Right here is crucial components of the brand new demo-script MultiExchangeWatch.mq5 (it is supplied together with the beta-version of the library).
NB: If the script is working very first time, it can ask to unpack (manually) CCXT Software Server (extracted as ccxtappsrvbundle.jsc from built-in useful resource), and run Node.js with it.
First, embody the headers.
#embody "ccxtjsmtlib.mqh" #embody "ccxtutil.mqh" #embody <MQL5Book/Feedback.mqh>
Within the inputs, the Node server setup needs to be performed.
enter group "Connection settings" enter string NodeServer = "http://127.0.0.1:8124"; enter string NodeAuth = "";
Subsequent, specify quite a few exchanges you need to monitor, a ticker, and a kind of the watch. By default, the script watches for order books for BCH/USDT.
To fill in these inputs correctly with most well-liked values, it is best to in all probability must output the record of supported exchanges and their markets beforehand. It may be considered in one other instance script CcxtAppSrvShowcase.mq5, supplied with the lib.
enter string Exchanges = "ascendex,bitmart,binance"; enter string Ticker = "BCH/USDT"; enter string Watch = "watchOrderBook"; enter uint WatchingDuration = 10;
Then OnStart occasion handler does its job. Inline feedback clarify the method. The imported features, lessons and strategies from the library are highlighted in yellow.
string Alternate[]; void OnStart() { PrintFormat("CCXT AppSrvLibrary model: %.2f", AppSrvLibraryVersion()); const static string standing[] = {"Cannot deploy", "App server ZIP is deployed, however not extracted", "App server information are deployed"}; const int d = DeployCcxtAppServer(); Print(standing[d + 1]); if(d <= 0) { return; } SetNodeServer(NodeServer, NodeAuth); CcxtLink *hyperlink = GetLink(); AutoPtr<CcxtJsExchangeProIntf> ccxt[]; const int n = StringSplit(Exchanges, ',', Alternate); ArrayResize(ccxt, n); for(int i = 0; i < n; i++) { ccxt[i] = CreateExchangePro(Alternate[i]); if(hyperlink.getLastHttpCode() != 200 || !ccxt[i][] || ccxt[i][][].t >= JS_NULL) { Print("Building failed for change: ", Alternate[i]); return; } const bool isPro = !!*ccxt[i][]["pro"]; if(!isPro) { PrintFormat("WARNING! %s is not PRO, there is no such thing as a websocket help", Alternate[i]); } if(!ccxt[i][]["has"][Watch].get<bool>()) { PrintFormat("WARNING! %s doesn't help '%s' subscriptions", Alternate[i], Watch); } } int lively = 0; for(int i = 0; i < n; i++) { if(ccxt[i][].improve()) { if(!ccxt[i][].watchAnything(StringFormat("%s("%s")", Watch, Ticker))) { PrintFormat("Cannot begin %s for %s", Watch, Alternate[i]); ccxt[i][].shut(); ccxt[i] = NULL; } else { lively++; } } else { if(ccxt[i][].isConnected()) { Print("Cannot improve to websockets"); string headers[][2]; if(ccxt[i][].ws().getHeaders(headers)) { } ccxt[i][].ws().shut(); ccxt[i][].shut(); ccxt[i] = NULL; } } } if(!lively) return; PrintFormat("* Monitoring %d subscriptions", lively); const uint begin = GetTickCount(); whereas(!IsStopped() && (!WatchingDuration || GetTickCount() - begin < WatchingDuration * 1000)) { for(int i = 0; i < n; i++) { if(ccxt[i][] && ccxt[i][].isConnected()) { AutoPtr<JsValue> j = ccxt[i][].readMessage(false); if(j[]) { ChronoComment(j[].stringify(0, 0)); } } } } Print("* Unsubscribing..."); for(int i = 0; i < n; i++) { if(ccxt[i][] && ccxt[i][].isConnected()) { ccxt[i][].un().watchAnything(StringFormat("%s("%s")", Watch, Ticker)); } } GracefullClose(ccxt); // not introduced right here within the weblog GracefullClose(ccxt, 5, true); Remark(""); }
When the script is working, the record of incoming order books (json-messages) is outputted and actively up to date on the chart.
On high of such a dataflow it is simple to implement varied arbitrage methods and calculate mixed statistics.