AR Forecast Scatterplot

// This source code is subject to the terms of the Mozilla Public License 2.0 at https://mozilla.org/MPL/2.0/
//  /$$$$$$   /$$                                                         /$$                                            
// /$$__  $$ | $$                                                        | $$                                            
//| $$  \__//$$$$$$    /$$$$$$  /$$    /$$ /$$$$$$   /$$$$$$   /$$$$$$$ /$$$$$$    /$$$$$$  /$$    /$$ /$$$$$$   /$$$$$$$
//|  $$$$$$|_  $$_/   /$$__  $$|  $$  /$$//$$__  $$ /$$__  $$ /$$_____/|_  $$_/   /$$__  $$|  $$  /$$//$$__  $$ /$$_____/
// \____  $$ | $$    | $$$$$$$$ \  $$/$$/| $$$$$$$$| $$  \__/|  $$$$$$   | $$    | $$$$$$$$ \  $$/$$/| $$$$$$$$|  $$$$$$ 
// /$$  \ $$ | $$ /$$| $$_____/  \  $$$/ | $$_____/| $$       \____  $$  | $$ /$$| $$_____/  \  $$$/ | $$_____/ \____  $$
//|  $$$$$$/ |  $$$$/|  $$$$$$$   \  $/  |  $$$$$$$| $$       /$$$$$$$/  |  $$$$/|  $$$$$$$   \  $/  |  $$$$$$$ /$$$$$$$/
// \______/   \___/   \_______/    \_/    \_______/|__/      |_______/    \___/   \_______/    \_/    \_______/|_______/ 

//       ___________________
//      /                   \
//     /  _____        _____ \
//    /  /     \      /     \  \
// __/__/       \____/       \__\_____
//|           ___________           ____|
// \_________/           \_________/
//            \  /////// /
//             \/////////
// © Steversteves
//@version=5
indicator("AR Forecast Scatterplot [SS]", overlay=true, max_lines_count = 500, max_labels_count = 500, max_bars_back = 1000)

import Steversteves/SPTS_StatsPakLib/1 as spts 

//@peacefulLizard50262
flush(source)=>
    if source.size() > 0
        for i = source.size() - 1 to 0
            source.get(i).delete()
            source.remove(i)

t1 = "This will plot a line from the first Result to the last Result for the forecast length"
t2 = "Showing results and variances will plot both the results (blue) and the standard errors based on the variance in the dataset. Only plotting reuslts will only plot the most likely outcome without the variance."
t3 = "This will omit the need for you to identify the lookback length yourself and will autofind the best lookback period"
// Inputs 
src =       input.source(close, "Forecast Source"),      train1 = input.int(150, "Train Time"),        len = input.int(14, "Forecast Length"),    
typ =       input.string("Scatter Plot", "Scatter Plot Type", ["Line Plot", "Scatter Plot"]),            show_stats = input.bool(true, "Show Model Statistics")
line_bf =   input.bool(true, "Show Trendline", tooltip = t1),       show_others = input.string("Show Results and Variances","Show Results/Variance", ["Show Results and Variances", "Show Results Only"], tooltip = t2)
autolen =   input.bool(true, "Use Auto lookback based on strongest trend", tooltip = t3)
onlytrend = input.bool(false, "Only Show Trendline") 
// Arrays 
results = array.new_float() 
ucl = array.new_float() 
lcl = array.new_float() 

// Find Trend
trend_finder(len) => 
    ta.correlation(close, time, len) 

t50 = trend_finder(50),     t100 = trend_finder(100),       t150 = trend_finder(150),       t200 = trend_finder(200),   t250 = trend_finder(250),       t300 = trend_finder(300),       t350 = trend_finder(350),
t400 = trend_finder(400),   t450 = trend_finder(450),       t500 = trend_finder(500),       t550 = trend_finder(550),   t600 = trend_finder(600),       t650 = trend_finder(650),       t700 = trend_finder(700),
strongest_trend = math.max(t50, t100, t150, t200, t250, t300, t350, t400, t450, t500, t550, t600, t650, t700) 
int find_trend = strongest_trend == t50 ? 50 : strongest_trend == t100 ? 100 : strongest_trend == t150 ? 150 : strongest_trend == t200 ? 200 : strongest_trend == t250 ? 250 : strongest_trend == t300 ? 300 : strongest_trend == t350 ? 350 : strongest_trend == t400 ? 400 : strongest_trend == t450 ? 450 : strongest_trend == t500 ? 500 : strongest_trend == t550 ? 550 : strongest_trend == t600 ? 600 : strongest_trend == t650 ? 650 : strongest_trend == t700 ? 700 : 100

int train = 0 
if autolen 
    train := find_trend 
else 
    train := train1


// Forecast 
[forecast_result, b, c] = spts.f_forecast(src, src[1], train, len, results, ucl, lcl)

// Model Statistics 
cor = ta.correlation(src, src[len], train) 
r2 = math.pow(cor,2) 
max_val = array.max(ucl) 
min_val = array.min(lcl) 

f_primary_labels(id) => 
    label.new(bar_index + id, array.get(results, id), text = "", color = color.blue, style = label.style_circle, size = size.tiny)
f_secondary_labels(id) => 
    label.new(bar_index + id, array.get(ucl, id), text = "", color = color.lime, style = label.style_circle, size = size.tiny)
f_secondary_labels2(id) => 
    label.new(bar_index + id, array.get(lcl, id), text = "", color = color.red, style = label.style_circle, size = size.tiny)    
f_primary_line(id) => 
    line.new(bar_index + id, y1 = array.get(results, id), x2 = bar_index + 1 + id, y2 = array.get(results, id), color = color.blue, width = 3)
f_secondary_line(id) =>     
    line.new(bar_index + id, y1 = array.get(ucl, id), x2 = bar_index + 1 + id, y2 = array.get(ucl, id), color = color.lime, width = 3)
f_secondary_line2(id) =>    
    line.new(bar_index + id, y1 = array.get(lcl, id), x2 = bar_index + 1 + id, y2 = array.get(lcl, id), color = color.red, width = 3)   

var label1_ar = array.new<label>() 
var label2_ar = array.new<label>() 
var label3_ar = array.new<label>() 
var line1_ar = array.new<line>() 
var line2_ar = array.new<line>() 
var line3_ar = array.new<line>() 

flush(label1_ar)
flush(label2_ar)
flush(label3_ar)
flush(line1_ar)
flush(line2_ar)
flush(line3_ar)

var forecast_line = array.new<line>()
flush(forecast_line)

if barstate.islast and line_bf 
    if forecast_result.size() > 0
        value = 0.0
        for i = 1 to forecast_result.size() - 1
            forecast_line.push(line.new(bar_index + i - 1, forecast_result.get(i - 1), bar_index + i , forecast_result.get(i), color = color.purple, width = 3))


// Plots 
var table data = table.new(position.middle_right, 2, 7, bgcolor = color.blue, frame_color = color.rgb(0, 0, 0), frame_width = 4)
if show_stats 
    table.cell(data, 1, 1, text = "Model Data", bgcolor = color.blue, text_color = color.white) 
    table.cell(data, 1, 2, text = "Correlation: " + str.tostring(math.round(cor,2)), bgcolor = color.blue, text_color = color.white) 
    table.cell(data, 1, 3, text = "R2: " + str.tostring(math.round(r2, 3)), bgcolor = color.blue, text_color = color.white) 
    table.cell(data, 1, 4, text = "Max Forecasted Value: " + str.tostring(math.round(max_val, 2)), bgcolor = color.blue, text_color = color.white) 
    table.cell(data, 1, 5, text = "Min Forecasted Value: " + str.tostring(math.round(min_val, 2)), bgcolor = color.blue, text_color = color.white) 
    table.cell(data, 1, 6, text = "Lookback Length: " + str.tostring(train), bgcolor = color.blue, text_color = color.white) 
if barstate.islast and typ == "Scatter Plot" and not onlytrend
    for idx = 0 to len 
        label1_ar.push(f_primary_labels(idx))
        if show_others == "Show Results and Variances"
            label2_ar.push(f_secondary_labels(idx))
            label3_ar.push(f_secondary_labels2(idx))

if barstate.islast and typ == "Line Plot" and not onlytrend
    for idx = 0 to len 
        line1_ar.push(f_primary_line(idx))
        if show_others == "Show Results and Variances"
            line2_ar.push(f_secondary_line(idx))
            line3_ar.push(f_secondary_line2(idx))