, _ := CreateTradePlan(allocations, pricing, account)

Putting it all together, we can use all of the components to run a full backtest of a strategy against the historical Timeseries data. Combining multiple Strategy objects along with the Allocator allow for complex strategies to be modelled over time.

// we are still using everyting defined above...

backtest := NewBacktest([]Strategy{strategy}, allocator, account)
history, _ := backtest.Run()

Running a backtest will about an AccountHistory object, which contains all of the relevant backtest snapshots/results over time. The account history can be used to perform analysis of the strategy's performance metrics.

profit := history.TotalProfit()

Credits

Techan is heavily influenced by the great ta4j. Many of the ideas and frameworks in this library owe their genesis to the great work done over there.

License

Techan is released under the MIT license. See LICENSE for details.