Skip to content

Latest commit

 

History

History
78 lines (58 loc) · 5.34 KB

kite_connect.md

File metadata and controls

78 lines (58 loc) · 5.34 KB
Raw

R Package for Kite Connect API

Prodipta Ghosh 2018-12-18

Kite Connect is a set of REST-like HTTP APIs from Zerodha that expose many capabilities required to build a complete stock market investment and trading platform. It lets you execute orders in real time (equities, commodities, mutual funds), manage user portfolios, stream live market data over WebSockets, and more. This R package wraps a select set of functionaliies of the version 3 (latest) of this API. This package allows one to do the following:

  • Place, modify and cancel orders
  • Query specific order details, as well query holdings, positions, available margins and other related portfolio management operations.
  • Obtain market quotes as well as historical data

The parts of the API that this package does not implement:

  • Websocket streaming (may be in future!).
  • Mutual fund related functionalities (may be in future!).

Requirements

The use of the API is based on Kite Connect app based authentication, using api_key and api_secret. This requires subscription to the API and registering an app. After an app is registered, the api_key and the api_secret will be available on the app page. In addition, to use the historical data option, one needs an additional subscription to the historical data API.

Workflow

The workflow starts with the creation of a kite connect object and completing a successful login flow. To create a kite connect object the following may be used. The login url then may be obtained using the get_login_url function.

require(kiteconnect3)
kite = create_connection_object(list(api_key="xxxx",api_secret="yyyy"))
login_url = get_login_url(kite)
print(login_url)

Navigating to this login url and completing a successful login will generate a request_token. A successful login comes back with the request_token as a URL query parameter to the redirect URL registered on the developer console for that api_key. It is not necessary to have a service running at the redirection URL to capture the request_token. In fact, individual developers can simply specify the local host IP (127.0.0.1) (or anything for that matter) as the redirect URL. After manually navigating and entering the login credentials, the browser will be redirected with the request_token in the URL as a query parameter, which can simply be copied from the browser address bar. This request_token may be used for subsequent handshake with the API.

API Handshake

Once the request_token is available, the set_request_token function can be used to update the kite connect object and then call the fetch_access_token to complete the API handshake and obtain an access_token. A successful handshake will set the access_token slot inside the kite connect object and return the object itself. From this point onward the handshake is complete and the API functions are available to the user.

kite = set_request_token(kite,request_token)
kite = fetch_access_token(kite)

Example of Use

The following displays the call signatures for some useful functions. For details on each function, see help (e.g. ?place_order).

# portfolio functions
orders <- get_orders(kite)
trades <- get_trades(kite)
positions <- get_positions(kite)
holdings <- get_holdings(kite)
order_hist <- order_history(kite,orders$order_id[1])
specific_trades <- order_trades(kite,orders$order_id[1])
margins <- get_margins(kite,"equity")
# trade functions
order_id <- place_order(kite,"NFO","NIFTY18MARFUT","SELL",75)
order_id <- modify_order(kite,"regular",order_id,quantity = 150)
order_id <- cancel_order(kite,"regular",order_id)
# market data functions
instruments <- get_instruments(kite,"NFO")
q <- get_quote(kite,c("NFO:NIFTY18FEBFUT"))
ohlc <- get_ohlc(kite,c("NFO:NIFTY18FEBFUT","NFO:NIFTY18MARFUT"))
ltp <- get_ltp(kite,c("NFO:NIFTY18FEBFUT","NFO:NIFTY18MARFUT"))

Important Points to Note

Please note, this package is intended to provide low level functionalities of the API. Most of the functions do minimum error checks and input validations. As a result, output can be different from what you expected. If you are using this package, be aware of this and use necessary error handling and validation and test this works as per your requirement. Also use some functions with cautions - for example the market data functions (e.g. get_quote). The API possibly has a API call rate restrictions and polling for market data too frequently will probably ban your IP and/ or api_key. It is best to use the websocket streaming functionality (not implemented here, may be in future!) if you need continuous market update. Also, querying about the status of an order is most efficiently done by using a postback or using the websocket route. Also check the licesing.

Welcome to systematic trading in R with Kite Connect. In case you are interested to learn more about systematic and quantitative trading from the domain experts, you can check out QuantInsti or their online platform Quantra.

"Do or do not, there is no try"