ixjb94 / indicators Goto Github PK

I'm not a professional programmer, so my code may seem "dirty" :)
This is a good popular indicator. ( https://www.tradingview.com/script/YHZimEz8-B-Xtrender-Puppytherapy/ )
But I still want to share!

async function bXtrender(close, short_l1, short_l2, short_l3, long_l1, long_l2) {
    const ema_short_l1 = await ta.ema(close, short_l1);
    const ema_short_l2 = await ta.ema(close, short_l2);
    const ema_lohg_l1 = await ta.ema(close, long_l1);
    const subResult = ema_short_l1.map((v, i) => v - ema_short_l2[i]);
    const longXtrend = await ta.rsi(ema_lohg_l1, short_l3);
    const shortXtrend = await ta.rsi(subResult, long_l2);
    const shortTermXtrender = shortXtrend.map(v => v - 50);
    const longTermXtrender = longXtrend.map(v => v - 50);
    const len = 5;
    const src = shortTermXtrender;

    let xe1_1 = await ta.ema(src, len);
    let xe2_1 = await ta.ema(xe1_1, len);
    let xe3_1 = await ta.ema(xe2_1, len);
    let xe4_1 = await ta.ema(xe3_1, len);
    let xe5_1 = await ta.ema(xe4_1, len);
    let xe6_1 = await ta.ema(xe5_1, len);

    let b_1 = 0.7;
    let c1_1 = (-b_1 * b_1 * b_1);
    let c2_1 = (3 * b_1 * b_1 + 3 * b_1 * b_1 * b_1);
    let c3_1 = (-6 * b_1 * b_1 - 3 * b_1 - 3 * b_1 * b_1 * b_1);
    let c4_1 = (1 + 3 * b_1 + b_1 * b_1 * b_1 + 3 * b_1 * b_1);

    let nT3Average_xe6_1 = xe6_1.map(v => v * c1_1);
    let nT3Average_xe5_1 = xe5_1.map(v => v * c2_1);
    let nT3Average_xe4_1 = xe4_1.map(v => v * c3_1)
    let nT3Average_xe3_1 = xe3_1.map(v => v * c4_1);

    const nT3Average = nT3Average_xe6_1.map((v, i) => v + nT3Average_xe5_1[i] + nT3Average_xe4_1[i] + + nT3Average_xe3_1[i]);
    return [longTermXtrender, nT3Average];
}

Do it:

                const short_l1 = 5;
                const short_l2 = 20;
                const short_l3 = 15;
                const long_l1 = 20;
                const long_l2 = 15;
                const results = await bXtrender(close, short_l1, short_l2, short_l3, long_l1, long_l2);
                console.log(results);

I checked with the original indicator on tradingview - they are like twins :)

If you have time, please add it to your library

P.S. I tried to make a kvo indicator like here (https://www.tradingview.com/script/Ka0gS4Kh-Klinger-Volume-Oscillator-KVO/), but it didn’t work

I will be glad if you can do this.

Thank you for the excellent library of indicators!

ERROR in ./node_modules/technical-indicators/src/core/indicators.ts 10:46
Module parse failed: Unexpected token (10:46)
File was processed with these loaders:

• ./node_modules/source-map-loader/dist/cjs.js
  You may need an additional loader to handle the result of these loaders.
  | * @returns
  | */

async normalize(originalLength: number, source: Array | string, empty = NaN): Promise<Array> {
| const diff = originalLength - source.length
|

Getting this error when i used your library in my react.js project.

Already tried every solution mentioned here
https://stackoverflow.com/questions/63423384/you-may-need-an-additional-loader-to-handle-the-result-of-these-loaders.

could not get out of the issue.

It could be very helpful for me if you solve this error.

If i comment the errored code. i will get error at other code lines

Hi there

First of all thanks because this is great for those of us in TS :)

I have 2 questions:

1. Is the data returned from say ta.rsi(close5m, rsiPeriod) is in desc or asc order? meaning is index 0 the most recent?
2. Why is it that the data is SO different from the tradingview indiators?

I think I am doing things correctly but please try to double check:

I fetch kline data for a good chunk (7 to 14 days) on 5 min, 15 min and 1h interval. For each of those individual datasets, I run the RSI as shown before and AO as follows: ta.ao(high15m, low15m).

The results are wildly different when i compare the numbers with the indicators on tradingview on the same coin - how come?

RSI at index 0 on 5m tf for OPUSDT shows 29.9 ta.rsi I call but on tradingview is like 60. The last index is 67, and although closer is still quite far off id say.

Any idea?

Hi there.

I tried using ta.ao(high, low) but nothing is returned. I can get data on the same dataset from ta.rsi but not .ao, everything is undefined. Why? Can you explain?

Also I wonder why RSI is quite off from how for example the TMA v2 indicator or VMV cipher b (or any tradingview indicator really) calculates it....

Any help is greatly appreciated.

Add benchmark

Thank you for this library! I hope it won't be too cheeky to offer a wee bit of feedback -- marking the functions as async offers no advantage as all the code I can see is all synchronous. Users will end up awaiting the results for no discernible benefit. Not a big deal, but it can complicate things if running inside complex loops, causing the caller to have to await every layer up the call stack up to that point.

Hello, I'm wondering if floating point error/decimal calculations are handled under the hood?

Hi again,

sorry to bother you again. Everything is working great now but i was trying to use bollinger bands and everything it returns is NaN, what am I doing wrong? Any Idea?

I tried passing an array of objects like these:

 {
        time: parseInt(x[0]),
        open: parseFloat(x[1]),
        high: parseFloat(x[2]),
        low: parseFloat(x[3]),
        close: parseFloat(x[4]),
        volume: parseFloat(x[5]),
      };

and also an array of arrays, but no cigar.

I am calling it like this: ta.bbands(historicalData1min, 20, 3)

Hi there

Any idea why the exact same underlying data can result in such different outcomes for even simpler things like regular emas?

Hi, i just i noticed cosh - Vector Hyperbolic Cosine and qstick are not implemented and marked as available in readme.md, fyi.

Hello! Can you help me with the PSAR indicator? I can't get the same result as TradingView. I can't figure out what length the high and low input data should be?

I would suggest optionally executing the indicators on a separate worker/thread

Something like

ta.ema(close, 20, { useWorker: true, }).then(data => console.log(data));

This would allow calculating many indicators with better performance and without blocking the main Node.js/Browser process.

References
https://nodejs.org/api/worker_threads.html

Hi

Trust you are doing good. We are a commercial project looking for talents like yourself.

How can we contact you? We want to avail your services for a commercial project. We are interested in similar work but somewhat very different which focuses on indicator optimizations

You can email us at [email protected]

I would suggest adding unit tests starting from the most used indicators

Hi again,

I have been trying to get the MACD working like on my tradingview but to no avail.

Tradingview's macd shows last 4 histograms values as -3, 1.3, 0.9, 3 vs what I get is -0.3164, -0.25768, -0.24459, -0.13909.

I configured my macd in the library as ta.macd(close1m, 12, 26, 9) and passing the close bars of the btc 1min klinedata and I belive the configuration is as it should according to my tradingview setup

Can you perhaps tell me what I am doing wrong?

While using IndicatorsNormalized and a indicator the returned data is not being processed correctly by other indicators, because the normalization returns a array with empty/undefined/NaN, which breaks the logic, see example below:

const close = [0, 10, 20, 30, 40, 0, 10, 20, 30, 40, 0, 10, 20, 30, 40, 0, 10, 20, 30, 40, 0, 10, 20, 30, 40, 0, 10, 20, 30, 40, 0, 10, 20, 30, 40, 0, 10, 20, 30, 40, 0, 10, 20, 30, 40, 0, 10, 20, 30, 40]
const ta = new IndicatorsNormalized()
const rsiSma = await ta.sma(await ta.rsi(close, 7), 20)
console.log(rsiSma) // returns [18 x empty items, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN]

Instead the other indicators should skip empty/undefined/NaN values in the calculation.