GDAX api, create candle graph

Question

I would like to create a candle graph using GDAX api. I am currently using the HTTP request for historical data https://docs.gdax.com/#get-historic-rates bug this is marked that I should use websocket API. Unfortunately I don't know how to handle historic data through Gdax websocket api https://github.com/coinbase/gdax-node. Could someone help me ?

Answer 1

Here are 1m candlesticks built from match channel using GDAX websocket

"use strict";
const
  WebSocket = require('ws'),
  PRECISION = 8

function _getPair(pair) {
  return pair.split('-')
}

let ws = new WebSocket('wss://ws-feed.pro.coinbase.com')

ws.on('open', () => {

  ws.send(JSON.stringify({
    "type": "subscribe",
    "product_ids": [
      "ETH-USD",
      "BTC-USD"
    ],
    "channels": [
      "matches"
    ]
  }))
})

let candles = {}
let lastCandleMap = {}

ws.on('message', msg => {
  msg = JSON.parse(msg);
  if (!msg.price)
    return;

  if (!msg.size)
    return;

  // Price and volume are sent as strings by the API
  msg.price = parseFloat(msg.price)
  msg.size = parseFloat(msg.size)

  let productId = msg.product_id;
  let [base, quote] = _getPair(productId);

  // Round the time to the nearest minute, Change as per your resolution
  let roundedTime = Math.floor(new Date(msg.time) / 60000.0) * 60

  // If the candles hashmap doesnt have this product id create an empty object for that id
  if (!candles[productId]) {
    candles[productId] = {}
  }


  // If the current product's candle at the latest rounded timestamp doesnt exist, create it
  if (!candles[productId][roundedTime]) {

    //Before creating a new candle, lets mark the old one as closed
    let lastCandle = lastCandleMap[productId]

    if (lastCandle) {
      lastCandle.closed = true;
      delete candles[productId][lastCandle.timestamp]
    }

    // Set Quote Volume to -1 as GDAX doesnt supply it
    candles[productId][roundedTime] = {
      timestamp: roundedTime,
      open: msg.price,
      high: msg.price,
      low: msg.price,
      close: msg.price,
      baseVolume: msg.size,
      quoteVolume: -1,
      closed: false
    }
  }

  // If this timestamp exists in our map for the product id, we need to update an existing candle
  else {
    let candle = candles[productId][roundedTime]
    candle.high = msg.price > candle.high ? msg.price : candle.high
    candle.low = msg.price < candle.low ? msg.price : candle.low
    candle.close = msg.price
    candle.baseVolume = parseFloat((candle.baseVolume + msg.size).toFixed(PRECISION))

    // Set the last candle as the one we just updated
    lastCandleMap[productId] = candle
  }

})

Answer 2

What they're suggesting is to get the historic rates https://docs.gdax.com/#get-historic-rates from this endpoint, then keep your candles up to date using the Websocket feed messages - whenever a 'match'/ticker message is received, you update the last candle accordingly.

From the docs: "The maximum number of data points for a single request is 300 candles. If your selection of start/end time and granularity will result in more than 300 data points, your request will be rejected." so probably why you can't get more than 2 days worth of data.

ps. I have a live orderbook implemented here and a basic candle stick chart - lots not hooked together fully still but its at least available for preview until its complete https://github.com/robevansuk/gdax-java/