First, an Apology

There is a very high chance that the indicator you want to visualize has not been implemented yet. So far, I've only implemented the few indicators that I personally use. Fortunately, new indicator visualizations are easy to implement.

How to Implement a New Indicator Visualization

Implement a denominated_price function.

Is the output of the indicator a price value?

Moving averages are the best example of this. It takes a price as an input and generates a price value as an output, so demominated_price should return true.

What about the opposite case?

Oscillators typically answer false for this. Think of RSI that takes close prices as input but its output is a value between 0 and 100 regardless of how high or low the price is. The domain and range are different for the RSI function (and many other oscillators) so denominated_price should return false.

denominated_price(sma::SMA) = true
denominated_price(rsi::RSI) = false

Implement a visualize function.

A visualize function takes 3 parameters.

1. An indicator instance
2. A Dict of visualization options. By convention, nothing is also allowed, and that means to use the defaults.
3. A DataFrame containing all the values generated for the Chart

Its job is to return something that LightweightCharts knows how to draw. The most common return value is an lwc_line.

Here's what visualize for SMA looks like. It's one of the simplest indicators to visualize, because it's a single line.

"""$(TYPEDSIGNATURES)

Return an lwc_line for visualizing an SMA indicator.
"""
function visualize(sma::SMA, opts::Union{AbstractDict,Nothing}, df::DataFrame)
    start = sma.period
    name = indicator_fields(sma)[1]
    kwargs = Dict(
        :label_name => "SMA $(sma.period)",
        :line_color => "#B84A62",
        :line_width => 2
    )
    if opts !== nothing                 # (opts == nothing) means use the defaults
        merge!(kwargs, opts)
    end
    return lwc_line(                    # the end goal is to return an lwc visual
        df.ts[start:end],
        [df[!, name][start:end]...];
        kwargs...
    )
end

If you need to return more than one line, return them in a vector. A good example of this is BB.jl which returns 3 lwc_lines.

You're also not limited to lwc_line. You can use any visualization function supported by LightweightCharts.jl.

Examples

All indicator visualizations fall into one of these four categories.

When implementing a new visualization, feel free to look at the source for these and use them as a starting point for your own work.

Finally

• Add an include for your newly implemented indicator at the bottom of src/TechnicalIndicatorCharts.jl.
• Add a test for your indicator. A basic test that makes sure it doesn't crash when visualize is run is enough.
• Mark that the indicator is done in the README.md for this project.

Thanks

Thanks to all future indicator visualization authors.