Write your own exchange scraper

Add your own scraper

Before you begin writing a scraper, please check whether the exchange under consideration offers integration through websockets. In this case please implement your scraper using the websocket instead of a RESTful API.

Now, let's assume you want to scrape a data source that provides trade information. Create a new file in exchange-scrapers/ and call it MySourceScraper.go. In order for us to be able to call your scraper from our system, you should introduce a type MySourceScraper struct that implements the APIScraper interface from the scrapers package:

type APIScraper interface {
// ScrapePair returns a PairScraper that continuously scrapes trades for a
// single pair from this APIScraper
ScrapePair(pair dia.Pair) (PairScraper, error)
// FetchAvailablePairs returns a list with all available trade pairs (usually
// fetched from an exchange's API)
FetchAvailablePairs() (pairs []dia.Pair, err error)
// Channel returns a channel that can be used to receive trades
Channel() chan *dia.Trade

From the MySourceScraper type you derive a MySourcePairScraper type which restricts the scraper to a specific pair. Next, you should write a function with signature NewMySourceScraper(exchangeName string) *MySourceScraper initializing a scraper. We suggest that this function calls a method func (s *MySourceScraper) mainLoop() in a go routine, constantly receiving trade information through the trade channel of MySourceScraper as long as the channel is open. The collection of new trading information inside the mainLoop() should be done by an update method with signature func (s *MySourceScraper) Update(). Finally, in order to implement the interface APIScraper you should include ScrapePair returning a MySourcePairScraper for a specific pair, so our main collection method can iterate over all possible trading pairs.

Also, please take care of proper error handling and cleanup. More precisely, you should include a method Error() which returns an error as soon as the scraper's channel closes, and methods Close() and cleanup() handling the closing/shutting down of channels.

Furthermore, in order for our system to see your scraper, add a reference to it in Config.go in the dia package, and to the switch statement in APIScraper.go in the scrapers package:

func NewAPIScraper(exchange string, key string, secret string) APIScraper {
switch exchange {
case dia.MySourceExchange:
return NewMySourceScraper(key, secret, dia.MySourceExchange)

Before running the scraper install the redis database on localhost and execute main.go from cmd/services/pairDiscoveryServices. Then, collector.go in the folder cmd/exchange-scrapers/collector will try to create a scraper for each exchange and collect the data pairs present in config/exchange-scrapers.json written by the method fetchAvailablePairs().

Finally, run the scraping executable flagged as follows:

go run collector.go -exchange MySource

For an illustration you can have a look at the KrakenScraper.go.