In JoeFernando/rIG:

<<<<<<< HEAD

title: "Untitled" author: "Joe Fernando" date: "23 July 2018" output: html_document

knitr::opts_chunk$set(echo = TRUE)

Install and load up libraries

To get started you will need loging id/password from your IG account along with your api key. The current link is pointed at the Dev verson of the API.

The usual warnings apply for use of this library, I built this for my personal use, there is no regular support for this library. I have published this to help and advance the R community. Use at own risk.

To get started install rIG along with its dependancies:

library(ghit)
install_github("JoeFernando/rIG")

library(rIG)
library(tidyverse)
library(lubridate)
library(stringr)
library(jsonlite)
library(httr)
library(rlist)
library(glue)

Login/Logout

b   <- '{"identifier":"your_login_id", "password": "insert_password_here"}'
api <- "insert_api_key_here"

ig_login(b, api) 

ig_logout() # this will end your session

Searching for Markets

A tibble is retuned for all matching markets:

ig_search("Copper")
ig_search("AUDUSD")

Time resolutions for history data

There is a helper time_resolution vector to list out the options available. Presently, SECONDis not working.

rIG::time_resolution

Downloading history

When you query the API the returned data gives both bid and ask. Hence, there is a two step process to getting data in the familiar ohlcv form, however the process is straighforward. If you prefer you could wrap the whole thing within a function. I prefer to have these separately. The result is presented in tibble format:

raw_dl <- ig_history(epic = "CS.D.GBPUSD.CFD.IP", mkt.name = "GBP/USD", 
                          from = "2018-07-02", to = "2018-07-03", res = "DAY")

ohlc_dl <- get_ohlc(raw_dl, "ask")                    # you can change between "bid" or "ask" - default is set for "ask"
ohlc_dl_rename <- get_ohlc_rename(raw_dl, "ask")     # this assigns mkt.name as prefix to ohlcv column names

Additionally, the mkt.name is set as the attribute name for the returned data. This is very handy to name within lists to hold data for analysis.

epic_names <- read_csv("epic_names.csv")   # table with epic names and instrument names


raw_list <- map2(epic_names$epic, epic_names$instrumentName, ~ig_history(epic = .x, mkt.name = .y,
                                                                   from = "2018-07-01", to = "2018-07-03", 
                                                                   res = "DAY"))

names(raw_list) <- map_chr(raw_list, ~attr(.x, "name")) # This is to name the tibbles within the list


ohlc_list         <- map(raw_list, ~get_ohlc(.x))
ohlc_list_renamed <- map(raw_list, ~get_ohlc_rename(.x))  


# Check names
names(raw_list)
names(ohlc_list)
names(ohlc_list_renamed)

I download all my data in one go, so to speed-up your download use the furrr package, change map2 to future_map2.

Updating EOD Data

Two functions are available to update EOD data, ig_eod and ig_eod_list. As I hold my data in a list, I use the latter function. The function is pretty simple, it assumes that you have a table containing epic codes and mkt.name as used in history download function. usage is as follows:

ig_eod <- (mkt.name,
to.date = (Sys.Date() + lubridate::ddays(1)),   # this is pre set to download up to the latest date
df.w.hist,                                      # tibble/df containing historic data
epic.table,                                     # this should be a df with 2 columns, epic and its name
epic.code.col_name = "epic",                    # these two epic.table field names 
epic.name.col.name = "instrumentName")



ig_eod_list <- (mkt.name,
to.date = (Sys.Date() + lubridate::ddays(1)),
list.of.df,                                   # list containing your data
epic.table,
epic.code.col_name = "epic",
epic.name.col.name = "instrumentName")



# This is how I use this function. All my historic data is held inlist "bb"
bbb <- map(names(bb), ~ig_eod_list(df.name = ., 
                                    list.of.df = bb,
                                    epic.table = epic_names))

names(bbb) <- purrr::map_chr(bbb, ~attr(.x, "name"))

=======