Financial markets have periods of illiquidity. If your plugin receives a request for a bar interval that has no volume or trades, do not return dummy values with zero prices. Instead, carry forward the previous close price or let AmiBroker pad the gaps automatically by returning the accurate count of valid records. 6. Compilation and Deployment To compile your source code into a functional .dll file:
Because AmiBroker is a 32-bit or 64-bit multi-threaded application, thread safety is paramount. Developers must use mutexes or critical sections when accessing shared data structures to prevent crashes. Furthermore, memory management must be impeccable; leaking memory in a data plugin will eventually lead to system instability, especially during long trading sessions where millions of ticks may be processed.
struct Quotation DATE DateTime; float Price; float Open; float High; float Low; float Volume; float OpenInterest; ; Use code with caution. amibroker data plugin source code top
This is what separates basic plugins from "top" ones.
This is the heart of the data plugin. When a user opens a chart or runs a backtest, AmiBroker calls GetQuotesEx to request a specific symbol's data array. Financial markets have periods of illiquidity
When a market tick arrives, the background thread parses the price data and pushes it into an internal circular queue.
: Telling AmiBroker whether the plugin supports intraday, daily, real-time, or historical backfills. receives this message
void GetSymbol(int index, char* symbol) /* return symbol */ int GetBarCount(const char* symbol) /* return bar count */ return 100; void GetBar(const char* symbol, int barIndex, float* open, float* high, float* low, float* close, float* volume) /* return bar data */ void GetQuote(const char* symbol, float* bid, float* ask) /* return quote */ ;
Every AmiBroker plugin must export these three core functions: GetPluginInfo
In this architecture, the Python script (like the vendor_class_wrapper.py in some projects) connects to the broker's API, fetches data, and builds bars. It then sends this formatted data as a JSON message over WebSocket to the C++ plugin. The plugin, in turn, receives this message, parses the JSON packet in situ for memory performance, and feeds it to AmiBroker. This setup allows the plugin to remain broker-agnostic, while the Python "adapters" can be swapped out or modified to connect to any data source.
