Got it, let's build out a fully working client-server setup for your MetaTrader (MQL4/5) + Python + ZeroMQ sync. You already have a solid Python server foundation, so I'll pair that with a tested MQL client and walk through all the key details to get you up and running.

Python ZeroMQ Server (REP Socket) #

Your existing code is a great start—here's the full, polished version that handles string messages cleanly and maintains the request-response cycle:

import zmq
import time

context = zmq.Context()
socket = context.socket(zmq.REP)
socket.bind("tcp://127.0.0.1:9999")
print("Server started, waiting for MQL client messages...")

while True:
    # Wait for incoming message from MQL client
    message = socket.recv_string()
    print(f"Received from MQL: {message}")
    
    # Add your custom processing logic here (e.g., data analysis, database calls)
    time.sleep(1)  # Simulate work
    
    # Send a response back to the MQL client
    response = f"Server processed: {message}"
    socket.send_string(response)
    print(f"Sent response: {response}")

I switched to recv_string() and send_string() instead of raw bytes because it plays nicer with MQL's string-based messaging, making debugging and data transfer smoother. The REP socket type ensures every request gets a response—critical for keeping your MQL and Python code in sync.

MQL4/5 ZeroMQ Client (REQ Socket) #

First, you'll need the ZeroMQ MQL binding installed:

• For MQL5: Open MetaEditor, go to Tools > MetaQuotes Language Market, search for "ZeroMQ", and install the official library.
• For MQL4: Download the ZeroMQ library from MetaQuotes' resources and place it in your MQL4/Libraries folder.

MQL5 Client Code #

//+------------------------------------------------------------------+
//|                                                  MQL_ZMQ_Client.mq5 |
//|                        Copyright 2024, MetaQuotes Software Corp. |
//|                                             https://www.mql5.com |
//+------------------------------------------------------------------+
#property copyright "Copyright 2024, MetaQuotes Software Corp."
#property link      "https://www.mql5.com"
#property version   "1.00"
#property strict

#include <Zmq/Zmq.mqh>

ZMQContext context;
ZMQSocket  socket;

//+------------------------------------------------------------------+
//| Expert initialization function                                   |
//+------------------------------------------------------------------+
int OnInit()
{
    if(context.Create() != ZMQ_OK)
    {
        Print("Failed to create ZMQ context: ", context.LastError());
        return INIT_FAILED;
    }
    
    if(socket.Create(context, ZMQ_REQ) != ZMQ_OK)
    {
        Print("Failed to create REQ socket: ", socket.LastError());
        context.Destroy();
        return INIT_FAILED;
    }
    
    if(socket.Connect("tcp://127.0.0.1:9999") != ZMQ_OK)
    {
        Print("Failed to connect to server: ", socket.LastError());
        socket.Destroy();
        context.Destroy();
        return INIT_FAILED;
    }
    
    Print("MQL client connected to Python server successfully");
    return INIT_SUCCEEDED;
}

//+------------------------------------------------------------------+
//| Expert deinitialization function                                 |
//+------------------------------------------------------------------+
void OnDeinit(const int reason)
{
    socket.Disconnect();
    socket.Destroy();
    context.Destroy();
}

//+------------------------------------------------------------------+
//| Expert tick function                                             |
//+------------------------------------------------------------------+
void OnTick()
{
    // Example: Send current symbol and bid price to the server on every tick
    string message = StringFormat("Symbol: %s, Bid: %.5f", _Symbol, SymbolInfoDouble(_Symbol, SYMBOL_BID));
    
    if(socket.Send(message, 0) != ZMQ_OK)
    {
        Print("Failed to send message: ", socket.LastError());
        return;
    }
    Print("Sent to server: ", message);
    
    // Wait for response from the Python server
    string response;
    if(socket.Recv(response, 0) != ZMQ_OK)
    {
        Print("Failed to receive response: ", socket.LastError());
        return;
    }
    Print("Received from server: ", response);
}
//+------------------------------------------------------------------+

MQL4 Client Code #

Nearly identical to MQL5, with minor syntax adjustments:

//+------------------------------------------------------------------+
//|                                                  MQL_ZMQ_Client.mq4 |
//|                        Copyright 2024, MetaQuotes Software Corp. |
//|                                             https://www.mql4.com |
//+------------------------------------------------------------------+
#property copyright "Copyright 2024, MetaQuotes Software Corp."
#property link      "https://www.mql4.com"
#property version   "1.00"

#include <Zmq/Zmq.mqh>

ZMQContext context;
ZMQSocket  socket;

//+------------------------------------------------------------------+
//| Expert initialization function                                   |
//+------------------------------------------------------------------+
int OnInit()
{
    if(context.Create() != ZMQ_OK)
    {
        Print("Failed to create ZMQ context: ", context.LastError());
        return(INIT_FAILED);
    }
    
    if(socket.Create(context, ZMQ_REQ) != ZMQ_OK)
    {
        Print("Failed to create REQ socket: ", socket.LastError());
        context.Destroy();
        return(INIT_FAILED);
    }
    
    if(socket.Connect("tcp://127.0.0.1:9999") != ZMQ_OK)
    {
        Print("Failed to connect to server: ", socket.LastError());
        socket.Destroy();
        context.Destroy();
        return(INIT_FAILED);
    }
    
    Print("MQL client connected to Python server");
    return(INIT_SUCCEEDED);
}

//+------------------------------------------------------------------+
//| Expert deinitialization function                                 |
//+------------------------------------------------------------------+
void OnDeinit(const int reason)
{
    socket.Disconnect();
    socket.Destroy();
    context.Destroy();
}

//+------------------------------------------------------------------+
//| Expert tick function                                             |
//+------------------------------------------------------------------+
void OnTick()
{
    string message = StringFormat("Symbol: %s, Bid: %.5f", Symbol(), Bid);
    
    if(socket.Send(message, 0) != ZMQ_OK)
    {
        Print("Send error: ", socket.LastError());
        return;
    }
    Print("Sent: ", message);
    
    string response;
    if(socket.Recv(response, 0) != ZMQ_OK)
    {
        Print("Receive error: ", socket.LastError());
        return;
    }
    Print("Received: ", response);
}
//+------------------------------------------------------------------+

Key Setup & Troubleshooting Tips #

• Socket Pairing: We're using ZMQ_REP (server) and ZMQ_REQ (client) because they enforce a strict request-response pattern—this is ideal for synchronous communication where MQL sends data and waits for Python to process it before moving on.
• Network Configuration: The example uses localhost (127.0.0.1) for testing. If you want to run the server on a remote machine, replace the IP with the server's public address and make sure port 9999 is open in both firewalls.
• Error Handling: The code includes basic error checking, but you can expand it to handle timeouts, reconnections, or invalid message formats based on your specific use case.
• Testing Order: Always start the Python server first, then launch the MQL client in MetaTrader—this ensures the client can connect immediately.

内容的提问来源于stack exchange，提问作者Kent Winston